Busca

... Medical & Biological Engineering & Computing (2024) 62:12771311 https://doi.org/10.1007/s11517-024-03020-3 REVIEW ARTICLE Obstructive sleep apnea detection during wakefulness: a comprehensive methodological review Ali Mohammad Alqudah1 Ahmed Elwali2 Brendan Kupiak3 Farahnaz Hajipour4 Natasha Jacobson5 Zahra Moussavi1,3 Received: 25 June 2023 / Accepted: 11 January 2024 / Published online: 27 January 2024 The Author(s) 2024 Abstract Obstructive sleep apnea (OSA) is a chronic condition affecting up to 1 billion people, globally. Despite this spread, OSA is still thought to be underdiagnosed. Lack of diagnosis is largely attributed to the high cost, resource-intensive, and timeconsuming nature of existing diagnostic technologies during sleep. As individuals with OSA do not show many symptoms other than daytime sleepiness, predicting OSA while the individual is awake (wakefulness) is quite challenging. However, research especially in the last decade has shown promising results for quick and accurate methodologies to predict OSA during wakefulness. Furthermore, advances in machine learning algorithms offer new ways to analyze the measured data with more precision. With a widening research outlook, the present review compares methodologies for OSA screening during wakefulness, and recommendations are made for avenues of future research and study designs. Keywords Sleep apnea Classification Home sleep study OSA Polysomnography Questionnaire Screening wakefulness 1 Introduction * Zahra Moussavi Zahra.Moussavi@umanitoba.ca Ali Mohammad Alqudah alqudaha@myumanitoba.ca Ahmed Elwali aelwali@marian.edu Brendan Kupiak kupiakb@myumanitoba.ca Farahnaz Hajipour Hajipouf@myumanitoba.ca Natasha Jacobson Natasha.Jacobson@umanitoba.ca 1 Biomedical Engineering Program, University of Manitoba, 66 Chancellors Cir, Winnipeg, MB R3T 2N2, Canada 2 Biomedical Engineering Program, Marian University, 3200 Cold Sprint Road, Indianapolis, IN 462221997, USA 3 Electrical and Computer Engineering Department, University of Manitoba, 66 Chancellors Cir, Winnipeg, MB R3T 2N2, Canada 4 SkipTheDishes, Winnipeg, Canada 5 Biosystems Engineering Department, University of Manitoba, 66 Chancellors Cir, Winnipeg, MB R3T 2N2, Canada Sleep apnea is a prevalent respiratory disorder defined as periods of airflow cessation (apnea) or reduced airflow by more than 30% (hypopnea) associated with at least a 3% drop in blood oxygen saturation level (SpO2) [1, 2]. Sleep apnea is differentiated into three types: obstructive, central, and mixed apnea. Obstructive sleep apnea (OSA) is the most common type, accounting for more than 75% of cases. In contrast, central sleep apnea is relatively rare (about 5% of cases), with 20% of cases presenting with a mix of central and OSA [1, 2]. OSA is characterized by repeated [3] episodes of complete (apnea) and partial (hypopnea) cessation of breathing due to pharyngeal collapse during sleep [3]. OSA severity is measured by the apnea/hypopnea index (AHI). AHI is the number of apneic and hypopneic events per hour of sleep. To constitute an event, apnea/hypopnea should last at least 10 s and be associated with more than a 3% drop in blood oxygen saturation [4]. An OSA event may be accompanied by an increased heart rate and blood pressure, as well as subsequent arousals from sleep to restore the upper airway (UA) patency [3]. Signs and symptoms of OSA include daytime sleepiness, tiredness, depression, morning headaches, nighttime gasping and choking, and Vol.:(0123456789) 1278 snoring. While it is debatable whether OSA causes other health issues, OSA has been linked with cardiovascular and cerebrovascular disease [58], reduced attention, an increased risk for car accidents [9], perioperative morbidity, and post-surgery mortality [10]. Commonly, AHI ratings between 05, 515, 1530, and > 30 are referred to as non-, mild-, moderate-, and severe-OSA, respectively [10]. 1.1Brief history of sleep apnea research during wakefulness Research suggests that a significant percentage of the OSA population (> 80%) remains undiagnosed and untreated [10, 11]. In the USA, the direct and indirect costs associated with untreated OSA are estimated to be between $65 and $165 billion, annually [12]. Under-diagnosis and, in turn, treatment delays for severe cases of OSA may further increase indirect costs while increasing the risk of associated vehicle accidents [13]. The main reasons for OSA under-diagnosis are the limited resources in healthcare (i.e., sleep laboratories, sleep technicians) and the inefficiency of existing diagnostic tooling. The gold standard diagnostic assessment of OSA is an overnight full polysomnography (PSG), in which more than 15 biological signals such as brain waves, respiratory flow, blood oxygen saturation, heart and muscle signals, and snoring sounds are recorded and monitored by a certified sleep technician at a sleep center. Though PSG remains the most accurate OSA diagnostic test, it is an expensive, time-consuming, and labor-intensive procedure. As a result, there is typically a long waiting list (occasionally up to 1 year [14]) to conduct an overnight PSG study. The second-best diagnostic assessment is the home sleep test (HST): a portable, simplified version of the PSG device that allows patients to self-monitor overnight at home [14] although the diagnosis is being made by a specialist after screening data of the HST [14]. The HST records a significantly smaller number of signals (34) than PSG and is less precise as it relies on the patient to perform the recording themselves. A clinically approved HST device records at least three main signals: nasal airflow, EEG to detect sleep stage, and blood oxygen saturation. Beyond PSG and HST studies, there also exist several overnight acoustic sleep studies that use tracheal breathing sounds in addition to blood oxygenation; they have resulted in high accuracy (> 96%) compared to gold standard PSG [15]. Given the significant time and labor-intensive demand of PSG and HST, daytime testing (during wakefulness) to detect OSA is a very desirable alternative. During wakefulness, OSA is clinically assessed by questionnaires such as STOP-BANG [16] or the Epworth Sleepiness Score [17]. Despite the simplicity of the questionnaires over traditional overnight testing, OSA risk estimation by such questionnaires has very low specificity around 25%, thus, implying a Medical & Biological Engineering & Computing (2024) 62:12771311 very high false-positive rate [18, 19]. However, many clinicians, particularly anesthesiologists, continue to use quick OSA assessment questionnaires as OSA is a major risk factor for complications after general anesthesia [20]. The history of OSA during wakefulness using other methods than the abovementioned questionnaires has been intricately tied to the exploration of anatomical structures and physiological changes, primarily through the lens of advanced imaging studies [21]. Hypotheses initially centered on the role of upper airway collapsibility and the anatomical predispositions contributing to its occurrence during wakefulness. With the advent of imaging techniques such as magnetic resonance imaging (MRI) and computed tomography (CT), researchers gained the ability to visualize and analyze the upper airways structural complexities in both OSA and non-OSA subjects [22]. These imaging studies have provided insights into the differences in the upper airway dimensions, soft tissue characteristics, and anatomical variations between OSA and non-OSA groups, offering critical clues to the pathophysiology of OSA during wakefulness. Subsequent investigations have continued to refine our understanding of the intricate interplay between anatomical predispositions and the development of OSA, fostering advancements in diagnostic approaches and therapeutic interventions for this complex sleep-related disorder [21, 22]. Thus, several research groups have attempted to analyze breathing or vocal sounds recorded in a few minutes during wakefulness to estimate OSA severity based on the anatomical differences effects on these sounds. Contrary to OSA detection using breathing sounds during sleep, the challenge of wakeful OSA detection is the lack of noticeable breathing difficulty, even during exercise, while an individual is awake [23]. This is most likely attributed, in general, to an increase in the UA dilator muscles activity (especially for the genioglossus muscle) during wakefulness, which compensates for the physiological UA changes due to OSA [2224]. Nevertheless, imaging studies during wakefulness have confirmed the existence of morphological and mechanical differences in individuals with various OSA severities [21, 22, 24]. The sleep-related narrowing and increased compliance or collapsibility of the UA are critical contributors to the pathogenesis of OSA [25]. Further, compared to their healthy counterparts, individuals with OSA have been shown to have an increased pharyngeal length, a thick posterior, a long and thick soft palate, and a more compliant airway [22, 24]. Overall, OSA individuals tend to have a circular velopharynx shape rather than an elliptic shape with the long axis oriented in the lateral plane as in non-OSA individuals [21]. To navigate the morphologic changes of OSA, the UA anatomy is illustrated in Fig. 1. As mentioned above, despite these morphological and structural changes in the UA, OSA individuals do Medical & Biological Engineering & Computing (2024) 62:12771311 1279 Fig. 1Anatomy of the upper airway [26] (CC BY 4.0) not experience any breathing difficulty while awake. However, given that breathing and vocal sounds are generated by the f low of air in the UA, it has been hypothesized [27] that the breathing sounds and vocal sounds reflect UA abnormalities during wakefulness. As such, OSA prediction during [2631] wakefulness has received significant attention in recent years, leading to the development of seven notable OSA detection technologies: (1) imaging techniques, (2) negative expiratory pressure, (3) facial image landmarks, (4) acoustic pharyngometer and nasal airway pressure, (5) Table 1List of abbreviations used in this paper in alphabetical order breathing sound analysis, (6) speech signal analysis, and (7) questionnaires. 1.2Objective With the increasing demand for efficient and accurate OSA screening tools during wakefulness, it is the objective of the present manuscript to review existing OSA detection technologies during wakefulness and suggest avenues for future research. A list of acronyms used throughout this paper is presented in Table 1. Term Definition Term Definition AHI AUC BMI CNN CT CPAP EEG ESS FVC GA GMM HST LASSO LDA LR MFCCS MLP Apnea/hypopnea index Area under curve Body mass index Convolution neural network Computed tomography Continuous positive airway pressure Electroencephalography Epworth Sleepiness Scale Flow-volume curve Genetic algorithm Gaussian mixture model Home sleep test Least absolute shrinkage and selection operator Linear discriminant analysis Logistic regression Mel Frequency Cepstral Coefficients Multi-layer perceptron MRI NEP ORP OSA PSG QI RF ROC REM SDB SpO2 SVM SVR TSI UA UARS Magnetic resonance imaging Negative expiratory pressure Odds ratio products Obstructive sleep apnea Polysomnography Quantitative index Random forest Receiver operating characteristics Rapid eye movement Sleep-disordered breathing Mean oxygen saturation Support vector machine Support vector regression Tracheal sound intensity Upper airway Upper airway resistance syndrome 1280 2Methods We included all journal papers related to OSA screening in adults during wakefulness; among conference papers, only those with results not published in a journal paper were included. The search for relevant papers in this review encompassed the period from 1980 to 2023. Relevant manuscripts were identified by searching Google Scholar, PubMed, Science.gov, IOPscience, ScienceDirect, MPDI, Hindawi, and Taylor & Francis. The collection of papers was performed by four researchers (co-authors) through an online shared folder. The search strategy involved the utilization of combined keywords related to sleep apnea and various aspects of screening, daytime symptoms, wakefulness, and specific diagnostic techniques such as polysomnography, acoustics, imaging, and breathing. Figure 2 shows the flow diagram of the searching algorithm. Notably, the search approach emphasized the combination of keywords rather than their isolated use to ensure a comprehensive and Fig. 2Flow diagram of paper inclusion and exclusion Medical & Biological Engineering & Computing (2024) 62:12771311 targeted search process. The search was performed using a combination of the following keywords: sleep apnea, screening, daytime, wakefulness, questionnaire, diagnosis, polysomnography, acoustics, imaging, and breathing. A sample of queries used during collecting papers is sleep apnea AND screening AND diagnosis, sleep apnea AND wakefulness AND acoustics, sleep apnea AND wakefulness AND breathing, and sleep apnea AND wakefulness AND imaging. The assessment of search outcomes was broadened, ceasing when no more pertinent results were evident, with the recognition that significant findings were usually discovered within the initial pages of the search results. The inclusion of papers in the preliminary set was based on a hierarchical assessment, beginning with an examination of titles indicating potential relevance and further scrutiny of corresponding abstracts, which resulted in an initial set of 7147 papers reduced to 1365 after removing duplications. The selection process of the papers involved a meticulous review by at least two co-authors for each technology and ensuring the Medical & Biological Engineering & Computing (2024) 62:12771311 inclusion of high-quality and pertinent research. The following exclusion criteria were applied: 1. Exclusion of sleep apnea studies in children because OSA pathology is different in adults and children. 2. Exclusion of studies during sleep since the focus of this paper is detection during wakefulness. 3. Exclusion of studies using invasive techniques because the focus of the paper is on non-invasive methods. After applying the above exclusion criteria, the number of included papers was reduced to 143. From those, 71 fulltext articles were included after applying other exclusion criteria such as removing those with incomplete or missing data, lack of full-text access, or publication in a language other than English. The remaining articles were then passed through an assessment process to focus solely on OSA detection during wakefulness without any analysis during sleep, except for polysomnography (PSG) for the apneahypopnea index (AHI), which is a calculation often used for accuracy measure of the algorithms. As a result, 57 papers were selected for review in this manuscript. These papers highlight various techniques such as breathing sound analysis (n = 11), speech signal analysis (n = 11), imaging techniques (n = 11), facial image landmarks (n = 5), pharyngometry and nasal airway pressure (n = 6), negative expiratory pressure analysis (n = 6), and analysis of OSA-related questionnaires (n = 7). The outcomes of the techniques used in each category are presented separately, followed by a general discussion. Figure 3 shows the distribution of articles over topics selected in this paper. Quesonnaires 12% Negave Expiratory Pressure 11% Breathing Sounds Analysis 19% Phyrangometery and nasal airway pressure 11% Facial Images landmarks 9% Speech Signals 19% [CATEGORY NAME] [PERCENTAGE] Fig. 3Distribution of articles over topics in this review 1281 It should be noted that not all the studies reviewed in this manuscript intended to detect OSA; for example, the imaging studies only investigated the UA physiological and structural changes during wakefulness without the intention of imaging to develop a screening tool for OSA detection. Regardless, these imaging studies are reviewed here because their findings support further research on OSA screening technologies during wakefulness. Since the imaging studies findings have provided the main rationale for OSA screening technologies during wakefulness, we present those findings first and then present existing technologies for OSA. Figure 4 shows a graphical summary of the techniques used for OSA detection during wakefulness that are covered and discussed in this paper. In this review, the reason for not employing the PRISMA guidelines is mainly due to our focus on OSA non-invasive screening techniques during wakefulness in adults. The inclusion criteria were designed to include both journal and conference papers with unpublished results in the journal, while exclusion criteria were designed to exclude studies that involved children, invasive techniques, and sleep analysis. The developed search methodology utilized a combination of keywords which were different from the strict PRISMA, ensuring that the outcomes aligned with the studys objectives and maintained methodological transparency. 2.1Imaging Different medical imaging techniques have long been key to providing insights into the deep anatomical and functional information about human organs, and through them, researchers can track changes in an organs size, shape, or dimensions. Several imaging modalities such as cephalometric X-rays, computed tomography (CT), ultrasound, magnetic resonance imaging (MRI), and endoscopy have been used to investigate the UA structural and morphological changes in patients with OSA. For diagnostic purposes using X-ray cephalometric, different physiological aspects and parameters may be measured via screening the mandibular deficiencies [28]. Commonly, cephalometric analysis is performed by measuring the angles and distances between various cephalometric landmarks, and it is very useful for detecting anatomical changes due to OSA in an individual, but During Wakefulness 1.Breathing Sounds Analysis 1.Speech Signals Analysis 1.Imaging Techniques Fig. 4OSA detection during wakefulness technologies 1.Facial Images landmarks 1.Phyrangomete ry 1.Negave Expiratory Pressure 1.Quesonnaires 1282 those changes are not enough to identify OSA accurately [28]. Nevertheless, imaging studies using X-rays on 15 participants have found the length of the soft palate to be longer, and position of the hyoid bone to be inferior, and a narrower posterior airway compared to those in healthy individuals [28]. Unlike X-rays, CT scans provide axial images, and volumetric scans and measurements instead of planar images. Different studies have used either traditional CT scans [29] or Cine CT (high-speed CT imaging) to study the UA changes due to OSA [30]. These studies explored the changes in the UA anatomy during wakeful and tidal breathing, and the presence of a sawtooth pattern in the flow-volume curves due to OSA [29] and the pharyngeal area during a wakeful respiratory cycle rack the respiratory cycle [30]. The results of the study in [29] showed that there is a strong relationship between pharyngeal area and OSA, where the mean area of the nasopharynx, oropharynx (the most severely narrowed part), and hypopharynx was significantly less in OSA individuals than those in healthy controls. Moreover, those regional measurements correlated with AHI and blood oxygen saturation levels without correlating with age or obesity. Also, the results, congruent with those findings reported in [28], showed no definitive evidence relating a sawtooth pattern in FVC and OSA. Also congruent with the outcomes of the study in [28], the results of [29] showed that low retropalatal region presented the greatest difference due to OSA, and the median minimal airway size was smaller in snorers than that in controls. However, that difference was not statistically significant and was also found to be correlated with body mass index (BMI), neck circumference, AHI, and blood oxygen saturation level. Furthermore, it was noticed that the OSA patients UA had overall larger dimensional changes during the respiratory cycle than that of both the snorer and control groups, thus, indicating a more distensible UA in OSA individuals than that in healthy ones. Finally, as expected, most of the dimensional changes occurred during expiration, during which the airway expanded greatly and then collapsed [28]. Lately, MRI scans have been used to study the UA changes in OSA subjects. Compared to CT scans, MRI provides higher resolution and more details about the anatomical structures in the human especially the soft tissues. Also, like CT, MRI can provide volumetric images and measurements but with higher resolution. An imaging study was conducted during sleep as well as wakefulness using ultrafast MRI (one image per 0.8 s); the study aimed to examine the UA structure between 17 OSA individuals and their age-matched 8 healthy controls [31]. The major finding of the study was that the velopharynx of apneic individuals was smaller than that of healthy ones during the respiratory cycle. The variation of the velopharynx area was greater in apneic patients, particularly during Medical & Biological Engineering & Computing (2024) 62:12771311 sleep. The authors suggested this difference could be due to the increased compliance of the velopharynx in apneic individuals. Additionally, the apneic individuals during sleep showed a more circular velopharynx. Overall, it was concluded that changes in the velopharynx area and diameter during the respiratory cycle were greater in apneic individuals than those in healthy controls; this trend was more pronounced during sleep [31]. Another research examined the prognostic value of the lateral pharyngeal wall (LPW) thickness for the detection of OSA using two both US and MRI [32]. One hundred individuals with and without OSA (36 healthy controls and 64 OSA subjects) were enrolled in the study and performed an overnight PSG, and then the LPW thickness was measured using 1.5-T MRI and ultrasound during wakefulness. The ultrasound assessment was conducted during rest and Mllers maneuver. The MRI results showed a significantly greater LPW thickness in the OSA group, while the ultrasound results showed a significant difference between the two groups only during the left side with Mllers maneuver. Also, in general, a significant correlation was observed between LPW thickness and BMI, where patients with high BMI showed higher LPW thickness. Moreover, in terms of sex, males showed higher LPW thickness than females either in MRI or ultrasound assessment using left-sided Mllers maneuver. Overall, AHI was correlated with LPW thickness, and the obstruction severity of LPW was correlated with LPW thickness; in addition, the LPW collapse was significantly correlated with AHI. Finally, 93% effectiveness in OSA prognostication was achieved using anthropometric data and the LPW thickness measured by ultrasound and 89% effectiveness using only LPW thickness. Moreover, using the MRI for detecting OSA during wakefulness and LPW-based obstruction resulted in 90% and 84% accuracy, respectively. Ultrasound data analysis successfully detected LPW-based collapse severity in 67% of cases. Another study focused on comparing the findings of the drug-induced sleep endoscopy (DISE) with the modified Mallampati score and Mllers maneuver evaluation [33]. The comparison was done based on noseoropharynxhypopharynxlarynx (NOHL) for 43 individuals with moderate to severe OSA. The results showed the degree of collapsibility was significant only at the hypopharyngeal level, where 41.8% of the individuals during wakefulness and 88.3% in DISE (p < 0.0001) showed a hypopharyngeal obstruction. Moreover, 18.6% and 4.6% of the subjects showed laryngeal obstruction during wakefulness and DISE examination, respectively. However, the DISE succeeded in identifying the incidence of multilevel collapses (p = 0.001), while the incidence of oropharyngeal obstruction in patients with Mallampati scores I and II was significantly higher in Medical & Biological Engineering & Computing (2024) 62:12771311 DISE compared to that measured by Mllers maneuver (p = 0.021). Some other studies have focused on the ability to use ultrasound or MRI imaging of the tongue for OSA prediction [32, 34]. One research [34] showed the ability of OSA prediction using AI applied to ultrasound imaging of the UA and subcutaneous adipose tissues (SAT) in the regions of the neck, chest, and abdomen measurements [35]. The data was collected from 100 individuals, 36 without OSA and 64 with different OSA severity (32 mild, 32 moderately severe) based on their overnight PSG scores, while the DISE was used to determine the obstruction location and configuration. The results showed using the SAT ultrasound and anthropometric data; the oropharyngeal and tongue-based obstructions could be predicted with 64% and 72% accuracies, respectively. In oropharyngeal obstruction prediction, the most important features were found to be BMI, abdominal and hip circumferences, and submental SAT and SAT above the second intercostal space on the left. Furthermore, for tongue-based obstruction, the most important features were found to be height; SAT measured 2 cm above the umbilicus and submental SAT. Overall, the OSA prediction using the parameters mentioned above had a sensitivity of 100% and a specificity of 91.7%. The second research [34] extracted geometrical parameters of the tongue from ultrasound and MRI images in OSA subjects based on sex, age, and BMI among 100 individuals (64 with OSA and AHI 5). The quadratic discriminant analysis was performed, and the results showed males compared to females had higher tongue volumes and axial diameter during Mllers maneuver of ultrasound and coronal diameter of the MRI. All the examined MRI parameters were found significantly correlated with AHI among females with OSA; also, BMI showed a stronger correlation with AHI in females than in males. Using tongue parameters and anthropometric values, ultrasound analysis showed a sensitivity of 94% and a specificity of 98%, while MRI analysis showed 56% sensitivity and 92% specificity. Another study used MRI to analyze upper airway changes during tidal breathing in the OSA group and healthy controls [36]. The study used dynamic MRI where subjects were free to breathe during acquisition. While overall structure differences were minimal, OSA subjects had a narrower airway at a specific level. Significant variations in the upper airway size change over tidal breathing were observed in the OSA group, particularly in the low retropalatal/high retroglossal region during wakefulness and sleep. In OSA subjects, the collapsed airway during sleep aligned with the region showing the greatest changes in caliber while awake with tidal breathing. These results suggest a potential application for dynamic OSA imaging during wakefulness. 1283 In the surveyed literature, three studies used MRI imag- ing to observe the UA regions during wakefulness [31, 32, 36, 37]. The findings in [36] were consistent with those using CT imaging [29, 30]. The results showed narrower low retropalatal/high retroglossal regions in OSA individuals in comparison to those in healthy controls. Greater changes in the UA during respiration were observed, while they were also seen in the low retropalatal/high retroglossal regions during both sleep and wakefulness. Unlike the conventional CT imaging study [30], the subjects of this study [36] were free to breathe through their nose or mouth. Additionally, the study found that the differences in the UA area between the OSA and non-OSA groups could be completely attributed to differences in the anteriorposterior diameter as there was no significant difference in the lateral diameter between the groups. Thereafter, the finding by [37] using volumetric MRI showed that the volumes of all measured soft tissues were significantly greater in OSA subjects than those in controls, and each of these volumes was associated with an increased risk of OSA according to odds ratio products (ORP) [38]. Moreover, the study in [37] found total tongue and total lateral wall volumes were significant independent risk factors of OSA. Also, the average area and minimum area of the retropalatal region [36] were found significantly smaller in OSA subjects [37]. Contrary to previous work [36], both the lateral and anteriorposterior diameters were found smaller in OSA subjects in [37]; this could be due to the small sample size in the study reported in [36], noting that, although the difference in lateral diameter was significant, it was smaller than the difference in anteriorposterior diameter. Furthermore, 2D soft tissue measurements showed that the lateral pharyngeal wall was larger in OSA subjects and was associated with an increased risk of OSA [32]; this is congruent with the findings in [32] using 1.5-T MRI that showed larger LPW thickness in OSA group, while the associated risk of the volumetric measurements was substantially greater [32]. While ultrasound has the worst resolution compared to MRI and CT, some researchers have shown that using only ultrasound image analysis has the potential to identify OSA individuals effectively [3234, 39]. The imaging technologies (excluding ultrasounds) discussed above provide valuable insights into the anatomic and physiological changes of the UA due to OSA and provide evidence for various risk factors of OSA. However, the main drawback of these technologies is their high cost and their invasive nature (i.e., exposure to radiation or magnetic fields lead); although they are considered minimally invasive, they are not considered a necessary diagnostic tool for OSA. Moreover, medical images are passed through a series of processing 1284 Medical & Biological Engineering & Computing (2024) 62:12771311 Fig. 5Detailed process of medical image analysis Digital Image techniques before starting the analysis [39]. Figure 5 shows the process sequence for the analysis of medical images. As a result, the sample size in imaging studies has been limited. That said, the studies presented have shown that the UA of OSA patients exhibits many of the same properties during sleep and wakefulness. As such, there is significant Image Denosing and Preprocessing Region Of Interest n Region of interest measurements n research interest in potential OSA screening tools exploiting the anatomic and physiological indicators found in imaging during wakefulness and making it a feasible option. Table 2 presents a summary of all investigated papers in this section. In summary, medical imaging techniques play a critical role in the detection and analysis of OSA detection and Table 2Summary of key findings of the investigated papers for medical imaging technique Paper Technique Sample size Key result summary [40] Ultrasound 21 with AHI < 15 and 20 with AHI > 15 [28] [29] X-ray scan and flow-volume curve 10 with AHI > 10 and 5 healthy controls for X-ray and 12 with AHI > 10 and 5 healthy controls for flow-volume loop CT scan and flow-volume curve 20 with AHI > 5 and 10 healthy controls [30] CT scan and flow meter 15 with AHI < 2, 14 snorers/mild with 2 < AHI < 15 and 13 with AHI > 15 [37] Volumetric MRI 48 with AHI < 5 and 48 with AHI 15 [36] [31] MRI, electroencephalography (EEG), and nasal/oral flow Ultra-fast MRI 7 with AHI 15 and 7 BMI matched with AHI < 5 17 OSA subjects with AHI 10 and 8 control Subjects [32] Ultrasound and MRI 36 non-OSA subjects and 64 OSA subjects [33] Endoscopy 43 subjects with AHI 15 [35] Ultrasound 36 with AHI < 5, 32 with 5 AHI < 15 and 32 with AHI 15 [34] Ultrasound and MRI 36 with AHI < 5 and 64 with AHI 5 A significant correlation between the length of the tongue base and AHI even after controlling for BMI. Training sensitivity and specificity of 80% and 67% OSA is characterized by mandibular deficiency, narrow posterior airway, long soft palate, and inferiorly positioned hyoid bone OSA was characterized by narrow nasopharynx, oropharynx, and hypopharynx and identified sawtooth patterns in flow-volume curves Moderate/severe OSA is characterized by a big change in the transition between inspiration and expiration and vice versa, constant cross-sectional area during inspiration, and a high overall percentage change in UA size during respiration OSA is characterized by a smaller retropalatal region, larger tongue, soft palate, lateral walls and total soft tissues, and thicker lateral pharyngeal wall OSA is characterized by smaller UA, especially in the low-retropalatal/high-retroglossal region OSA subjects have a smaller area of the velopharynx during part of the respiratory cycle. This variation was greater in apneic patients than in controls, particularly during sleep The OSA group significantly has greater LPW thickness using MRI, while the significant difference between the two groups appears in the ultrasound only during left side MM. MRI for detecting OSA results in 90% and LPW-based obstruction in 84% The collapsibility was significant only at the hypopharyngeal level (41.8% during wakefulness and 88.3% in DISE). 18.6% and 4.6% of patients showed laryngeal obstruction during awake and DISE examinations, respectively OSA prediction using the ultrasound parameters achieved a sensitivity of 100% and a specificity of 91.7% Using tongue parameters and anthropometric values, an ultrasound sensitivity of 94% and a specificity of 98% were achieved while MRI has 56% sensitivity and 92% specificity Medical & Biological Engineering & Computing (2024) 62:12771311 provide insights into the intricate anatomical and physiological changes associated with the condition that are used to form the hypothesis for other methods. Moreover, highlights the varying capabilities of each imaging modality in delineating the characteristics of OSA. Cephalometric X-rays, for instance, are valuable in detecting specific mandibular deficiencies, yet their ability to provide conclusive evidence for OSA diagnosis appears limited. In contrast, CT scans offer detailed axial and volumetric images, revealing strong correlations between pharyngeal area and OSA, demonstrating their potential for precise anatomical analysis. MRI, with its high-resolution imaging of soft tissues, emphasizes significant differences in the velopharynx area and diameter between OSA patients and healthy individuals during the respiratory cycle. However, ultrasound, while providing less detailed images, has been utilized for measuring the length of the tongue base, which has shown associations with OSA severity. Moreover, endoscopy provides the best solution for obstructive localization. Considering the varied performances of these imaging techniques, it is imperative to adopt a multifaceted approach for a comprehensive assessment of OSA. Integrating the strengths of each modality, such as the detailed anatomical insights from CT scans and the high-resolution soft tissue visualization from MRI, could enhance the accuracy of OSA detection and characterization. Moreover, combining imaging data with other relevant risk factors, such as BMI and neck circumference, could provide a more holistic understanding of the underlying pathophysiology. Future research endeavor should be directed towards developing non-invasive or low-risk imaging alternatives that can provide precise and comprehensive data for OSA diagnosis, especially for large-scale screening. By addressing these recommendations, the field can advance towards more effective OSA management strategies, facilitating early intervention and improved patient outcomes. 2.2Negative expiratory pressure Negative expiratory pressure (NEP) is a technology designed to evaluate airflow limitations and as such its ability to detect UA collapsibility. NEP is obtained by finding the slope of the pressure-flow relationship when applying negative pressure during wakefulness. UA collapsibility is a main anatomical feature of OSA [41]. Based on that, one may predict the severity of sleep-disordered breathing (SDB); thus, different studies have been conducted to study the ability to use NEP features for detecting OSA during wakefulness. The earliest study for OSA screening using NEP was conducted in 2005 [41], where they investigated the usage of the NEP as a screening tool for SDB, including OSA (moderate and severe) and upper airway resistance syndrome (UARS). Also, a comparison study to find the differences in FVCs of 1285 normal expiration and expiration during NEP between various SDBs was done by the same study [41]. Sleep apnea can be differentiated into three types: obstructive. The subjects airflow was recorded, while breathing normally and then while breathing with NEP applied during expiration in both sitting and supine positions and with different NEP values. After that, a quantitative index (QI) factor was defined as the ratio of the area under the expiratory FVCs between NEP and normal breathing and was analyzed among the subjects. The study found a significant difference in QI (decreased with the severity of SDB) between the groups after controlling for age, BMI, and expiratory reserve volume via analysis of covariance (ANCOVA). The QI was lower in both values of NEP in most SDB patients than those in healthy controls. The outcomes of this study were congruent with those of medical imaging studies [28, 29] in that FVC features were not found to be as effective as QI in terms of detecting OSA. Moreover, FVCs were considered correlated with OSA if a sawtooth pattern was present [28, 42], though this correlation was greater with overweight subjects than with OSA subjects [28] where the adipose tissue is deposited in the pharyngeal wall [42]. The same outcome has been noted by Sanders et al. [42], where the fluttering of tissue was associated with a sawtooth pattern in the FVC. Another study was performed [43] to study the UA collapsibility of OSA patients using NEP during wakefulness [43]. The same measurement procedure presented in [41] was applied; the results confirmed the outcome of [41] that there was a correlation between the area under the flow-volume curve and UA collapsibility, especially in a seated position. A similar idea of using NEP for detecting OSA during wakefulness was studied in [44] and [45] in which airway collapsibility was assessed by analyzing expired volume during NEP application. The study in [44] measured the flow drop as a percentage of peak flow immediately after NEP and the expired volume 200 ms after NEP application (V0.2). The results showed that flow drop was significantly higher V0.2 was significantly lower in severe OSA patients, and the predictive effectiveness of flow rate was very high. Moreover, the study in [45] was conducted based on the results of [44] to assess the airway collapsibility by measuring expired volume within different times 200 ms (V0.2), 500 ms (V0.5), and 1 s (V1.0) after NEP. The outcomes were like those in [44] that V0.2 and V0.5 were significantly lower in OSA subjects and that they were both associated with OSA severity. Another study investigated whether negative pressure (NP)induced airflow alteration applied to participants upper airways during wakefulness was related to OSA severity [46]. While awake, for five full breaths, the 18 participants were orally twice exposed to 3 cm H2O of NP. Then, the NP ratio (NPR) was calculated as the ratio of the breathing volumes of the last two breaths during NP exposure to 1286 Medical & Biological Engineering & Computing (2024) 62:12771311 the last two breaths before NP exposure. The results showed a strong relationship between participants OSA severity (measured by AHI) and the exponentially transformed NPR (ExpNPR) (R2 = 0.55, p < 0.001). In the multivariable model using the ExpNPR, age, body mass index, and sex as independent variables showed that these variables accounted for 81% of the variability in AHI (p < 0.001). Finally, a leave-one-subject-out cross-validation analysis showed that the multivariable model could predict the AHI and it had a strong relation with the actual AHI from PSG (R2 = 0.72, p < 0.001). Overall, the relationship between ExpNPR and AHI remained robust, independent of demographic factors commonly associated with AHI. Table 3 presents a summary of all investigated papers in this section. Overall, the application of NEP in evaluating airflow limitations and detecting OSA during wakefulness has emerged as an encouraging technique for predicting the severity of SDB and OSA. NEP allows for the assessment of the pressure-flow relationship during expiration, providing insights into the collapsibility of the upper airway. The discussed studies indicate that NEP features, such as the QI and expired volume measurements, exhibit notable differences between individuals with SDB, including OSA, and control groups even after controlling for various confounding factors. This suggests that NEP holds potential as a reliable and effective method for detecting and assessing UA collapsibility and providing valuable insights into the pathophysiology of OSA during wakefulness. Comparatively, while medical imaging techniques offer insights into anatomical changes associated with OSA, NEP appears to focus more directly on the functional aspects of the UA collapsibility, complementing the information provided by imaging technologies. The discussed studies highlight NEPs ability to capture dynamic changes in airflow during expiration in correlation with the severity of OSA, hence, offering a functional perspective that complements the structural insights provided by imaging techniques. To leverage the full potential of NEP in OSA screening during wakefulness, future research efforts should prioritize the expansion of sample size to enhance the generalizability and robustness of findings. 2.3Facial images landmarks Recently, with the advancement of digital image processing and computer vision methodologies especially using machine learning and deep learning, researchers have applied these techniques in different medical fields [47] and in particular to OSA detection during wakefulness using facial landmarks [47, 48]. It is well known that people with craniofacial abnormalities are often diagnosed with OSA [4850], based on that OSA detection might be possible by using craniofacial characteristics using landmarks from different image postures [51]. Figure 6 shows a standard landmark system used for OSA detection using facial image analysis. The first research for using facial marks to detect OSA was proposed in [53] and then was enhanced in [54]. In [53], a logistic regression for selections and classification and regression trees were used to predict OSA among 180 subjects (114 as OSA with AHI 10). The subsequent Fig. 6Facial image landmarks, frontal view, and profile [52] (CC BY 4.0) Table 3Summary of key findings of the investigated papers for NEP Paper Sample size [41] [43] [44] [45] [46] 15 healthy controls, 6 UARS with AHI 2.5, 9 with 15 AHI < 30 and 20 with AHI > 30 20 with AHI < 1.5 and 7 with AHI 5 Key result summary OSA is characterized by the same or decreased flow during the first 75% of the expiratory period and a lower area under the curve of the NEP flow-volume loop A high correlation with the hypotonic and activated slope of pressure-flow relationship measurements. While the seated position showed the strongest correlation 24 with AHI < 5 and 24 with AHI > 30 OSA is characterized with more flow drop and less volume exhaled. Training area under curve (AUC) of receiver operating characteristics (ROC) was up to 99% with 91.7% sensitivity and 95.8% specificity 29 with AHI < 5, 28 with 5 AHI < 15, 34 with Exhaled volume percentages at 200 ms and 500 ms were significantly lower in all 15 AHI < 30 and 64 with AHI > 30 OSA severity groups. At a threshold of AHI > 15, training AUC of ROC was up to 90% with 93.9% sensitivity and 74.1% specificity 18 subjects with AHI values between 1.7 and 87.9 The relationship between ExpNPR and AHI remained robust, even when considering demographic factors commonly associated with AHI Medical & Biological Engineering & Computing (2024) 62:12771311 improvement in [54] was achieved by using the cascade method to automatically detect facial landmarks by using a support vector machine (SVM) classifier to detect the object, and then a cascade regression technique was used for landmark detection [54] in addition to the manual detection used in previous work [53]. The method was applied to 365 subjects (142 controls with AHI < 10 and 223 apneic patients with AHI > 10). The outcomes of the methods showed an accuracy of 70% using manual features (face width, eye width, cervicomental angle, and mandibular length 1) and 69% using automatic landmark detection method. In addition, the authors trained a neural network model for automatic OSA classification with automatic features as input. The neural network model showed an accuracy of 62%. Another research combined craniofacial morphology points with ML to detect OSA in the Chinese population [55]. However, ethnicity differences as an OSA risk factor may affect facial predictors. The study involved 200 subjects (146 OSA, 54 non-OSA), with calibrated frontal and profile facial photographs collected before PSG. Various facial, demographic, and anthropometrical variables were considered in predicting OSA. Logistic regression modeling showed that cervicomental angle (OR 1.06 per degree; 95% CI, 1.031.09; p < 0.001) and face width (OR 1.7 per centimeter; 95% CI, 1.12.7; p = 0.02) were OSA predictors (AUC = 0.76). Tree analysis identified the cricomental space area, mandibular width, mandibular plane angle, and neck soft tissue area as predictors with an AUC of 0.81. Moreover, recent research employed scanned 3D maxillofacial shapes on 280 Caucasian men with suspected OSA [56]. Alongside PSG, anthropometric data, comorbidities, and medication were also collected at baseline. The valid 3D craniofacial scans of 267 out of 280 subjects were processed using geometric morphometrics and passed to 13 different ML algorithms that were trained and tested. The results showed the ML algorithm achieved a specificity of 56% for detecting those with AHI > 15 (derived from PSG). 1287 When combing the 3D geometric data with patients anthropometrics, a 0.75 AUC score and sensitivity of 80% with the XGBoost classifier were achieved. More recently, using pre-trained models from companies such as Vgg from Google [57], different pre-trained facial recognition deep networks have been studied for OSA classification using transfer learning [58]. Since the pre-trained models are trained on the ImageNet dataset and weights and hyperparameters need to be updated, the authors of that study [rf] adjusted the hyperparameters of the pre-trained facial recognition deep networks VGGFace, PAMs-VGG19, and PAMs-AlexNet to classify OSA from facial depth maps (obtained from 3D facial photographs). The results scored a low performance, which was anticipated since the available datasets are very small, while deep learning requires large datasets. Table 4 presents a summary of all investigated papers in this section. In summary, the integration of digital image processing, computer vision methodologies, and machine learning techniques has ushered in a new era of possibilities in the medical field, particularly in the realm of OSA detection. Notably, the association between craniofacial abnormalities and OSA has motivated the use of facial landmarks as representations of these distinctive craniofacial features, thereby enabling OSA screening during wakefulness. Nevertheless, the accuracy of these technologies has been around 70% [56]. These findings emphasize the need for larger datasets to improve their accuracy. 2.4Acoustic pharyngometer and nasal airway pressure An acoustic pharyngometer is a non-invasive device that emits acoustic pulses through the mouth into the UA and measures their reflections to determine the minimal crosssectional area versus distance along the UA [59]. Figure 7 shows the acoustic pharyngometry schematic diagram and Table 4Summary of key findings of the investigated papers for facial image landmarks Paper Sample size [53] [54] [55] [56] [58] 66 with AHI < 10 and 114 with AHI 10 Key result summary Using clinical variables combined with the photographic measurements resulted in 79.4% validation accuracy, 85.1% specificity, and 69.7% sensitivity Dataset 1: 65 with AHI < 10 and 104 with AHI 10 AUC of ROC of 69% for either automatic or manual marking. A testing classificaDataset 2: 77 with AHI < 10 and 119 with AHI 10 tion accuracy of 69.8%, a sensitivity of 68.5%, and a specificity of 76.4%. The logistic regression classifier outperformed the neural network classifier 200 subjects (54 non-OSA with AHI < 10 and 146 Classification and regression using tree analysis identified cricomental space area, OSA with AHI 10) mandibular width, mandibular plane angle, and neck soft tissue area as predictors with an of AUC 0.81 267 subjects with AHI mean of 23.7 (0.5 to 99.5) AUC score of 0.75 AUC using the XGBoost classifier when combining 3D geometric data with patient anthropometrics 69 with an AHI threshold of 15, and Testing classification accuracy using VGG face, PAMs-VGG19, and PAMs-Alex 14 for testing were 67.42%, 57.14%, and 59.37%, respectively 1288 Fig. 7Acoustic pharyngometry: A schematic diagram and B important anatomical sites [61] (CC BY 4.0) recorded signal with related anatomical sites. Since OSA patients have often morphological changes in their UA, it is possible to use pharyngometry during wakefulness [60]. Typically, the pharyngometric measurements are done by performing different UA landmarks (usually 5), while subjects are in upright, supine, and left and right lateral positions [60]. Pharyngometry has been used to measure the minimum cross-sectional area of the UA in 60 OSA subjects in an upright position [62]. The results showed that the minimum cross-sectional area yielded a very good performance in discriminating OSA subjects. Also, by analyzing pharyngometric measurements using different landmarks, [59] the vast majority of measurements in OSA subjects were found to be smaller, while the oropharyngeal junction area measured in the supine position was the most discriminant measurement [59]. These findings agree with the imaging studies [2830]. The mathematical formulas used in the Kushida Index, as proposed in [60], are derived from a comprehensive analysis of relevant clinical data. These formulas are specifically designed to predict the likelihood of patients developing OSA and to distinguish OSA subjects from controls. The model uses readily obtainable clinical features including measurements of the oral cavity, BMI, and neck circumference. While the exact details of the formulas are beyond the scope of this review, they are based on a study [60] that demonstrated excellent performance in identifying OSA among 300 individuals (46 with AHI < 5 and 254 with AHI > 15). The high sensitivity (97.6%), specificity (100%), positive predictive value (100%), and negative predictive value (88.5%) attest to the reliability of the Kushida Index in identifying OSA risk. In addition, it was demonstrated that the Kushida Index did not correlate with any pharyngometric measurements [59]. The Kushida Index offers a non-invasive and easily implementable approach for screening OSA, particularly in settings where PSG, the gold standard for OSA diagnosis, may not be readily available [59]. Unfortunately, the specific derivation of the formulas was not elaborated. Medical & Biological Engineering & Computing (2024) 62:12771311 The indexs effectiveness in distinguishing between OSA and non-OSA individuals highlighted its potential utility in clinical practice. However, further research is warranted to validate the indexs performance in diverse populations and to elucidate the underlying mechanisms contributing to its predictive accuracy. Recently, a study has provided valuable insights into the use of acoustic pharyngometry in preventive otorhinolaryngological programs [63]. The authors compared the anthropometric and pharyngometric measurements of participants aged under and over 40 years in relation to sleep disorders. Then, try to identify the most common oral cavity alterations using acoustic pharyngometry. However, the prevalence of low soft palate and elongated uvula emerged as the predominant oral cavity anomalies pinpointed through acoustic pharyngometry. Additionally, the study revealed notable variations in both anthropometric measures (including BMI, neck circumference, and adjusted neck circumference) and pharyngometric parameters (encompassing cross-sectional area, minimal cross-sectional area, minimal distance, oral cavity length, and volume) between male and female subjects, signifying statistically significant distinctions. Further analysis uncovered compelling positive and negative correlations among these parameters, underscoring the intricate interplay between various physiological factors. Also, the results suggest that acoustic pharyngometry can be a useful tool in the screening and diagnosis of OSA, where the mean cross-sectional areas and airway volumes in any segments are statistically significantly smaller in OSA patients than in healthy controls. The nasal airway pressure measured in [64] a nasal breathing tube combined with a pressure transducer has also been used to screen OSA during wakefulness [64]; the authors applied nonlinear and nonstationary signal analysis methods on the Hilbert transform of normalized nasal airway pressure signal. The outcome of the research showed a 100% sensitivity and 100% specificity for data set 1 and 85.7% sensitivity and 100% specificity for data set 2 to screen OSA during wakefulness. Table 5 presents a summary of all investigated papers in this section. In summary, acoustic pharyngometry proves to be a rapid and efficient non-invasive method for assessing the upper airway (UA) in patients with obstructive sleep apnea (OSA). By emitting acoustic pulses and analyzing reflections, the pharyngometer quickly measures minimal cross-sectional area versus distance along the UA. Studies emphasize its discriminant capabilities, particularly in identifying OSA subjects based on the minimal cross-sectional area. Notably, the oropharyngeal junction area, especially in the supine position, emerges as an effective feature for discriminating against OSA subjects. Additionally, detecting nasal airway pressure with a breathing tube and pressure transducer holds promise for efficiently Medical & Biological Engineering & Computing (2024) 62:12771311 1289 Table 5Summary of key findings of the investigated papers for pharyngometer and nasal airway pressure Paper Sample size [62] [59] [60] [64] [63] Key result summary 30 with AHI < 15 and 51 with AHI 15 Cross-sectional area odds ratio 54.21 for multivariate analysis with 87% positive predictive value and 87% negative predictive value 16 with AHI < 5 and 54 with AHI > 5 Using a Kushida Index of 70 resulted in 89% training sensitivity with 94% specificity 46 with AHI < 5 and 254 with AHI > 15 The model resulted in 97.6% testing sensitivity with 100% specificity and 99.6% AUC of ROC 17 with AHI < 5 & 17 with AHI 5 OSA is characterized by an upward shift in frequencies. The study resulted in 85.7% testing sensitivity with 100% specificity 100 subjects Low soft palate and uvula elongation belonged to the most common oral cavity alterations identified by acoustic pharyngometry detecting OSA during wakefulness. While these methods show practicality, further investigations with larger cohorts are essential for validation and generalization, ensuring their reliability in clinical OSA assessment and management, ultimately improving patient care and treatment outcomes. 2.5Breathing sound analysis After successful OSA detection during sleep using only tracheal breathing sounds with or without bloods oxygen saturation level with very high (~ 96%) accuracy compared to PSG [10, 15, 65], the researchers started to use linear and nonlinear analyses of tracheal breathing sounds recorded in a few minutes during wakefulness to screen OSA [15, 27, 6574]. The rationale for breathing sound analysis for OSA screening during wakefulness is that structural and morphological changes in the UA, as shown by imaging studies [21], do reflect on the generated breathing sounds in the UA that are detectable by a sensitive microphone and advanced signal processing [70]. One of the first research on this topic investigated the effect of OSA on tracheal sound intensity while breathing normally in both upright and supine positions [75]. The results of sound power analysis showed a higher power of expiratory breathing sounds in the OSA group (AHI > 35) across the frequencies of 2003000 Hz compared to controls (AHI < 20). This study was one of the first to inspect features of breathing sounds in relation to OSA. However, due to its very small sample size (15 in total), no strong conclusions could be drawn. Another early study on tracheal breathing sound analysis to screen OSA used formants as the features to classify 10 mild moderate OSA subjects (AHI < 30) from 13 severe OSA subjects (AHI > 30) during wakefulness [76]. The local maxima of the spectral envelope, frequency, relative amplitude, and attenuation of each format, as well as the breath-to-breath variability of these features, were calculated and used as inputs to a linear discriminant analysis (LDA) classifier. The classifier automatically selected the frequency of F4 (say its frequency in Hz here) which was confirmed by the authors previous studies using six formants analysis [77]. The results show the breath-to-breath variability of its amplitude as its features and obtained a classification accuracy of 77.3% with a sensitivity of 76.9% and specificity of 77.8%. When BMI was added as an additional feature, accuracy increased to 81.8% with a sensitivity of 76.9% and specificity of 88.9% [77]. Another group [65] used a feedforward neural network to automatically detect breathing segments within a speech recording, and from these breathing segments extracted 104 spectral features like Mel Frequency Cepstral Coefficients (MFCC) to classify 43 OSA subjects (mean AHI 29.1) and 47 controls (mean AHI 4.7). The feature set was reduced to three spectral features using fast-forward selection, and an SVM classifier using these three features achieved 76.5% accuracy with 100% specificity, but only 55% sensitivity which is insufficient for an effective screening tool. A research group in Manitoba has performed substantial research in breathing sound analysis. In one of their first studies [71], mouth and nose breathing sounds of 35 OSA subjects (AHI > 5) and 17 controls (AHI < 5) were recorded in both the upright and supine positions. Signals were segmented into 50 ms windows, and the variance and median of the power spectrum density (PSD), Katz fractal dimension, and Kurtosis were calculated. Feature selection and reduction were performed using analysis of variance followed by maximum relevancy minimum redundancy (MRMR) algorithm to reduce to two features: (1) median of the average PSD and (2) variance of the average Kurtosis of nasal inspiration in the upright position. LDA and quadratic discriminant analysis classifiers were trained on these two features to discriminate between OSA and non-OSA (AHI threshold of 15), as well as between severe OSA and non-OSA (AHI < 5 or AHI > 30). Quadratic discriminant analysis performed the best in both cases with accuracy, sensitivity, and specificity of 83.3%, 85.0%, 81.3% 91.7%, 92.9%, and 87.5%, respectively. Furthermore, in a subsequent study [27], breathing sound features were compared and combined with anthropometric features for OSA detection. Breathing sounds were recorded from 69 OSA subjects (AHI > 10) and 61 controls (AHI < 5), from which 26 power spectrum-based features 1290 were extracted. Unpaired t-tests and SVM classifiers were used to reduce the feature set to two features which were then used to train an SVM classifier for OSA classification resulting in 83.9% testing accuracy, 82.6% sensitivity, and 85.2% specificity. Adding anthropometric data such as gender, height, weight, and Mallampati score to the classifier increased these results by only about 1%. The research was furthered by investigating the relationship between breathing sound features to various anthropometric features [69]. Breathing sounds were recorded in 48 OSA individuals (AHI > 15) and 66 controls (AHI < 15) during wakefulness in the supine position. From this dataset, 412 features were extracted from the PSD, bispectrum, Hurst exponent, and Katz and Higuchi fractal dimensions. A two-step feature reduction phase using p-value, area under curve (AUC) of ROC, SVM classification, and correlation coefficients was then performed to reduce the feature set to ten features. For analysis, the subjects were split into two subsets (training and testing) of 105 participants chosen at random (56 non-OSA and 49 OSA) based on sex, BMI, neck circumference, and age for training. Within each subset, each of the 10 features was evaluated using p-value, correlation with AHI, and the classification accuracy of SVM classification. Next, a feature selection method based on the coefficients of variation of the AUC of ROC was performed. It was observed that some features had great variability of AUC of ROC and classification accuracy between the subsets. Thus, the conclusion was made that breathing sound features are influenced by anthropometric parameters (sex, BMI, neck circumference, and age) and could, therefore, increase classification performance if anthropometric subsets were selected in the future. This conclusion was built on in a later study [73], in which two separate feature selection and classification schemes were evaluated in classifying 71 OSA subjects (AHI > 15) and 51 controls (AHI < 15) using 235 extracted breathing sound features. In the first scheme, SVM classifiers were used, and the coefficient of variation and positive impact of the classification accuracies were used to select the features least sensitive to anthropometric variables. SVM classification using various sets of one, two, and three of the least sensitive selected features resulted in maximum classification accuracy of 72.1% (sensitivity 64.7%, specificity 77.5%) using a set of only two features. In the second scheme, subjects were grouped into anthropometric subsets based on age, sex, BMI, and Mallampati score. In this second scheme, the features resulting in the highest classification accuracy in each of the subsets were selected as the most sensitive features to that anthropometric parameter. A new classification method was used in which each subject was classified using four separate SVM classifiers (one for each anthropometric subset), and the results of each classifier Medical & Biological Engineering & Computing (2024) 62:12771311 were combined using a weighted average. The outcome of the paper claimed a maximum classification accuracy of 83.6%, a sensitivity of 74.5%, and a specificity of 90.1%, by grouping participants according to each anthropometric measure into smaller groups and using a voting process the final result is generated. Due to the success of the anthropometric-based subset classification, a novel classification algorithm was introduced and tested by the same previous Manitoba research group [70]. Breathing sounds were recorded from 90 OSA subjects (AHI > 15) and 109 controls (AHI < 15) from which 250 features were extracted. Subjects were then split into subgroups as in the previous study, namely training and testing. Feature reduction and selection were done for each subgroup using p-values, SVM, and random forest (RF) classifiers to select 34 features for each subgroup. For classification, subjects were classified using four RF classifiers (one for each subgroup they fell into), each of which output a decision score: 1 multiplied by the classifier sensitivity for an OSA classification or 1 multiplied by the classifier specificity for a non-OSA classification. These four decision scores were then summed to obtain the ultimate classification which resulted in 81.4% accuracy with 82.1% sensitivity and 80.9% specificity. In addition to the SVM and random forest classifiers previously discussed, regularized logistic regression with the least absolute shrinkage and selection operator (LASSO) was also investigated for feature selection and classification [74]. The study used the same subjects as in work [70] but split the subjects into 90 moderatesevere OSA subjects (AHI > 15), 35 mild OSA subjects (5 < AHI < 15) (for testing only), and 74 controls (AHI < 5). A total of 78 PSD-based features were extracted from recorded breathing sounds, as well as 7 anthropometric features which were then reduced to five features via LASSO logistic regression. Performing LASSO linear regression classification on the entire dataset using the five selected features resulted in 81.1% accuracy, 84.4% sensitivity, and 77.0% specificity making it comparable to the previously studied methods. Later LASSO logistic regression and RF methods were compared for both feature selection and classification using the same dataset as in [74]. LASSO logistic regression where it is selected the same feature set as in the previous work [74], while the random forest algorithm selected a similar feature set, but more self-correlated [72]. Based on strictly classification results, RF (accuracy 82.1%, sensitivity 84.2%, specificity 79.5%) outperformed LASSO logistic regression (accuracy 79.3%, sensitivity 82.2%, specificity 75.8%), although logistic regression was faster [72]. Furthering research on the random forest algorithm, the group investigated predicting various additional PSG Medical & Biological Engineering & Computing (2024) 62:12771311 measurements, aside from only AHI, using breathing sound features [67]. A subset of 145 subjects from previous studies [70] who completed PSG were used in this study, from each of whom 36 PSG parameters were measured. A threshold between OSA and non-OSA was determined for each PSG parameter by comparing the average power spectra of sound signals of subjects above and below various candidate thresholds. The candidate threshold resulting in the greatest gap between the 95% confidence intervals of the average spectra was chosen as the threshold. Unique sound features were selected for each PSG parameter by first removing any features that were highly correlated with other features and then finding the features with the greatest significance between the two groups based on the threshold of the PSG parameter. Then, using a combination of anthropometric and sound features, bilinear polynomial models were developed to estimate each of the PSG parameters resulting in correlation coefficients up to 0.84. The estimated parameters were each individually used as the input to a single feature RF classifier to classify between OSA and non-OSA resulting in classification accuracies up to 88.8% showing definite potential for future use of these predicted parameters for OSA classification [67]. Table 6 presents a summary of all investigated papers in this section. By conceptualizing the UA as an acoustic 1291 medium, researchers have aimed to detect physiological abnormalities by analyzing the sound signals generated by breathing during wakefulness. Various studies have explored the potential of different sound features in accurately discerning OSA from non-OSA subjects during wakefulness, providing valuable insights into the diagnostic potential of breathing sound analysis. Initial studies using tracheal breathing sounds during wakefulness started by simply showing the differences in the sounds intensity or spectral features between healthy and OSA groups [75, 77]. Later, researchers applied advanced signal processing techniques and classifications to tracheal breathing sounds recorded for a few minutes during wakefulness to identify OSA from non-OSA groups [65, 75, 77]. Overall, the outcomes of these researches highlight the effectiveness of breathing sound analysis as a potential diagnostic tool for OSA, with the successful integration of sound-based parameters pointing towards promising technologies for quick, reliable, and accurate OSA assessment during wakefulness. However, further validation and larger-scale studies are needed to solidify the efficacy and reliability of these methods for potential integration into clinical practice, improving OSA diagnosis and management. Table 6Summary of key findings of the investigated papers for breathing sounds Paper Sample size [73] [69] [70] [27] [67] [74] [72] [71] [75] [65] [76] 51 with AHI < 15 and 71 with AHI > 15 Key result summary The least sensitive features to anthropometric factors resulted in 72.1% testing accuracy with 64.7% sensitivity and 77.5% specificity. Most sensitive features to anthropometric factors resulted in 83.6% testing accuracy with 74.5% sensitivity and 90.1% specificity 66 with AHI < 15 and 48 with AHI > 15 Different risk factors affect the breathing sounds independent of OSA severity. Some sound features showed consistent performance among different anthropometric groups 109 with AHI < 15 and 90 with AHI > 15 AWakeOSA algorithm resulted in a blind testing accuracy of 81.4%, a sensitivity of 80.9%, and a specificity of 82.1% 61 with AHI < 5 and 69 with AHI > 10 OSA is characterized by low power at low frequencies (< 350 Hz) and high power at high frequencies (> 350 Hz). Combining anthropometric and sound features resulted in 84.5% validation accuracy with 88.2% sensitivity and 80.9% specificity 80 with AHI < 15 and 65 with AHI 15 Many sound and anthropometric features had significant correlations (up to 0.6) with PSG parameters. Developed models had correlations up to 0.84. Testing accuracies to predict PSG parameters were up to 88.8% 109 with AHI < 15 and 90 with AHI > 15 Testing accuracy of 79.3% with 82.2% sensitivity and 75.8% specificity 109 with AHI < 15 and 90 with AHI > 15 RF classifier had low variance and outperformed the regularized LR in terms of blind testing accuracy, specificity, and sensitivity with 3.5%, 2.4%, and 3.7% improvement, respectively. The regularized LR was found to be faster than the RF and resulted in a more parsimonious model 17 with AHI < 5 and 35 with AHI > 5 Individuals with higher AHI had a higher average power spectrum. Using an AHI threshold of 15 resulted in a validation accuracy of 83.3% with 85% sensitivity and 81.25% specificity 8 with AHI < 20 and 7 with AHI > 35 OSA is characterized by higher inspiratory TSI difference between upright and supine positions 43 OSA subjects (mean AHI 29.1) and Using fast-forward selection and an SVM classifier using these three features obtained 76.5% 47 controls (mean AHI 4.7) accuracy, 100% specificity, and 55% sensitivity for testing 10 with AHI < 30 and 13 with AHI > 30 OSA is characterized by low format frequency in the range of 925 Hz to 1400 Hz. Using LDA resulted in 86.4% validation sensitivity with 88.9% specificity 1292 2.6Speech sound analysis Using the same rationale as those studying breathing sounds (that a structural change in the UA should be reflected in the breathing sounds), speech sounds have also been studied for OSA classification. Speaking involves the manipulation of the shape of the UA to produce different sounds, and the way that these manipulations affect the sound and relate to each other has been investigated to detect OSA. In 2009, early detection of severe apnea cases using effective automatic speech recognition-based detection by employing Gaussian mixture model (GMM)based speaker recognition technique to distinguish between severe OSA and non-OSA subjects by modeling vowels in nasal and non-nasal phonetic contexts. GMMs were trained using 12 Mel Frequency Cepstral Coefficients (MFCCs) plus energy, as well as their first and second derivatives extracted from voice recordings of four Spanish sentences [78]. Participants in this study consisted of 40 severe OSA (AHI > 30) and 40 controls (AHI < 10); all participants were males, but BMI and age of the OSA group were higher than those of the control group. The results showed that the methodology has an 81% accuracy, 77.5% sensitivity, and 85% specificity. In a follow-up study by the same research group [79], an incremental subset analysis was used to determine the most discriminative features from a set of 16 features extracted from four recorded sentences in Spanish from 62 severe OSA (AHI > 30) and 60 controls (AHI < 10). Both multiple linear regression and LDA were used for classification using the 6, 7, 8, and 9 most discriminative selected features, as well as using all 16 features. The greatest results were obtained using LDA on the eight most discriminative features; which yielded 82.9% accuracy with 85.0% sensitivity and 75.0% specificity. It is worth noting that the test set in this study matched age and BMI between OSA and non-OSA groups, although the training set did not. The authors also studied the correlation between age and BMI with the eight selected speech features and did find a significant correlation among some of the features with both age and BMI but only in the training set. This could indicate that the selected features correlated with AHI also were correlated with age and BMI. Building further on their previous research, a group investigated the use of both facial image analysis and spectral features of speech to predict subjects AHI [51]. MFCCs were extracted from recordings of 258 males with AHI between 0 and 84.4 uttering four Spanish sentences and a set of sustained vowels. The features were transformed using GMMs into a lower dimensional i-vector which was used as the input to a support vector regression (SVR) to estimate AHI. The SVR was tested using various dimensions of i-vectors both with and without clinical variables (age, height, weight, BMI, and cervical perimeter). The best correlation to true AHI using MFCCs and clinical variables was found Medical & Biological Engineering & Computing (2024) 62:12771311 to be 0.38 with an MAE of 12.43 and was obtained using a 300-dimensional i-vector with clinical variables. This result was no better than when using only clinical variables (correlation of 0.40 and MAE of 12.32) which contrasted the results in other work [78] in which MFCCs were also solely used to predict OSA. This finding thus revealed that MFCCs, alone, may not be a useful tool for OSA screening. AHI estimation using facial features and clinical variables had the best result with a correlation coefficient of 0.45 and an MAE of 11.97, and when using a threshold of AHI = 10, resulted in a classification accuracy of 79.4% with sensitivity of 85.1% and specificity of 69.7%. This research was furthered by comparing the ability to predict AHI with the ability to predict other clinical variables such as age and BMI from the same speech signals [80]. Nineteen MFCCs and their first derivatives were extracted from recordings of four Spanish sentences and a set of sustained vowels read by 426 male subjects (AHI between 0 and 102). These features were then transformed using GMMs into high-dimensional super vectors and lower-dimensional i-vectors. Next, SVR was used on the super vectors, and various dimensional i-vectors to attempt to predict AHI, as well as age, height, weight, BMI, and cervical perimeter. Overall, the results showed higher correlation coefficients and lower MAE for the estimation of all the anthropometric variables over AHI for different scenarios using super vectors or i-vectors and SVM and SVR. The best AHI prediction was found using a 100-dimensional i-vector with a linear kernel SVR which yielded a correlation of 0.3 and MAE of 13.23. Two studies investigated the correlation between formant frequencies and their bandwidths with AHI [52, 81]. In one study [81], sustained vowel recordings from 241 males (AHI, 0 to 84) with the same Spanish dialect were used, whereas in a second study [52], sustained vowel recordings from 129 females (AHI, 0 to 108) with the same Spanish dialect were used. In both cases, the Spearman correlation coefficient was used to assess the correlation between the first to third formant frequencies and their bandwidths with AHI. In the male population, only a very weak correlation was found between two of the bandwidths and AHI, but, overall, clinical variables alone provided a stronger correlation, supporting previous findings [51]. For the female population, the frequency of the second formant of /i/ vowel showed a weak correlation with AHI without showing a correlation with any of the clinical variables. Overall, no significant correlations were seen from these studies point towards an underlying correlation between clinical variables and voice features that may provide misleading results in similar studies. In an earlier study, recordings were taken from 93 subjects uttering short sentences, a set of sustained vowels, and answering yes/no questions in Hebrew [82]. The recordings Medical & Biological Engineering & Computing (2024) 62:12771311 were manually segmented into 30 ms frames to isolate the vowel sounds and the /n/ and /m/ phonemes. From these segments, 100 short-term and 28 long-term features were extracted, and sequential forward floating selection was used for feature selection among each feature set. Using the selected features with various GMMs, a decision score was calculated for both the short-term and long-term feature sets, and the two scores were then fused using multiplication and 1 nearest neighbor. This yielded a sensitivity of 79% and specificity of 83% for males (using an AHI threshold of 10) and 84% and 86%, respectively, for females (using an AHI threshold of 5). As in studies reviewed [78, 80], the age and BMI of the subjects in this study were both higher among OSA subjects than non-OSA subjects. A later study performed by the same research group of study [82] applied a similar system fusion approach but used different subsystems [83]. In this research, this research used a sustained vowel system, a continuous speech system, and a breathing sound system taken from 208 OSA subjects (AHI > 15) and 190 controls (AHI < 15). The breathing sound system specifically took into account features of the breathing sounds between speaking (i.e., breaths within the speech recordings as in previous work [65]). For the breathing sound system, the forward selection was used to find the most discriminative of a set of 104 extracted features, and then SVR was used to predict AHI. For the sustained vowel system, 12 MFCCs were extracted for each of the vowel sounds, and a convolution neural network (CNN) was used on each set of 12 MFCCs to predict AHI. For the continuous speech system, 12 MFCCs were extracted, and a long short-term memory neural network was used on all 12 MFCCs to estimate AHI. It was found that the best prediction of CNNs for the sustained vowel system results was using vowels /a/ and /u/. Moreover, the prediction of the sustained vowel system was fused with the predictions from the continuous speech and breathing sound systems, and age and BMI to produce a final prediction. Multiple fusion models were tested, and it was found that linear regression with an intersection term provided the best result. AHI was predicted with a Pearson correlation coefficient of 0.61 and an MAE of 8.80. The predicted AHI was also used to classify OSA and non-OSA (using a threshold of AHI = 15), resulting in 77.14% accuracy with 75% sensitivity and 79% specificity. In another study [84], differentiated severe OSA from healthy subjects using speech signals, one Spanish sentence and a set of sustained vowels from 121 severe OSA subjects (AHI > 30), and 127 healthy controls (AHI < 5) were recorded. Then, 253 features were extracted from the recordings and various techniques for feature reduction and classification. MannWhitney U test, principal component analysis, LDA, and genetic algorithm (GA) were each used for feature selection, and multi-layer perceptron (MLP), SVM, AdaBoost, K-nearest neighbor, and Bayesian 1293 classification were each used for classification. Based on fivefold cross-validation, the best-performing feature selection classification pairs were found to be GA with Bayesian classification, GA with SVM, and LDP with MLP. The classifications of these three systems were then combined by a majority vote to produce the final decision yielding 82.9% accuracy, 81.49% sensitivity, and 84.69% specificity in detecting severe OSA. It is worth noting a substantial age difference between groups in the study; the mean age of the severe OSA group was 54, whereas the mean age of the healthy group was 29.7 years. The authors found a Pearson correlation coefficient of 0.60 between age and AHI in the subjects of this study, which could further explain the strong classification results. More recently, a study focused on studying the use of higher frequency range (> 6 kHz) components of the speech signals and their effect on the detection of OSA during wakefulness [85]. The authors extracted traditional higher-order speech features but added higher-frequency components of the speech signals during awake for a better characterization of OSA patients speech. The features included an optimized version of traditional features for higher frequency energy with PCA-based sequence forward feature selection (PCASFFS) for feature selection. The features were extracted from 66 OSA patients. The results show that the new optimized feature for the whole frequency range achieves an accuracy of 84.85% using fivefolds for multi-class OSA detection using the QDA classifier. Another research studied the nonlinear structure of the OSA subjects speech for detection purposes applied to the Turkish population [86]. The characteristics were studied and evaluated for vowels (/a/, /i/, //, and /u/) and 24 consonants (/ca/, /ci/, /c/, /cu/, /ga/, /gi/, /g/, /gu/, etc.); then, different trials were applied to search in which voice group the nonlinear features were more discriminant in OSA. The nonlinear analysis was applied to a wide variety of voice samples having vocal tract components recorded from 40 subjects (20 OSA and 20 healthy subjects). The results showed the consonants to be more effective for classification than the vowels. Using the whole dataset and employing fivefold cross-validation, the best OSA detection performance using vowels was 83.5% using KNN, and the best performance using consonants only was 96% using SVM. Moreover, 82.5% accuracy was achieved with only six features from consonants using KNN on a blind subset of data. The study supports the hypothesis that the nonlinear characteristics of vocal tract changes in subjects with OSA. Table 7 presents a summary of all investigated papers in this section. Many studies have explored the use of speech sound analysis for classifying and predicting OSA, leveraging changes in the upper airway structure due to OSA that can affect speech sounds. Various features extracted from speech recordings demonstrate the potential of 1294 Medical & Biological Engineering & Computing (2024) 62:12771311 Table 7Summary of key findings of the investigated papers for speech analysis Paper Sample size Key result summary [81] 241 males with AHI between 0 and 84 [79] 60 with AHI < 10 and 62 with AHI > 30 [80] 426 male subjects with AHI between 0 and 102 [51] 258 males with AHI between 0 and 84.4 [78] 40 with AHI < 10 and 40 with AHI > 30 [82] Males (12 with AHI < 10 and 48 with AHI > 10) Females (14 with AHI < 5 and 19 with AHI > 5) [83] 190 with AHI < 15 and 208 with AHI > 15 [84] 127 with AHI < 5 and 121 with AHI > 30 [52] 129 females (AHI 0 to 108.4) [85] 66 Subjects (31 subjects for AHI < 5; 13 subjects for 5 AHI < 15;10 subjects for 15 AHI < 30; and 12 subjects for AHI 30) 40 subjects (20 healthy and 20 OSA with AHI > 9) Weak correlations (up to 0.21) were seen between AHI and voice features (formants) The eight most discriminative features with LDA result in 82.9% accuracy with 85.0% sensitivity and 75.0% specificity for testing. Also, it is noted that the test set in this study matched age and BMI between OSA and non-OSA groups, while the training set did not 19 MFCCs and their first derivatives were transformed using GMMs into high-dimensional super vectors and lower-dimensional i-vectors. The best AHI prediction was found using a 100-dimensional i-vector with a linear kernel SVR which yielded a maximum correlation of 0.3, and the testing classification accuracy was 71% with 92% sensitivity and 20% specificity Transformed MFCCs using GMMs into high-dimensional super vectors and lower-dimensional i-vectors used as the input to an SVR, the best correlation was found to be 0.45 with a testing classification accuracy of 79.4% with 85.1% sensitivity and 69.7% specificity 12 Mel Frequency Cepstral Coefficients (MFCCs) plus energy, as well as their first and second derivatives extracted and then passed to GMMs results in 81% accuracy, 77.5% sensitivity, and 85% specificity for validation From 30 ms segments, 100 short-term and 28 long-term features were extracted, and sequential forward floating selection was used for feature selection among each feature set. The method yielded a testing sensitivity of 79% and testing specificity of 83% for males, and 84% and 86%, respectively for females 12 MFCCs were extracted for each of the vowel sounds, and a CNN to predict AHI. Then, the predicted AHI was also used to classify OSA and non-OSA (using a threshold of AHI = 15), resulting in a correlation of 0.61 between AHI and the fused predicted AHI and 77.1% accuracy with 75% sensitivity and 79% specificity for testing 253 features from the recordings and tested various techniques for feature selection and classification. The best feature selection classification pairs were found to be GA with Bayesian classification, GA with SVM, and LDP with MLP. Then by combining these by a majority vote yielding 82.9% accuracy, 81.49% sensitivity, and 84.69% specificity in detecting severe OSA for validation Formant frequencies showed a weak correlation (up to 0.26) with AHI The new optimized feature for the whole frequency range achieves an accuracy of 84.85% for multi-class OSA detection using the QDA classifier The nonlinear characteristics of vocal tract changes in subjects with OSA can be used as discriminant features, especially for consonants. Vowels feature only were 83.5% using KNN; consonants only were 96% using SVM; and 85.5% accuracy using a blind test set with six consonant features [86] speech sound analysis to effectively discern OSA patients using machine learning techniques and feature selection methods [7981]. Overall, comprehensive speech sound analysis shows promising potential for effective OSA detection and prediction, with multimodal approaches and advanced machine learning techniques providing robust results. Further research and larger-scale studies are necessary to solidify the reliability and clinical applicability of these methods. 2.7Questionnaires Many studies have been published to evaluate the performance of OSA screening during wakefulness by Medical & Biological Engineering & Computing (2024) 62:12771311 questionnaires [87]. Here, we review the studies on the reliability of OSA screening using the following questionnaires: the Epworth Sleepiness Scale (ESS), the Berlin questionnaire, the STOP-Bang questionnaire, and the STOP questionnaire [18, 88]. A meta-analysis was applied to evaluate and compare the clinical screening tests of OSA and build a case for using them before surgery. For each screening test in this evaluation, diagnostic odds ratios were utilized as summary metrics of accuracy, and false-negative rates were used as markers of missed diagnosis [87]. The results reveal that test accuracy in many validation studies of the same screening test is inconsistent, suggesting an underlying heterogeneity in either the clinical presentation or the measured clinical components of these models. Moreover, the false-negative rates show that a large fraction of patients with OSA were missed by most clinical screening tests [87]. For screening the OSA using questionnaires, a study was conducted [89], the study assessed the questionnaires capability to detect increased apnea activity, and an epidemiologic investigation of OSA with 465 people was conducted. A questionnaire consisting of 56 questions about sleeping patterns, feeling sleepy, and performance during the day and an in-home sleep study was completed by subjects and their roommates separately. The responses were analyzed using LR, factor analysis, and ROC. The results of factor analysis showed that 16 questions, grouped into five variables (functional impact of drowsiness, self-reported breathing abnormalities, roommate-observed breathing disturbances, driving impairment, and insomnia), were found to account for 67% of the variance in the questionnaire data. Moreover, for nine out of ten questions, there was some degree of agreement between the subjects and his or her partners self-reported responses (kappa statistics, 0.34 to 0.57). Also, three questions about snoring intensity, choking a roommate saw, and dozing off while operating a vehicle were found to be the most accurate predictors of increased apnea activity, according to logistic regression analysis (ROC area, 0.78). Symptoms combined with information on gender and BMI increased prediction power by 10% (ROC area, 0.87). In population surveys of OSA, questionnaire data thus offer a reliable method of characterizing symptom distributions. Multiple questions or a separate roommate questionnaire do not greatly improve predictive ability, but gender and BMI information do. Another study to evaluate the performance of the STOPBang questionnaire as a screening tool for OSA was conducted [19]. They evaluate the studies that have been done on adults, the performance of the results was validated using PSG, and the subjects OSA was defined as AHI 5. A systematic review of 17 studies with a total number of 9206 patients was done; the results showed that the STOP-Bang questionnaire has a sensitivity value of 90% for any OSA (AHI 5), while the sensitivity increased for higher AHI 1295 values to 94% for moderate-to-severe OSA (AHI 15) and 96% severe OSA (AHI 30). Moreover, the NPV was 46%, 75%, and 90% for the same AHI values, respectively. While, in the sleep clinic population, 25% of the severe OSA have a STOP-Bang score of 3 with rising proportionally probability of 35%, 45%, 55%, and 75% with a stepwise increase of score to 4, 5, 6, and 7/8. The analysis concluded by demonstrating that the likelihood of having moderate-to-severe OSA increases with increasing the STOP-Bang score. A study to evaluate the STOP-Bang questionnaire was done [88]; the authors used a dataset of 856 subjects attending a sleep clinic for PSG which was used to evaluate the performance of the questionnaire. The authors used four of eight STOP-BANG questionnaire features, and these features were combined using a logistic regression (LR) model, where the model performance was evaluated using 8020% train and test sets. The outcome of the study in the test set was 83.3%, 45.8%, 62.2%, and 71.7% for sensitivity, specificity, positive predictive value, and negative predictive value, respectively, while the AUC has a mean value of 0.717. Moreover, it is also shown that a subset of STOPBang questionnaire features has the same performance as the full features set which means that there are redundant questions/features in the STOP-Bang questionnaire. A cross-sectional study to compare different questionnaires for screening of OSA was conducted [18], the study included 234 patients attending the sleep clinic for overnight PSG, and then the patients were administered four sleep questionnaires (Berlin, Epworth Sleepiness Scale [ESS], STOP, and STOP-Bang). The results showed that the STOP-Bang has the highest sensitivity among all OSA severity categories (97.55%, 97.74%, and 98.65% for OSA, moderate-to-severe, and severe, respectively), then the Berlin (95.07%, 95.48%, and 97.3%) and STOP questionnaires (91.67%, 94.35%, 95.48%, and 95.95%). While regarding the specificity, ESS had the highest value to predict OSA (75%), moderate-tosevere OSA (48.15%), and severe OSA (46.43%) but with the lowest sensitivity values. Finally, the Berlin, STOP, and STOP-Bang questionnaire sensitivity was quite high, but because of their low specificity, they produced more falsepositive results and failed to exclude those who were at low risk. Moreover, to determine which patients are at risk and to determine the ideal combination of these tools, the clinical utility of five different questionnairesSTOP, STOPBang, Berlin questionnaire, Epworth Sleepiness Scale, and 4-Variable Screening Toolin a sleep clinic is being evaluated [90]. Like in the previous studies, the outcome shows that the highest specificity was found in 4-V, while SB had the highest sensitivity and AUC. Their predictive value was not increased by combining various surveys. Yet, another more recent study compared the reliability of different questionnaires in the detection of OSA on 201 subjects with different OSA severity, where the subjects completed 1296 Medical & Biological Engineering & Computing (2024) 62:12771311 five different types of questionnaires: the ESS questionnaire, the STOP-Bang questionnaire, the STOP questionnaire, the BQ questionnaire and the Pittsburgh Sleep Quality Index (PSQI) [91]. Moreover, the subjects were examined using limited PSG, and the performance of the questionnaires was evaluated. The results showed the highest sensitivity was achieved by STOP-Bang (81.6%), Berlin (78.7%), and STOP questionnaires (74.2%), while the PSQI and ESS sensitives were low (50.8% and 34.5%, respectively). For specificity, the highest values were achieved by ESS (82.6%), STOPBang (75%), STOP (61.9%), and Berlin questionnaires (61.9%). Based on the results, the STOP-Bang and Berlin questionnaires were found to be the most reliable screening tool. Also, the STOP questionnaire was found to have the most time-saving nature as is a short questionnaire. Table 8 presents a summary of all investigated papers in this section. Many studies have evaluated the performance of common OSA screening questionnaires like the ESS, Berlin questionnaire, STOP-Bang, and STOP questionnaire, revealing variations in accuracy and false-negative rates. While some studies highlighted specific predictors within questionnaires, others emphasized potential redundancies. Comparative evaluations consistently showed high sensitivity but very low specificity, leading to increased false positives and the inability to exclude low-risk individuals. Recent research identified STOP-Bang and Berlin questionnaires as having the highest sensitivity, while ESS exhibited the highest specificity. Despite their sensitivity, ongoing research and refinement are essential to address specificity limitations and optimize the clinical utility of OSA screening tools. 3Challenges This section investigates the challenges in OSA detection methods used during wakefulness. 3.1Sample size The sample size is an important part of any study; for OSA detection, this is one of the major issues since it is hard to collect other data from patients especially before and after the PSG recording [92]. Even though, recording PSG is hard to record since it is costly and requires the patient to sleep at the hospital sleep lab [14]. Most of the studies are Table 8Summary of key findings of the investigated papers for the questionnaire technique Paper Sample size [87] [88] [18] [89] [19] [90] [91] 26 different studies Key result summary Testing results: for AHI of 5, sensitivity (28.6%, 83.6%) and specificity (38.2%, 68.5%); for AHI of 10, sensitivity (30.4%, 77.8%) and specificity (64.2%, 98.8%); for AHI of 25, sensitivity (85.1%, 88.2%) and specificity (75.9%, 81.2%); and for AHI of 30, sensitivity (61.4%, 100%) and specificity (36%, 75%) 155 with AHI < 5, 106 with 5 AHI < 15, 123 with 15 AHI < 30, A performance of 83.3%, 45.8%, 62.2%, 71.7%, and 0.717 for sen215 with AHI 30, and 257 with snore sitivity, specificity, positive predictive value, negative predictive value, and AUC, respectively 234 with AHI < 5, 27 with 5 AHI < 15, 177 with 15 AHI < 30, Testing results: for AHI of 5, sensitivity (91.7%, 97.6%) and speciand 148 with AHI 30 ficity (25%, 26.3%); for AHI of 15, sensitivity (75.7%, 97.7%) and specificity (3.7%, 48.2%); for AHI of 30, sensitivity (79.7%, 98.7%) and specificity (5.4%, 46.4%) 184 control subjects and 281 OSA subjects 16 questions were found to account for 67% of the variance in the questionnaire data. Moreover, for nine out of ten questions, a degree of agreement between the subject's and his or her partners self-reported responses (kappa statistics, 0.34 to 0.57). Finally, combined with gender and BMI increased prediction by 10% (ROC area, 0.87) 17 articles with a 9206 total number of subjects Using STOP-Bang threshold of 3 for the clinic and surgical populations, listed respectively in the following: For AHI 5, 90 and 84% sensitivity and 49 and 43% specificity; for AHI 15, 94 and 91% sensitivity and 34 and 32% specificity; for AHI 30 were 96 and 96% sensitivity and 25 and 29% specificity 1853 patients Testing results: for AHI of 5, sensitivity (50%, 95%) and specificity (14%, 78%); for AHI of 15, sensitivity (54%, 97.6%) and specificity (12.7%, 74.4%); for AHI of 30, sensitivity (57%, 98.7%) and specificity (9.9%, 69.3%) 24 with AHI < 5, 28 with 5 AHI < 15, 65 with 15 AHI < 30, and STOP-Bang and Berlins questionnaires demonstrated the highest 84 with AHI 30 sensitivity (81.6% and 78.7%, respectively), while ESS exhibited the highest specificity (82.6%) Medical & Biological Engineering & Computing (2024) 62:12771311 performed on a relatively small sample size, and in many cases, the recorded datasets are imbalanced which causes biasing on the results towards one of the classes. Based on that, any future studies must include larger datasets. 3.2Affordability The cost of the process to diagnose patients with OSA is high when it is referred to PSG, imaging studies, and NEP [29, 37, 43]. Another process like detection using breathing sounds and speech sounds is very cheap [73]. Moreover, researchers in their future research methods should ensure that their proposed systems are affordable. The effective parameters on the affordability consist of the technology of the process, fabrication materials, required equipment, and the transmission technology if required. 3.3Ease of use Usually, OSA diagnosis devices are not easy to use and require a complex setup of the process and a specialized person for the interpretation of the acquired data [43]. This is a major challenge to the current methodologies; based on that, researchers must take into consideration the simplicity of the process setup and in it is best cases the patient can set it up on their side with easy instructions. 3.4Portability The portability of any proposed system for wakefulness detection is OSA which is a main challenge, since most of the reviewed systems and methods required high computational processes when using data processing [93]. Moreover, some techniques like medical imaging require large equipment for data acquisition [28]. Based on that, researchers in the future must be able to use cloud computing and wireless transmission of data to overcome such challenges. 1297 3.6Detection of different OSA severity Most of the proposed methods are focused on detecting if there is OSA or not based on a threshold applied to the recorded AHI [94]. One of the challenging things in the future development of OSA detection during wakefulness is to provide a system that can detect the severity of the OSA based on the predefined AHI value thresholds [45]. Such a system can help provide a full diagnosis system instead of a classification system. 3.7Performance Developing an advanced and high-performance method that is able to detect OSA with very high performance is the ultimate goal of any detection and classification system [42]. Additionally, based on the previously discussed challenges, this can be quite challenging based on the number of subjects included in the studies [82]. So, the performance of the detection system also remains an important challenge by focusing on the specificity and sensitivity not only the accuracy and needs to be investigated further in-depth in any future work. 3.8Providing physiological interpretation and information beyond AHI The primary outcome of the PSG system is the apnea/hypopnea index (AHI) which is the primary to detect OSA severity. However, other severity parameters like total arousal index and SpO2 are very important to provide a full diagnosis of the patient and decide on a treatment option [67]. PSG assessments and home sleep tests measure these parameters, but most wakefulness techniques are unable to estimate or predict these parameters; there has been only one study [67]. In future methods, there is a need to provide a system able to estimate or predict these parameters and to be investigated further in-depth in any future work. 3.5Measurement time 4Managing false positives in OSA detection Measuring time is one of the issues during studies and affects the number of the included subjects in the studies [40]. Moreover, it also affects the ability to apply any system in real time [53]. Also, since our main focus in this review is on the detection of OSA during wakefulness, it is important and challenging to make proposed methods to record required measurements in a short time and even generate the results in a short time [93]. Any future research should focus on the required time for measuring as one of the main challenges to be overcome. False positives can emerge in various OSA detection techniques; for each of these techniques, there are different reasons behind false positives, and different guidance on managing an excessive number of clinically irrelevant OSA detections is needed. These insights are essential for researchers, clinicians, and technologists striving to enhance the accuracy and reliability of OSA diagnosis during wakefulness [95]. By addressing the issue of false positives systematically across different OSA detection techniques, we aim to contribute to the development of more precise and 1298 clinically relevant methods [96]. The goal is to ensure that patients receive accurate diagnoses, appropriate treatment plans, and peace of mind, while healthcare resources are utilized efficiently and effectively. In the following, we will delve into specific techniques and their respective strategies for managing false positives [97]. For the use of imaging techniques, mitigating clinically irrelevant OSA detections involves implementing robust post-processing methods and automatically identifying and excluding artifacts [96]. It is crucial to set specific parameters during image acquisition and establish criteria for extracting anatomical features based on validated clinical data to distinguish between relevant and irrelevant findings. Regular calibration of imaging equipment, adherence to standardized protocols, and employing standard device setups are essential to minimize false positives [95, 97]. For various OSA detection methods, managing excessive clinically irrelevant detections necessitates specific strategies. In NEP tests, clear clinical guidelines defining thresholds for collapsibility and guiding repeat tests, or different interpretations are crucial [95]. Training healthcare professionals in NEP interpretation nuances can further reduce the likelihood of excessive irrelevant detections [97]. In facial landmarks analysis, refining algorithms and incorporating machine learning models based on large datasets enhance landmark detection accuracy [98]. Similar precision improvements can be achieved in pharyngometry by establishing normative data for airway dimensions, considering dynamic changes during sleep, and comparing patient data to norms [95, 97]. Advanced signal processing in breathing sound analysis, including using balanced groups dataset, appropriate recording protocols, and patient-specific characteristics, enhances accuracy [96, 99, 100]. Similarly, in speech signal analysis, focusing on specific speech features, considering contextual information, and employing continuous monitoring and real-time feedback systems contribute to accuracy [98, 99]. In questionnaires, refining designs, implementing scoring thresholds, and combining questionnaire data with physiological parameters improve diagnostic accuracy and reduce irrelevant detections [96, 99, 100]. Overall, integrating these tailored strategies into each OSA detection technique enhances precision, reliability, and clinical relevance [96]. 5Discussion Gold standard OSA diagnosis, an overnight PSG sleep study, has many drawbacks such as being labor-intensive, time-consuming, expensive, and lack of availability in remote areas. Thus, research interest in detecting OSA during wakefulness within a few minutes has been on the rise, especially in the last decade. This review has been dedicated Medical & Biological Engineering & Computing (2024) 62:12771311 to reviewing the studies dedicated to understanding OSA manifestation on the upper airway as well as technologies to screen and detect OSA during wakefulness. This review has presented 57 journal papers and conference papers; all papers related to screening children were excluded since children, and adults have significant disparities in sleep and respiratory physiology and their OSA pathology [101]. Having analyzed and condensed available literature, characteristics of a good OSA screening tool have been identified as (1) affordability, (2) ease of use, (3) portability, (4) executability during wakefulness, (5) prompt setup and measurement time, (6) large sample size testing, (7) noninvasiveness, (8) ability to screen for different OSA severity groups, (9) accuracy with high sensitivity and specificity, and (10) ability to provide physiological interpretation and information beyond AHI. Most of these characteristics are a challenge that faces the past and current development of OSA wakefulness technologies. Given these characteristics, imaging techniques would not meet the design specifications for a future OSA screening tool, as imaging methods remain bulky, expensive, and not readily available outside of a clinical setting. However, imaging techniques remain very helpful research tools to better understand the pathogenes of the disorder. Table 9 summarizes the investigated papers method characteristics. Of the 57 reviewed papers, 40 papers proposed a classification analysis methodology, while only 12 papers of these 40 introduced only training results, 11 papers introduced validation results, and 25 papers introduced testing results. Moreover, the number of participants per study was between 14 [36] and 597 [80] individuals, and this number is still small given the heterogeneity of the OSA population and its confounding variables and also compared to the number of samples that application of artificial intelligence and deep learning required to achieve reliable results. A major drawback with imaging techniques [2830, 36, 37, 58], negative expiratory pressure [41, 4345], and pyranometer-based studies [59, 60, 62] is that they did not introduce any testing classification results; these studies require further investigation with validation and blind testing results. On the other hand, facial-related papers provided testing classification accuracies, but they were relatively low: they were between 57.14% [58] using deep learning and 69.8% [54] using automatics landmark detection with NN classification. These results show that facial imaging still needs more development and may require combining extracted features from these techniques with other features such as anthropometric features to enhance the overall performance. In contrast to the above studies, the OSA detection performance was increased in breathing soundrelated papers, with testing classification accuracies between 72.1 and 83.6% [73], sensitivities between 55 [65] and 82.2% [72], and specificities between 75.8 [72] and 100% [65]. While tracheal breathing sound analysis has shown reasonably high Y Y 2004 Y 2004 Y Tamisier R, Wuyam B, Nicolle I, Ppin JL, Orliaguet O, Perrin CP N 2003 N [41] [59] [37] [75] [60] Y Y N 1993 N [30] 1996 Y 1997 Y N 1983 N [28] Y Y Y Y Y Y Y Y N 1983 N [29] Y Y 1981 Y Sanders MH, Martin RJ, Pennock BE, Rogers RM Haponik EF, Smith PL, Bohlman ME, Allen RP, Goldman SM, et al Riley R, Guilleminault C, Herran J, Powell N Schwab RJ, Gefter WB, Hoffman E a, Gupta KB, Pack AI Pasterkamp H Kushida CA, Efron B, Guilleminault C Schwab RJ, Pasirstein M, Pierson R, Mackley A, Hachadoorian R, et al Dae GJ, Hae YC, Grunstein RR, Yee B [42] Y Y N Y Y N Y N Y Wake Prompt Used during 15 min wakefulness measurement time Year Afford Port US$5000 Easily transported to home or clinic Reference Author Table 9Summary of methodology in reviewed works based on method characteristics N N N N Y N N N N Sample N 100 and N of one severity group is 30 N N N N Y N N N N Tested Evaluated a blind testing dataset Y Y Y Y Y Y Y Y Y Non-inv Noninvasive technology N N N N N N N N N Groups Can screen for 2 OSA severity groups N N N N Y N N N Y N N N N Y N N N Y N Y Y N N Y Y Y N Spec Sens More 80% 80% Additional metrics can be measured Medical & Biological Engineering & Computing (2024) 62:12771311 1299 Y Y Y Y Y Y Y Y 2006 Y 2009 Y 2009 Y 2009 Y 2009 Y 2010 Y 2011 Y 2011 Y Salisbury JI, Sun Y Lahav Y, Rosenzweig E, Heyman Z, Doljansky J, Green A, et al Lee RWW, Petocz P, Prvan T, Chan ASL, Grunstein RR, Cistulli PA Pozo RF, Murillo JLB, Gmez LH, Gonzalo EL, Ramrez JA, et al Ramachandran SK, Josephs LA Caseiro P, Fonseca-Pinto R, Andrade A Romano S, Salvaggio A, Hirata RP, Bue A Lo, Picciolo S, et al Romano S, Salvaggio A, Bue A Lo, Marrone O, Insalaco G [64] [44] [45] [93] [87] [78] [53] [40] Year Afford Port US$5000 Easily transported to home or clinic Reference Author Table 9(continued) Y Y Y Y Y Y Y Y N N Y Y Y Y Y Y Wake Prompt Used during 15 min wakefulness measurement time N N N - N Y N N Sample N 100 and N of one severity group is 30 N N N Y N N N Y Tested Evaluated a blind testing dataset Y Y Y Y Y Y Y Y Non-inv Noninvasive technology Y N N Y N N N N Groups Can screen for 2 OSA severity groups N N N N Y Y N Y N N N Y N N N Y N N N N N N Y N Spec Sens More 80% 80% Additional metrics can be measured 1300 Medical & Biological Engineering & Computing (2024) 62:12771311 [79] Y Y Y Y 2014 Y 2014 Y 2015 Y 2014 Y Carrera HL, Marcus CL, McDonough JM, Morera JCO, Huang J, Farre R Montero Benavides A, Fernndez Pozo R, Toledano DT, et al Y Y 2012 Y 2013 Y [43] [84] [76] [18] [62] Y Y Y Y Y Y Y Y 2012 Y [71] Y Y 2011 Y Goldshtein E, Tarasiuk A, Zigel Y Montazeri A, Giannouli E, Moussavi Z El-Sayed IH DeYoung PN, Bakker JP, Sands SA, BatoolAnwar S, Connolly JG, Butler JP Sol-Soler J, Fiz JA, Torres A, Jan R Sol-Casals J, Munteanu C, Martn OC, Barb F, Queipo C, Amilibia J [82] Y N Y Y Y Y Y Y Wake Prompt Used during 15 min wakefulness measurement time Year Afford Port US$5000 Easily transported to home or clinic Reference Author Table 9(continued) Y N Y N N N N Sample N 100 and N of one severity group is 30 Y N N N Y N N Y Tested Evaluated a blind testing dataset Y Y Y Y Y N Y Y Non-inv Noninvasive technology N N N N Y N N N Groups Can screen for 2 OSA severity groups N N Y Y N Y Y Y Y N Y Y Y Y Y Y N N N N N Y N N Spec Sens More 80% 80% Additional metrics can be measured Medical & Biological Engineering & Computing (2024) 62:12771311 1301 Y Y Y 2016 Y 2017 Y [88] [81] Y Y Y Y Y 2016 Y EspinozaCuadros F, FernndezPozo R, Toledano DT, et al Montero Benavides A, Blanco Murillo JL, Fernndez Pozo R, et al Behar JA, Palmius N, Daly J, Li Q, Rizzatti FG, et al Y Y [80] [19] [51] Y Y Y Y Y Y Y Wake Prompt Used during 15 min wakefulness measurement time Y 2014 Y Pataka A, Daskalopoulou E, Kalamaras G, Fekete Passa K, et al 2015 Y EspinozaCuadros F, FernndezPozo R, Toledano DT, AlczarRamrez JD, LpezGonzalo E, HernndezGmez LA 2015 Y Nagappa M, Liao P, Wong J, Auckley D, Ramachandran SK, et al Year Afford Port US$5000 Easily transported to home or clinic [90] Reference Author Table 9(continued) - Y Y - Y - Sample N 100 and N of one severity group is 30 Y N Y Y Y Y Tested Evaluated a blind testing dataset Y Y Y Y Y Y Non-inv Noninvasive technology Y N N Y N Y Groups Can screen for 2 OSA severity groups N N N N N N Y N Y Y Y Y N N N N N N Spec Sens More 80% 80% Additional metrics can be measured 1302 Medical & Biological Engineering & Computing (2024) 62:12771311 Darquenne C, Elliott AR, Sibille B, Smales ET, DeYoung PN, et al Islam SM, Mahmood H, Al-Jumaily AA, Claxton S Simply RM, Dafna E, Zigel Y Elwali A, Moussavi Z [69] [65] [58] Y N Y Y 1994 Y 2018 N 2018 Y 2018 Y Y Y 2017 Y 2019 Y Y 2016 Y Y Y Y Y Y Y Y [36] [89] [52] [27] Y Y Y N Y Y Y Y Y 2017 Y Balaei AT, Sutherland K, Cistulli PA, De Chazal P Elwali A, Moussavi Z Tyan M, EspinozaCuadros F, Pozo RF, Toledano D, Gonzalo EL, Ramirez JDA Kump K, Whalen C, Tishler P V., Browner I, Ferrette V, et al [54] Y Wake Prompt Used during 15 min wakefulness measurement time Year Afford Port US$5000 Easily transported to home or clinic Reference Author Table 9(continued) Y N N N - Y Y Y Sample N 100 and N of one severity group is 30 N Y Y N Y N N Y Tested Evaluated a blind testing dataset Y Y Y Y Y Y Y Y Non-inv Noninvasive technology N N N N Y N N N Groups Can screen for 2 OSA severity groups N Y N N N N Y N N N N N Y N Y N N N N Y N N Y N Spec Sens More 80% 80% Additional metrics can be measured Medical & Biological Engineering & Computing (2024) 62:12771311 1303 Y Y Y N N N 2023 N Molnr V, Molnr A, Lakner Z, Trnoki DL, Trnoki D, Jokkel Z, et al Bindi I, Ori M, 2022 N Marchegiani M, Morreale M, Gallucci L, Ricci G 2022 N Molnr V, Lakner Z, Molnr A, Trnoki DL, Trnoki D, Kunos L, et al [35] [33] Y 2020 Y [32] Y 2020 Y Y Y 2017 Y 2021 Y [67] [72] [83] [70] Y Y Y Y Y Y 2019 Y [73] Y Y 2019 Y Hajipour F, Jozani MJ, Elwali A, Moussavi Z Elwali A, Meza-Vargas S, Moussavi Z Elwali A, Moussavi Z Simply RM, Dafna E, Zigel Y Hajipour F, Jozani MJ, Moussavi Z Elwali A, Moussavi Z [74] N N N Y Y Y Y Y Y Wake Prompt Used during 15 min wakefulness measurement time Year Afford Port US$5000 Easily transported to home or clinic Reference Author Table 9(continued) N N N Y Y Y Y Y Y Sample N 100 and N of one severity group is 30 N N N Y Y Y Y Y Y Tested Evaluated a blind testing dataset N Y Y Y Y Y Y Y Y Non-inv Noninvasive technology Y N N N N N N N N Groups Can screen for 2 OSA severity groups Y N Y Y N N Y N N Y N Y Y Y N Y Y Y Y N Y Y N N Y N Y Spec Sens More 80% 80% Additional metrics can be measured 1304 Medical & Biological Engineering & Computing (2024) 62:12771311 [86] [63] [85] Y Y Y Y Y 2023 Y Y Y Y Y Y Y Y Y Y Y Y Y Y N N Wake Prompt Used during 15 min wakefulness measurement time N 2022 Y 2022 Y 2022 Y Monna F, Ben Messaoud R, Navarro N, Baillieul S, Sanchez L, Loiodice C, et al Shivarov G Pang K-G, Hsung T-C, Liao G, Ling W-K, Law AK-W, Choi WS Ylmaz D, Yldz M, Uyar Toprak Y, Yetkin S [56] [55] [46] 2022 N Molnr V, Lakner Z, Molnr A, Trnoki DL, Trnoki D, Kunos L, et al Lim J, Alshaer 2023 Y H, Ghahjaverestan NM, Bradley TD Sutherland K, 2016 Y Lee RWW, Petocz P, Chan TO, Ng S, Hui DS, et al Year Afford Port US$5000 Easily transported to home or clinic [34] Reference Author Table 9(continued) Y N Y Y Y N N Sample N 100 and N of one severity group is 30 Y N - Y Y N N Tested Evaluated a blind testing dataset Y Y Y Y Y Y Y Non-inv Noninvasive technology N N - N N Y N Groups Can screen for 2 OSA severity groups Y N - N N N Y Y N - N N N Y Y N - N N N Y Spec Sens More 80% 80% Additional metrics can be measured Medical & Biological Engineering & Computing (2024) 62:12771311 1305 Y Y Y Sample N 100 and N of one severity group is 30 Wake Prompt Used during 15 min wakefulness measurement time Y Y Y Y Y Y Medical & Biological Engineering & Computing (2024) 62:12771311 blind test sensitivity and specificity, studies have shown the accuracy can still be benefited by combining some anthropometric features with the sound analysis [70]. Similar to breathing sounds, speech sound analysis can also be used for OSA detection. More variation was noticed in speech signalrelated papers, with testing classification accuracies between 71 [80] and 79.4% [78], sensitivities between 75 [79] and 92.9% [71], and specificities between 20 [80] and 79% [83]. On the other hand, the greatest variation was seen in papers related to questionnaires, showing testing sensitivities between 30.4 [19] and 99.8% [18] and specificities between 3.7 [18] and 98.8%. The oral cavity and clinical measurements related article provided 97.6% testing sensitivity and 100% specificity [60]. However, the oral cavity and clinical measurement model has certain limitations that affect its accuracy and further development. These limitations arise from extreme values in the models variables and include factors such as age restrictions (persons younger than 15 years or older than 80 years), conditions like Marfan syndrome or major muscle disorders, oral abnormalities (cleft palate, severe malocclusion, or reconstructive surgery), coexisting serious medical conditions, and limited ethnic diversity in the patient sample. Air pressurerelated papers provided 85.7% testing sensitivity and 100% specificity [64]. However, the air pressure research paper was done on very small datasets; thus, further investigation and standardizing the instrumentation are required to confirm the robustness of the proposed methodology. Furthermore, there is interest in predicting other OSA-related parameters that a PSG overnight measures, by breathing sound analysis during wakefulness [67]. Overall, combining different methodologies for wider reporting metrics, in addition to improved accuracy, may provide a more well-rounded, comprehensive screening tool for future use. 2022 Y [91] Solecka , Matler K, Kostliv T, Kubec V, Tomkov H, Betka J Y Year Afford Port US$5000 Easily transported to home or clinic 6Conclusions Reference Author Table 9(continued) Tested Evaluated a blind testing dataset Non-inv Noninvasive technology Groups Can screen for 2 OSA severity groups Spec Sens More 80% 80% Additional metrics can be measured 1306 Non-invasive detection during wakefulness of OSA is important as it can resolve many current major issues such as long waiting time to have an overnight PSG and lack of OSA diagnosis by reducing the need for PSG assessment through a quick and accurate screening during wakefulness, thus, significantly reducing the economic burden of OSA on healthcare. In addition, a reliable, comprehensive OSA detection tool would reduce possible perioperative morbidity and mortality, as well as facilitate faster treatment. There exist many studies that have investigated OSA screening during wakefulness, and yet, as suggested throughout the present review, opportunities for improvement exist to provide a measure for severity rather than only screening for OSA and non-OSA populations. Medical & Biological Engineering & Computing (2024) 62:12771311 In this paper, different techniques for OSA detection during wakefulness are divided based on the main used methodology like imaging techniques, negative expiratory pressure, facial image landmarks, pharyngometry, breathing sound analysis, speech signal analysis, and questionnaires. For each technique, all related papers are reviewed and summarized to show the main outcome. This review also highlights the road map for the design specifications which are required or preferred in any feature methodology for the wakefulness technique of OSA detection. The future open path for research in this area will be the design of more comfortable, reliable, and accurate devices to provide comfortable, cost-effective, and accurate ways for wakefulness detection of OSA and its severity; these will reduce the need for PSG recordings, especially for the initial screening. In a nutshell, this review shows that there is an increased focus by researchers on developing techniques for OSA detection during wakefulness. Although there are promising results from surveyed papers, there is a need for more clinical validation of these methods on larger populations. Author contribution Conceptualization: AMA, AE, and ZM; methodology: AMA, AE, BK, and FH; writing and original draft preparation: AMA and AE; writing, review, and editing: AMA, AE FH, NJ and ZM; supervision: ZM; project administration: ZM; funding acquisition: ZM through the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Canadian Institutes of Health Research (CIHR) funding. All authors contributed to the manuscript revision, read, and approved the submitted version. Funding This review study was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Canadian Institutes of Health Research (CIHR). Declarations Conflict of interest The authors declare no competing interests. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the articles Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the articles Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. 1307 References 1. Colten HR, Altevogt BM (2006) Sleep disorders and sleep deprivation: an unmet public health problem. The national academies press, Washington, D.C. https://doi.org/10.17226/11617 2. Kushida CA, Littner MR, Morgenthaler T et al (2005) Practice parameters for the indications for polysomnography and related procedures: an update for 2005. Sleep 28:499523. https://doi. org/10.1093/SLEEP/28.4.499 3. American Academy of Sleep Medicine (2005) International classification of sleep disorders: diagnostic & coding manual, 2nd ed. American Academy of Sleep Medicine, Westchester, IL 4. Berry RB, Brooks R, Gamaldo CE et al (2012) The AASM manual for the scoring of sleep and associated events. Rules Terminology Techn Specifications Darien Illinois Am Acad Sleep Med 176:2012 5. Bradley TD, Floras JS (2013) Sleep apnea: implications in cardiovascular and cerebrovascular disease, 2nd edn. CRC Press, Boca Raton 6. Butt M, Dwivedi G, Khair O, Lip GYH (2010) Obstructive sleep apnea and cardiovascular disease. Int J Cardiol 139:716. https:// doi.org/10.1016/J.IJCARD.2009.05.021 7. Espiritu JRD (2021) Health consequences of obstructive sleep apnea. In: Management of obstructive sleep apnea. Springer International Publishing, Cham, pp 2343. https://doi.org/10. 1007/978-3-030-54146-0_3 8. Yoshihisa A, Takeishi Y (2019) Sleep disordered breathing and cardiovascular diseases. J Atheroscler Thromb 26:315327. https://doi.org/10.5551/JAT.RV17032 9. Noda A, Yasuma F, Miyata S et al (2019) Sleep fragmentation and risk of automobile accidents in patients with obstructive sleep apneasleep fragmentation and automobile accidents in OSA. Health N Hav 11:171181 10. Young T, Finn L, Peppard PE et al (2008) Sleep disordered breathing and mortality: eighteen-year follow-up of the Wisconsin Sleep Cohort. Sleep 31:1071. https://doi.org/10.1016/ s8756-3452(08)79181-3 11. American Academy of Sleep Medicine (2016) Hidden health crisis costing america billions: underdiagnosing and undertreating obstructive sleep apnea draining healthcare system. Frost & Sullivan, Darien, IL 12. The Harvard Medical School Division of Sleep Medicine (2010) The price of fatigue: the surprising economic costs of unmanaged sleep apnea. Harvard Medical School Division of Sleep Medicine Boston, Boston, MA 13. Tregear S, Reston J, Schoelles K, Phillips B (2009) Obstructive sleep apnea and risk of motor vehicle crash: systematic review and meta-analysis. J Clin Sleep Med 5:573. https://doi.org/10. 5664/jcsm.27662 14. Stewart SA, Skomro R, Reid J et al (2015) Improvement in obstructive sleep apnea diagnosis and management wait times: a retrospective analysis of a home management pathway for obstructive sleep apnea. Can Respir J 22:167170. https://doi. org/10.1155/2015/516580 15. Yadollahi A, Moussavi Z (2009) Acoustic obstructive sleep apnea detection. In: 2009 Annual international conference of the ieee engineering in medicine and biology society. IEEE, pp 7110 7113. https://doi.org/10.1109/IEMBS.2009.5332870 16. Chen L, Pivetta B, Nagappa M et al (2021) Validation of the STOP-Bang questionnaire for screening of obstructive sleep apnea in the general population and commercial drivers: a systematic review and meta-analysis. Sleep Breath 25:17411751. https://doi.org/10.1007/S11325-021-02299-Y/FIGURES/2 17. Mazzotti DR, Keenan BT, Thorarinsdottir EH et al (2022) Is the Epworth Sleepiness Scale sufficient to identify the excessively 1308 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. Medical & Biological Engineering & Computing (2024) 62:12771311 sleepy subtype of OSA? Chest 161:557561. https://doi.org/10. 1016/J.CHEST.2021.10.027 El-Sayed IH (2012) Comparison of four sleep questionnaires for screening obstructive sleep apnea. Egypt J Chest Dis Tuberc 61:433441. https://doi.org/10.1016/J.EJCDT.2012.07.003 Nagappa M, Liao P, Wong J et al (2015) Validation of the STOP-Bang questionnaire as a screening tool for obstructive sleep apnea among different populations: a systematic review and meta-analysis. PLoS One 10:e0143697. https://doi.org/10. 1371/JOURNAL.PONE.0143697 American Society of Anesthesiologists (2014) Practice guidelines for the perioperative management of patients with obstructive sleep apnea: An updated report. Anesthesiology 120:268 286. https://doi.org/10.1097/ALN.0000000000000053 Finkelstein Y, Wolf L, Nachmani A et al (2014) Velopharyngeal anatomy in patients with obstructive sleep apnea versus normal subjects. J Oral Maxillofac Surg 72:13501372. https://doi.org/ 10.1016/J.JOMS.2013.12.006 Lan Z, Itoi A, Takashima M et al (2006) Difference of pharyngeal morphology and mechanical property between OSAHS patients and normal subjects. Auris Nasus Larynx 33:433439. https:// doi.org/10.1016/J.ANL.2006.03.009 Younes M (2008) Role of respiratory control mechanisms in the pathogenesis of obstructive sleep disorders. J Appl Physiol 105:13891405. https://doi.org/10.1152/JAPPLPHYSIOL. 90408.2008 Malhotra A, Pillar G, Fogel R et al (2001) Upper-airway collapsibility: measurements and sleep effects. Chest 120:156161. https://doi.org/10.1378/CHEST.120.1.156 Ryan CM (1985) Bradley TD (2005) Pathogenesis of obstructive sleep apnea. J Appl Physiol 99:24402450. https://doi.org/10. 1152/JAPPLPHYSIOL.00772.2005 Betts JG, DeSaix P, Johnson E et al (2022) Anatomy and physiology, 2nd ed. OpenStax, Houston, TX Elwali A, Moussavi Z (2016) Obstructive sleep apnea screening and airway structure characterization during wakefulness using tracheal breathing sounds. Ann Biomed Eng. https://doi. org/10.1007/s10439-016-1720-5 Riley R, Guilleminault C, Herran J, Powell N (1983) Cephalometric analyses and flow-volume loops in obstructive sleep apnea patients. Sleep 6:303311. https://d oi.o rg/1 0.1 093/ sleep/6.4.303 Haponik EF, Smith PL, Bohlman ME et al (1983) Computerized tomography in obstructive sleep apnea. Correlation of airway size with physiology during sleep and wakefulness. Am Rev Respir Dis 127:221226. https://doi.org/10.1164/arrd.1983. 127.2.221 Schwab RJ, Gefter WB, Hoffman EA et al (1993) Dynamic upper airway imaging during awake respiration in normal subjects and patients with sleep disordered breathing. Am Rev Respir Dis 148:13851400. https://doi.org/10.1164/ajrccm/148.5.1385 Ciscar MA, Juan G, Mart nez V et al (2001) Magnetic resonance imaging of the pharynx in OSA patients and healthy subjects. Eur Respir J 17:7986. https://doi.org/10.1183/09031936. 01.17100790 Molnr V, Molnr A, Lakner Z et al (2023) The prognostic role of ultrasound and magnetic resonance imaging in obstructive sleep apnoea based on lateral oropharyngeal wall obstruction. Sleep Breath 27:319328. https://d oi.o rg/1 0.1 007/s 11325-0 22-0 2597-z Bindi I, Ori M, Marchegiani M et al (2022) Diagnosis of upper airways collapse in moderate-to-severe OSAHS patients: a comparison between drug-induced sleep endoscopy and the awake examination. Eur Arch Otorhinolaryngol 279:21672173. https:// doi.org/10.1007/s00405-021-07184-8 34. Molnr V, Lakner Z, Molnr A et al (2022) Ultrasound and magnetic resonance imaging of the tongue in obstructive sleep apnoea. Appl Sci 12:9583. https://d oi.org/10.3390/a pp121 99583 35. Molnr V, Lakner Z, Molnr A et al (2022) The predictive role of subcutaneous adipose tissue in the pathogenesis of obstructive sleep apnoea. Life 12:1504. https://doi.org/10.3390/life121015 04 36. Darquenne C, Elliott AR, Sibille B et al (2018) Upper airway dynamic imaging during tidal breathing in awake and asleep subjects with obstructive sleep apnea and healthy controls. Physiol Rep 6:19. https://doi.org/10.14814/phy2.13711 37. Schwab RJ, Pasirstein M, Pierson R et al (2003) Identification of upper airway anatomic risk factors for obstructive sleep apnea with volumetric magnetic resonance imaging. Am J Respir Crit Care Med 168:522530. https://doi.org/10.1164/RCCM. 200208-866OC 38. Bland JM, Altman DG (2000) The odds ratio. BMJ 320:1468. https://doi.org/10.1136/BMJ.320.7247.1468 39. Kim TY, Son J, Kim KG (2011) The recent progress in quantitative medical image analysis for computer aided diagnosis systems. Healthc Inform Res 17:143. https://doi.org/10.4258/HIR. 2011.17.3.143 40. Lahav Y, Rosenzweig E, Heyman Z et al (2009) Tongue base ultrasound: a diagnostic tool for predicting obstructive sleep apnea. Ann Otol Rhinol Laryngol 118:179184. https://doi.org/ 10.1177/000348940911800304 41. Tamisier R, Wuyam B, Nicolle I et al (2005) Awake flow limitation with negative expiratory pressure in sleep disordered breathing. Sleep Med 6:205213. https://doi.org/10.1016/j.sleep.2004.10.013 42. Sanders MH, Martin RJ, Pennock BE, Rogers RM (1981) The detection of sleep apnea in the awake patient: the saw-tooth sign. JAMA 245:24142418. https://doi.org/10.1001/JAMA. 1981.03310480030021 43. Carrera HL, Marcus CL, McDonough JM et al (2015) Negative expiratory pressure technique: an awake test to measure upper airway collapsibility in adolescents. Sleep 38:1783. https://doi. org/10.5665/SLEEP.5164 44. Romano S, Salvaggio A, Hirata RP et al (2011) Upper airway collapsibility evaluated by a negative expiratory pressure test in severe obstructive sleep apnea. Clinics 66:567572. https://doi. org/10.1590/S1807-59322011000400008 45. Romano S, Salvaggio A, Lo BA et al (2011) A negative expiratory pressure test during wakefulness for evaluating the risk of obstructive sleep apnea in patients referred for sleep studies. Clinics (Sao Paulo) 66:18871894. https://doi.org/10.1590/ S1807-59322011001100007 46. Lim J, Alshaer H, Ghahjaverestan NM, Bradley TD (2023) Relationship between airflow limitation in response to upper airway negative pressure during wakefulness and obstructive sleep apnea severity. Sleep Breath 1:19. https://doi.org/10.1007/ S11325-023-02892-3 47. Gao J, Yang Y, Lin P, Park DS (2018) Computer vision in healthcare applications. J Healthc Eng 2018:15. https://doi.org/10. 1155/2018/5157020 48. Siddiqi MH, Khan K, Khan RU, Alsirhani A (2022) Face image analysis using machine learning: a survey on recent trends and applications. Electronics 11:1210. https://d oi.o rg/1 0.3 390/ ELECTRONICS11081210 49. Okubo M, Suzuki M, Horiuchi A et al (2006) Morphologic analyses of mandible and upper airway soft tissue by MRI of patients with obstructive sleep apnea hypopnea syndrome. Sleep 29:909915. https://doi.org/10.1093/SLEEP/29.7.909 50. Miles PG, Vig PS, Weyant RJ et al (1996) Craniofacial structure and obstructive sleep apnea syndromea qualitative analysis and meta-analysis of the literature. Am J Orthod Dentofacial Orthop 109:163172. https://doi.org/10.1016/S0889-5406(96)70177-4 Medical & Biological Engineering & Computing (2024) 62:12771311 51. Espinoza-Cuadros F, Fernndez-Pozo R, Toledano DT et al (2015) Speech signal and facial image processing for obstructive sleep apnea assessment. Comput Math Methods Med 2015:114. https://doi.org/10.1155/2015/489761 52. Tyan M, Espinoza-Cuadros F, Pozo RF et al (2017) Obstructive sleep apnea in women: study of speech and craniofacial characteristics. JMIR Mhealth Uhealth 5:e8238. https://doi.org/10. 2196/MHEALTH.8238 53. Lee RWW, Petocz P, Prvan T et al (2009) Prediction of obstructive sleep apnea with craniofacial photographic analysis. Sleep 32:46. https://doi.org/10.5665/sleep/32.1.46 54. Balaei AT, Sutherland K, Cistulli PA, de Chazal P (2017) Automatic detection of obstructive sleep apnea using facial images. In: 2017 IEEE 14th International symposium on biomedical imaging (ISBI 2017). IEEE, pp 215218. https://d oi.o rg/1 0.1 109/ ISBI.2017.7950504 55. Sutherland K, Lee RWW, Petocz P et al (2016) Craniofacial phenotyping for prediction of obstructive sleep apnoea in a Chinese population. Respirology 21:11181125. https://doi.org/10.1111/ resp.12792 56. Monna F, Ben Messaoud R, Navarro N et al (2022) Machine learning and geometric morphometrics to predict obstructive sleep apnea from 3D craniofacial scans. Sleep Med 95:7683. https://doi.org/10.1016/j.sleep.2022.04.019 57. Ciaparrone G, Chiariglione L, Tagliaferri R (2022) A comparison of deep learning models for end-to-end face-based video retrieval in unconstrained videos. Neural Comput Appl 34:74897506. https://doi.org/10.1007/S00521-021-06875-X/FIGURES/7 58. Islam SMS, Mahmood H, Al-Jumaily AA, Claxton S (2018) Deep learning of facial depth maps for obstructive sleep apnea prediction. In: 2018 International conference on machine learning and data engineering (iCMLDE). IEEE, pp 154157.https:// doi.org/10.1109/iCMLDE.2018.00036 59. Dae GJ, Hae YC, Grunstein RR, Yee B (2004) Predictive value of Kushida index and acoustic pharyngometry for the evaluation of upper airway in subjects with or without obstructive sleep apnea. J Korean Med Sci 19:662667. https://d oi.org/10. 3346/jkms.2004.19.5.662 60. Kushida CA, Efron B, Guilleminault C (1997) A predictive morphometric model for the obstructive sleep apnea syndrome. Ann Intern Med 127:581587. https://doi.org/10.7326/0003- 4819-127-8_PART_1-199710150-00001 61. Gelardi M, Del Giudice AM, Cariti F et al (2007) Acoustic pharyngometry: clinical and instrumental correlations in sleep disorders. Braz J Otorhinolaryngol 73:257265. https://doi.org/ 10.1016/S1808-8694(15)31075-2 62. DeYoung PN, Bakker JP, Sands SA et al (2013) Acoustic pharyngometry measurement of minimal cross-sectional airway area is a significant independent predictor of moderate-tosevere obstructive sleep apnea. J Clin Sleep Med 9:11611164. https://doi.org/10.5664/jcsm.3158 63. Shivarov G (2022) Anthropometry and acoustic pharyngometry of the oral cavity in sleep-disordered breathing. J IMAB Ann Proc Sci Pap 28:43774383. https://doi.org/10.5272/JIMAB. 2022282.4377 64. Salisbury JI, Sun Y (2007) Rapid screening test for sleep apnea using a nonlinear and nonstationary signal processing technique. Med Eng Phys 29:336343. https://doi.org/10.1016/j. medengphy.2006.05.013 65. Simply RM, Dafna E, Zigel Y (2018) Obstructive sleep apnea (OSA) classification using analysis of breathing sounds during speech. In: 2018 26th European signal processing conference (EUSIPCO). IEEE, pp 11321136. https://doi.org/10.23919/ EUSIPCO.2018.8553353 66. Moussavi Z, Elwali A, University of Manitoba (2021) System and methods for screening obstructive sleep apnea during 1309 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. wakefulness using anthropometric information and tracheal breathing sounds. U.S. Patent Application 17/349,298. Elwali A, Moussavi Z (2021) Predicting polysomnography parameters from anthropometric features and breathing sounds recorded during wakefulness. Diagnostics 11:905. https://doi. org/10.3390/DIAGNOSTICS11050905 Hajipour F, Moussavi Z (2019) Spectral and higher order statistical characteristics of expiratory tracheal breathing sounds during wakefulness and sleep in people with different levels of obstructive sleep apnea. J Med Biol Eng 39:244250.https:// doi.org/10.1007/s40846-018-0409-7 Elwali A, Moussavi Z (2019) Determining breathing sound features representative of obstructive sleep apnea during wakefulness with least sensitivity to other risk factors. J Med Biol Eng 39:230237. https://doi.org/10.1007/s40846-018-0410-1 Elwali A, Moussavi Z (2019) A novel decision making procedure during wakefulness for screening obstructive sleep apnea using anthropometric information and tracheal breathing sounds. Sci Rep 9:11467. https://doi.org/10.1038/ s41598-019-47998-5 Montazeri A, Giannouli E, Moussavi Z (2012) Assessment of obstructive sleep apnea and its severity during wakefulness. Ann Biomed Eng 40:916924. https://doi.org/10.1007/ s10439-011-0456-5 Hajipour F, Jozani MJ, Moussavi Z (2020) A comparison of regularized logistic regression and random forest machine learning models for daytime diagnosis of obstructive sleep apnea. Med Biol Eng Comput 58:25172529. https://doi.o rg/10.1007/ s11517-020-02206-9 Elwali A, Meza-Vargas S, Moussavi Z (2019) Using tracheal breathing sounds and anthropometric information for screening obstructive sleep apnoea during wakefulness. J Med Eng Technol 43:111123. https://doi.org/10.1080/03091902.2019.1617799 Hajipour F, Jozani MJ, Elwali A, Moussavi Z (2019) Regularized logistic regression for obstructive sleep apnea screening during wakefulness using daytime tracheal breathing sounds and anthropometric information. Med Biol Eng Comput 57:26412655. https://doi.org/10.1007/s11517-019-02052-4 Pasterkamp H (1996) Posture-dependent change of tracheal sounds at standardized flows in patients with obstructive sleep apnea. Chest J 110:1493. https://doi.org/10.1378/chest.110.6. 1493 Sola-Soler J, Fiz JA, Torres A, Jane R (2014) Identification of obstructive sleep apnea patients from tracheal breath sound analysis during wakefulness in polysomnographic studies. In: 2014 36th Annual international conference of the ieee engineering in medicine and biology society. IEEE, pp 42324235. https://doi.org/10.1109/embc.2014.6944558 Sola-Soler J, Jane R, Fiz JA, Morera J (2008) Formant frequencies of normal breath sounds of snorers may indicate the risk of obstructive sleep apnea syndrome. In: 2008 30th Annual international conference of the IEEE engineering in medicine and biology society. IEEE, pp 35003503. https://doi.org/10. 1109/IEMBS.2008.4649960 Pozo RF, Murillo JLB, Gmez LH et al (2009) Assessment of severe apnoea through voice analysis, automatic speech, and speaker recognition techniques. EURASIP J Adv Sig Process 2009:111. https://doi.org/10.1155/2009/982531 Montero Benavides A, Fernndez Pozo R, Toledano DT et al (2014) Analysis of voice features related to obstructive sleep apnoea and their application in diagnosis support. Comput Speech Lang 28:434452. https://doi.org/10.1016/J.CSL.2013. 08.002 Espinoza-Cuadros F, Fernndez-Pozo R, Toledano DT et al (2016) Reviewing the connection between speech and 1310 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. Medical & Biological Engineering & Computing (2024) 62:12771311 obstructive sleep apnea. Biomed Eng Online 15:120. https:// doi.org/10.1186/S12938-016-0138-5 Montero Benavides A, Blanco Murillo JL, Fernndez Pozo R et al (2016) Formant frequencies and bandwidths in relation to clinical variables in an obstructive sleep apnea population. J Voice 30:2129. https://doi.org/10.1016/J.JVOICE.2015.01. 006 Goldshtein E, Tarasiuk A, Zigel Y (2011) Automatic detection of obstructive sleep apnea using speech signals. IEEE Trans Biomed Eng 58:13731382. https://doi.org/10.1109/TBME. 2010.2100096 Simply RM, Dafna E, Zigel Y (2020) Diagnosis of obstructive sleep apnea using speech signals from awake subjects. IEEE J Sel Top Sign Proces 14:251260. https://d oi.o rg/1 0.1 109/J STSP. 2019.2955019 Sol-Casals J, Munteanu C, Martn OC et al (2014) Detection of severe obstructive sleep apnea through voice analysis. Appl Soft Comput 23:346354. https://doi.org/10.1016/J.ASOC.2014.06. 017 Pang K-G, Hsung T-C, Liao G et al (2022) Obstructive sleep apnea detection using speech signals with high frequency components. J Commun 17:4955. https://d oi.o rg/1 0.1 2720/j cm.1 7.1. 49-55 Ylmaz D, Yldz M, Uyar Toprak Y, Yetkin S (2023) Obstructive sleep apnea detection with nonlinear analysis of speech. Biomed Signal Process Control 84:104956. https://doi.org/10. 1016/j.bspc.2023.104956 Ramachandran SK, Josephs LA (2009) A meta-analysis of clinical screening tests for obstructive sleep apnea. Anesthesiology 110:928939. https://doi.org/10.1097/ALN.0b013e31819c47b6 Behar JA, Palmius N, Daly J et al (2017) Sleep questionnaires in screening for obstructive sleep apnoea. Comput Cardiol (2010) 44:14. https://doi.org/10.22489/CinC.2017.233-188 Kump K, Whalen C, Tishler PV et al (1994) Assessment of the validity and utility of a sleep-symptom questionnaire. Am J Respir Crit Care Med 150:735741. https://doi.org/10.1164/ ajrccm.150.3.8087345 Pataka A, Daskalopoulou E, Kalamaras G et al (2014) Evaluation of five different questionnaires for assessing sleep apnea syndrome in a sleep clinic. Sleep Med 15:776781. https://doi. org/10.1016/J.SLEEP.2014.03.012 Solecka , Matler K, Kostliv T et al (2022) A comparison of the reliability of five sleep questionnaires for the detection of obstructive sleep apnea. Life 12:1416. https://doi.org/10.3390/ life12091416 Shahar E, Whitney CW, Redline S et al (2001) Sleep-disordered breathing and cardiovascular disease: cross-sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med 163:1925. https://doi.org/10.1164/AJRCCM.163.1.2001008 Caseiro P, Fonseca-Pinto R, Andrade A (2010) Screening of obstructive sleep apnea using Hilbert-Huang decomposition of oronasal airway pressure recordings. Med Eng Phys 32:561568 Young T, Palta M, Dempsey J et al (1993) The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 328:12301235. https://doi.org/10.1056/NEJM199304 293281704 Simas R, Maestri F, Normando D (2014) Controlling false positive rates in research and its clinical implications. Dental Press J Orthod 19:24. https://doi.org/10.1590/2176-9451.19.3.024-025. EBO 96. Olliaro P, Torreele E (2021) Managing the risks of making the wrong diagnosis: first, do no harm. Int J Infect Dis 106:382. https://doi.org/10.1016/J.IJID.2021.04.004 97. Burke JF, Sussman JB, Kent DM, Hayward RA (2015) Three simple rules to ensure reasonably credible subgroup analyses. The BMJ 351:15. https://doi.org/10.1136/BMJ.H5651 98. Kim HE, Kim HH, Han BK et al (2020) Changes in cancer detection and false-positive recall in mammography using artificial intelligence: a retrospective, multireader study. Lancet Digit Health 2:e138e148. https://doi.org/10.1016/S2589-7500(20) 30003-0 99. Meshkinfamfard S, Gorban A, Tyukin I (2018) Tackling rare false-positives in face recognition: a case study. In: 2018 IEEE 20th International conference on high performance computing and communications; IEEE 16th International conference on smart city; IEEE 4th International conference on data science and systems (HPCC/SmartCity/DSS). IEEE, pp 15921598. https:// doi.org/10.1109/HPCC/SMARTCITY/DSS.2018.00260 100. Shen Y, Shamout FE, Oliver JR et al (2021) Artificial intelligence system reduces false-positive findings in the interpretation of breast ultrasound exams. Nat Commun 12:113. https://doi.org/ 10.1038/s41467-021-26023-2 101. Alsubie HS, BaHammam AS (2017) Obstructive sleep apnoea: children are not little adults. Paediatr Respir Rev 21:7279. https://doi.org/10.1016/J.PRRV.2016.02.003 Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Ali Mohammad Alqudah is a Ph.D. student and graduate research assistant in biomedical engineering at the University of Manitoba. His scientific interests are Biomedical Signals and Images Processing and Analysis, Deep Learning, Machine Learning, and obstructive sleep apnea. Ahmed Elwali A former Ph.D. from University of Manitoba and currently an assistant professor of biomedical engineering at Marian University. His Scientific interests are in signal processing, machine learning, electronics, and obstructive sleep apnea. Medical & Biological Engineering & Computing (2024) 62:12771311 1311 Brendan Kupiak A MSc student in the electrical and computer engineering department at the University of Manitoba. His Scientific interests are in signal processing, machine learning, and o b s t r u c t i ve s l e e p a p n e a detection. Natasha Jacobson A former Ph.D. from University of McGill and post-doctoral fellow in the biomedical engineering program at the University of Manitoba. Currently, she is an instructor at the Department of Biosystems Engineering, University of Manitoba. Her Scientific interests are in soft tissue mechanics, human modelling, medical instrumentation. Farahnaz Hajipour A former Biomedical engineering PhD graduate (University of Manitoba). Her Scientific interests are in signal processing, machine learning, and obstructive sleep apnea detection. Currently, she is a data scientist ad SkipTheDishes, Canada. Zahra Moussavi is a professor and Canada Research Chair in Biomedical Engineering at the University of Manitoba. Her Scientific interests are in signal processing, obstructive sleep apnea and Alzheimers diagnosis and treatment.  ...

... bioRxiv preprint doi: https://doi.org/10.1101/2024.01.08.574670; this version posted January 9, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Effectiveness of a network Open House model to recruit trainees to post-baccalaureate STEM programs Scott Takeo Aoki1, Lindsay Lewellyn2, Sarah Justice1,3, Sarah Mordan-McCombs4, Neetu Tewari5, Jorge Cantu6, Robert Seiser7, Ahmed Lakhani8, Jennifer R. Kowalski2* Department of Biochemistry and Molecular Biology, Indiana University School of Medicine, 635 Barnhill Drive, Indianapolis, IN 46202, USA 2 Department of Biological Sciences, Butler University, 4600 Sunset Avenue, Indianapolis, IN 46208, USA 3 Department of Biology, Marian University, 3200 Cold Spring Road, Indianapolis, IN 46222, USA 4 Department of Biology, DePauw University, 1 E Hanna Drive, Greencastle, IN 46135 USA 5 Department of Biological Science, East-West University, 816 S. Michigan Avenue, Chicago, Illinois 60605 6 Department of Biology, Northeastern Illinois University, 5500 N St Louis Ave, Chicago, IL 60625 7 Department of Biological, Physical and Health Sciences, Roosevelt University, 1400 N Roosevelt Boulevard, Schaumburg, IL 60173 8 Department of Biomedical and Health Sciences, Calumet College of St. Joseph, 2400 New York Avenue, Whiting, IN, 46394 1 *Corresponding author: jrkowals@butler.edu Keywords (up to 7): post-baccalaureate, open house, underrepresented, diversity, recruitment, STEM, minority 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 bioRxiv preprint doi: https://doi.org/10.1101/2024.01.08.574670; this version posted January 9, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. Abstract Post-baccalaureate (post-bac) programs can have a positive impact on science training and STEM career opportunities for junior trainees. A goal for many of these sponsored programs is to increase research exposure for underrepresented minorities, a group that can include African American, Hispanic, Native American, and firstgeneration college students, among others. Recruiting underrepresented minorities to post-bac programs can be challenging, for reasons that include a lack of available research opportunities, time to pursue these experiences, and awareness of available programs. To this end, an Open House event was created to inform and excite potential students for future post-bac programs. Students were recruited from partnering Minority Serving Institutions (MSIs) to attend a two-day event at a primarily undergraduate institution (PUI) and a research- intensive R1 institution. The students visited both campuses, were informed about post-bac programs and potential research opportunities, and met with faculty, current graduate students, and a former post-bac scholar. Transportation, lodging, and meals were provided. Visiting students completed voluntary pre- and postsurveys. Results indicated that attendees, the majority of whom were underrepresented minorities in STEM, left the event with an increased understanding about post-bac programs and their benefits to a career in STEM and that their attendance at the event made it more likely they would apply to available post-bac programs. Thus, this work demonstrates that in-person events involving integrative partnerships across multiple universities are effective strategies for increasing awareness of opportunities available to students post-graduation and for recruiting underrepresented groups in STEM to post-bac programs. Introduction Starting a career in science depends on extensive hands-on experience. For many, laboratory research experience begins in their high school or undergraduate education, but for others, obligations outside the classroom prevent them from experiencing bench research firsthand. This challenge is often observed with students who identify as underrepresented minorities in science or have come through a community college system (1, 2), and it can limit individuals belonging to these groups from obtaining lab research experience necessary for graduate programs or employment in STEM careers. For example, graduate schools look for meaningful research experience in their candidates. In many programs, matriculating graduate students are years past their undergraduate education (3), giving them time to obtain relevant research experience that they might not have had the opportunity to pursue while working towards their bachelors degree. Developing opportunities for students to gain experience after their undergraduate training is central to recruiting a diverse, balanced population to the STEM workforce, but many of those who would benefit most from these opportunities may be unaware of their existence or benefits. Post-baccalaureate programs are one to two-year funded, research-intensive training experiences designed to prepare trainees for graduate school and STEM careers. Some of these programs have been active for several years. For example, the National Institutes of Health (NIH) Postbaccalaureate Research Education Program (PREP) program is in its third decade and supports post-bac trainees at a variety of research institutions across the country (4). This program has evolved new strategies to promote readiness for STEM graduate school (5, 6) and is incredibly successful. Currently, 65-97% of PREP scholars enter graduate school programs, and Ph.D. completion rates are > 65% above the rates reported for underrepresented groups in the life sciences (6-8). The American Cancer Society (ACS) and National Science Foundation (NSF) have recently developed post-bac programs with similar structural models (9, 10). All these programs recognize the need to support research experiences for underrepresented minorities in science. A funded research experience outside of schooling promises more opportunity to recruit a breadth of students from a wide demographic, but a challenge faced by post-bac programs is how to reach trainees who may not be familiar with the benefits of these programs or who are disconnected from pathways that lead to a successful STEM career. An Open House event invites candidate trainees on site to introduce a program and present opportunities available to them. These events are flexible by design and can be impactful well past the traditional K-12 use of such events. Targeted, personal Open House-like events can be helpful in recruiting individuals from specific demographics, like those who identify as female and African Americans (11). Students considering various undergraduate programs also have identified Open House events as an effective recruiting tool (12). Universities note that Open House events are a chance to present a positive image to visitors (13). Open Houses are a chance for real human connection, which can showcase the advantages of an educational program to groups of people missed through other advertising campaigns. 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 bioRxiv preprint doi: https://doi.org/10.1101/2024.01.08.574670; this version posted January 9, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. In this study, an Open House event was developed to introduce the benefits of post-bac programs, with an emphasis on reaching students from groups underrepresented in the biological sciences (14) with little previous research experience. Faculty and students from research-intensive R1s, primarily undergraduate institutions (PUIs), and minority serving institutions (MSIs) that form collaborative research networks are effective in undergraduate biology training (15), and personalized referrals are among the most effective strategies for recruiting students from underrepresented minority groups to STEM graduate school (11). In consideration of these factors, an event was crafted that leveraged the strengths of faculty partnerships across a network of MSI, PUI, and R1 institutions. The effort created an experience that reached a cohort of students from underrepresented minority groups in science and presented post-bac programs as a viable steppingstone for a STEM career. This strategy can be modified to present the strengths of any university, training program, or geographical area. Thus, STEM training programs may consider hosting similar events to increase the diversity of their applicant cohort. Materials and Methods Open House Event and Survey Format Recruitment for the Open House was performed through advertising and word of mouth. The advertising flyer was created in Canva (Canva, Sydney, Australia; www.canva.com), which contained a QR code linked to a Google Form (Google; Mountain View, CA; www.google.com) for registration. Students were selected on a first come, first serve basis. Partnering MSIs were given first access to registration, followed by students at the hosting institutions. In total, 17 students were recruited to the event, with 15 attending on both days. Students and faculty from their home institutions were responsible for arranging travel to Indianapolis, IN. Hotels were reserved through Butler University, the primary hosting institution. Day 1. Students and faculty arrived at Butler University, a PUI in Indianapolis, IN. Prior to scheduled events (Fig S1), students completed an anonymous pre-survey (Supplemental Information 1), approved by a Butler University IRB (Approval date: Sept. 18, 2023) and administered by Qualtrics (Qualtrics; Provo, UT), taking approximately 10-15 minutes to complete. This survey requested information regarding the participants demographics, science experiences, and familiarity with and interest in post-bac programs. A total of 17 students completed the pre-survey. Students then learned of the opportunities for post-bacs and those with science graduate degrees (e.g., M.S., Ph.D.), research opportunities at local PUIs, and resources available at Butler University. A tour of the Butler University campus was made available for those interested. Visiting students and faculty then were taken to dinner with faculty interested in hosting post-bacs and with graduate students from the Indiana University School of Medicine, an R1 Research Institution. Visiting faculty and students stayed at a local hotel sponsored by the program. Day 2. Students and faculty visited Indiana University School of Medicine; Indianapolis, IN (Fig S1). They were given an overview of an established post-bac program (https://iprep.iupui.edu/index.html) and research at Indiana University and interacted with a graduate student panel assembled by the local chapter of the Society for the Advancement of Chicanos/Hispanics and Native Americans in Science (SACNAS). Tours of the Centers of Electron Microscopy and Proteome Analysis facilities were given. A sponsored lunch was provided with Indiana University School of Medicine faculty members and graduate students. Visiting students were prompted to complete a Qualtrics exit survey consisting of the similar questions regarding post-bac programs (Supplemental Information 2). A total of 13 students completed this exit survey. Data Analysis Anonymized pre- and post-event survey data were aggregated separately and analyzed for statistical significance in GraphPad Prism version 10.1.1 for MacOS (GraphPad Software, Boston, Massachusetts USA). Figure 1A and B data were analyzed using a Mann Whitney U test to compare pre- and post-survey Likert score means converted to a 1-5 scale. Figure 1C data were analyzed using a One Sample Sign Test (One sample t and Wilcoxon test in Prism) with 3.0 neither at the middle of the 1-5 Likert scale set as the theoretical mean value. Figures were also made using Prism and Adobe Illustrator (Adobe, San Jose, CA). Qualtrics data for all survey questions are included in the Supplemental Information 1 and 2. Results The goal this project was to develop an event that could recruit applicants from a range of backgrounds to post baccalaureate programs. To this end, an Open House was created to advertise a potential post-bac program to 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 bioRxiv preprint doi: https://doi.org/10.1101/2024.01.08.574670; this version posted January 9, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. students in Indiana and the Chicago area. Partnerships were first established between three Indianapolis area PUIs that are proximal to a centrally located R1 institution. Next, additional partnerships were formed with four MSIs in the Northern Indiana/Chicago area. Faculty at these MSIs interact regularly with many students from underrepresented groups, as defined by both the NIH (16) and NSF (17). Each MSI had a faculty contact who facilitated event advertising and chaperoned students to the Open House. A full schedule of talks and social events were planned (Fig S1) and held at Butler University and Indiana University School of Medicine. Students learned about scientific research and professional opportunities for those entering post-bac programs and STEM careers. Discussion forums and meals were included, which allowed visiting students to discuss post-bac programs and graduate school with R1 graduate students from SACNAS and with faculty from PUI and R1 institutions. Voluntary, anonymous pre- and post-surveys were administered at the beginning and ending of the Open House. The pre-survey solicited demographic information of the students attending the event (Supplementary Information 1). Information was collected regarding age, year in school, sexuality, gender, disability, military service, education, science exposure, career goals, and the attendees knowledge of the concept of and opportunities available in post-bac programs. All results are provided for those who responded (Supplementary Information 1). Of note, 76.2% of total pre-survey respondents identified as an underrepresented racial/ethnic minority, including Black/African American (33.3%), Hispanic/Latinx/a/o/e (38.1%), or Indigenous/American Indian or Alaskan Native (4.8%). Additionally, 17.6% of respondents indicated they had a disability according to the NIH/NSF definition (16-19). Only 18.8% reported having a family member in the household with a 4-year degree or higher. While most respondents (94.1%) reported pursuit of a bachelors degree in science, less than half (47.1%) could identify a science role model. A similar percentage (57.1%) reported that they did not pursue independent research in their undergraduate education, either because it was not available or because they chose not to participate. The responses indicated that limited time due to work or personal obligations (32.1%) and access to knowledge regarding research activities (25.0%) were both significant factors in deciding whether to pursue undergraduate research. In sum, the students recruited to this Open House were members of groups typically underrepresented in science with limited exposure to science research. Analysis of pre- and post-survey data indicated that the attending students learned about and had a positive impression of the post-bac program. Responses showed that students gained a statistically significantly improvement in their understanding of post-bac training programs and what they entail after attending the Open House (Fig 1A; U (NPre=15, NPost=13) = 26.5, p = 0.0003). Students also expressed a strong interest in pursuing a post-bac opportunity (Fig 1B). Although the pre- to post-survey gains were not statistically significant for this question [(U (NPre=15, NPost=13) = 71, p = 0.192], this is likely due to both small sample sizes and the high number of students agreeing with the statement despite not being very familiar with post-bac programs in the presurvey. Nevertheless, more students strongly agreed they were interested in pursuing a post-bac program in the post-survey (MeanPre = 4.0; MeanPost = 5.0). The responses for the Open House event were universally positive and indicated a statistically significant increase in the likelihood attendees would apply for a post-bac program (Fig 1C) [one sample, t(df) = 10.65 (11); p < 0.0001]. The most positive experiences came from hearing about the benefits of a post-bac program (75%), an overview of a model post-bac program (75%), and the graduate student panel (83%). Anecdotally, student survey responders commented that they definitely sold me on (the location)and all the programs offered, that the event was really informative, and that the event was really fun and insightful. I found out more about post bac programs and the benefits. While some students commented in the pre-survey that they were worried about the location away from home, being at a predominantly white institution, and being unsure whether completing the post-bac program would lead to something, none of these concerns appeared in post-survey responses. Thus, the Open House may have been successful in addressing students concerns. In fact, one respondent in the post-survey stated, That being away from home and finding a new place to live and having to start out my own with this change is daunting but Im sure Im capable of doing it. In sum, the network-based Open House event delivered a positive experience and was successful in informing students about the benefits of a post-bac program to pursuing future careers in STEM. Discussion Post-baccalaureate recruitment of underrepresented minorities can be challenging due to a lack of science exposure and personalized interactions. To improve outreach to underserved populations in science, an open house event was established to advertise post-bac programs to students from MSIs and surrounding universities. 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 bioRxiv preprint doi: https://doi.org/10.1101/2024.01.08.574670; this version posted January 9, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. Students visited the campuses of a PUI and an R1 institution, heard about post-bac programs and graduate school, and had a chance to socialize with faculty and students. Pre- and post-surveys performed indicated that many of the students who visited represented underserved minorities in science and that the Open House both informed and left a positive impact on their impressions of post-bac programs. Hence, direct personalized events leveraging the strengths of multiple institutions is a viable strategy to encourage trainees to pursue post-bac opportunities. MSI partnership to enhance science outreach and development is a well-established strategy. Personal referrals are an effective means to recruit students to graduate programs (11). Furthermore, MSI partnerships have aided in recruitment of underrepresented minorities in sciences into a physical sciences graduate program (20), and encouraged participation in STEM research with the National Oceanic and Atmospheric Administration (NOAA) (21). National programs like the Leadership Alliance, comprised of 32 institutions ranging from Ivy League schools and R1s to MSIs, have been collaboratively mentoring underrepresented minority students from undergraduate through graduate training for 30 years (22). Similarly, this Open House event relied heavily on MSI faculty to recruit students through word of mouth and flyer distribution. MSI faculty members also accompanied their students to the two-day event. Personalized mentorship is known to enhance a students STEM experience and decision to enter STEM careers (23). Thus, personalized experiences, like invitations from faculty at their own institutions to an Open House event, are likely to increase the likelihood that students will apply to post-bac programs. Improvements will further refine the effectiveness of the Open House. First, while MSI student participants expressed many positive sentiments regarding their experience at the event, informal conversations with student and faculty attendees indicated that they would like additional time to explore the local area, including housing options and neighborhood information, as well as a more comprehensive overview of research departments and areas, while also ensuring that research talks are as accessible as possible to a wide range of students. Second, scheduling the Open House at a time that was mutually convenient for all institutions, each with their own unique academic calendars, while also avoiding local hotel event conflicts, was challenging. Continued communication and advance planning, as well as pairing the in-person event with virtual office hours and other campus visits by post-bac program faculty and student representatives should minimize these challenges in the future. Third, although advertising with the partnered MSIs was effective for recruiting Open House attendees, less effort was placed on recruiting students in the area. Local students represent an additional, potentially high yield population for a post-bac program, as they would not need extensive travel to attend the Open House, and some would likely identify as an underserved minority in science. Thus, recruiting local students to post-bac programs may be extremely fruitful, as they may be comfortable committing to a program in which they know the area, universities, and faculty members involved. More effort should be placed to advertise such Open House events to all students, near and far. Fourth, many students who attended the Open House event had already made career choices. Many students were interested in clinical professions, with less than half citing research as their career goal (Supplemental Information 1). Student mindset can change, but it may be advantageous to target college students who are undecided or leaning toward a non-clinical STEM career, as these students will be the strongest candidates for post-bac programs. Continued personalized invitations to such students from MSI, PUI, and R1 faculty, along with providing additional STEM-career focused information to candidates, will likely be most effective in achieving this goal (11). As designed, the Open House format permits flexibility for hosts to reconfigure and emphasize strengths of their geographical area, research programs, and partners to recruit their desired post-bac cohort. Conclusion Overall, this work provides evidence that having in-person Open House events is an effective way to inform students, and particularly those from groups underrepresented in STEM, about post-bac programs. Post-bac programs continue to gain traction because of their strengths in preparing students for graduate school. These training opportunities are promising avenues to recruit talent from all walks of life into STEM careers. Virtual office hours and flyer advertising on university boards or email are affordable and can be effective for the student knowledgeable about the next steps in a STEM career. However, to recruit students unaware of the possibilities in a science career, a more active recruitment process, such as an Open House event, may aid in identifying talent outside of the normal cohort. This Open House model, which capitalized on the synergy of a network of partner institutions (MSIs, PUIs, and RIs), is one method for successfully identifying post-bac 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 bioRxiv preprint doi: https://doi.org/10.1101/2024.01.08.574670; this version posted January 9, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. candidates from underrepresented groups and sharing with them the benefits of participating in a post-bac program as an integral step in their STEM career progression. Acknowledgments. The authors thank the students from partnering MSIs and hosting institutions for their attendance and participation. They also thank Drs. Ann Kimble-Hill, Evan Cornett, Qiuyan Chen, Emma Doud, Yangshin Park; Ms. Carmen Herrera-Sandoval, Moraima Noda; Mr. Rodney Claude, Derrick Gray, Miguel Barriera Diaz; and SACNAS (Indiana University School of Medicine (IUSM)), Center for Electron Microscopy (IUSM), and the Center for Proteome Analysis (IUSM) for speaking about their science and available post-bac programs, as well as Mr. Randall Ojeda and Ms. Mikala Lain (Butler Efroymson Diversity Center) for sharing diversity and inclusion resources and Dr. Rob Denton (Marian University) for speaking about his science. Additional thanks go out to Butler University and Indiana University School of Medicine for use of their facilities; faculty from Butler, Marian, and Indiana Universities for attending the dinner and lunch; and the Aoki Lab for help with lunch set up and clean up. Finally, the authors thank Dr. Andrew Stoehr (Butler University) for advice on statistical analysis. This project was funded by the Butler University Provosts Office. Figure Legends Figure 1. Effectiveness of an Open House event in educating and promoting post-baccalaureate programs. (A, B) Pre- and post-event surveys of (A) student familiarity with post-baccalaureate training programs and (B) student interest in participating in a post-baccalaureate training program (N = 15 pre; N = 13 post). (C) Post-survey responses regarding the impact of the Open House event on the likelihood of their future application to a post-baccalaureate training program (N = 12). See Results text for statistical analysis. Figure S1. Open House Agenda Supplemental Information 1. Open House Pre-Survey Results Supplemental Information 2. Open House Post-Survey Results. 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 bioRxiv preprint doi: https://doi.org/10.1101/2024.01.08.574670; this version posted January 9, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. References 1. Mahatmya D, Morrison J, Jones RM, Garner PW, Davis SN, Manske J, et al. Pathways to Undergraduate Research Experiences: a Multi-Institutional Study. Innovative Higher Education. 2017;42(5):491-504. doi: 10.1007/s10755-017-94013. 2. Hirst RA, Bolduc G, Liotta L, Packard BW-L. Cultivating the STEM Transfer Pathway and Capacity for Research: A Partnership Between a Community College and a 4-Year College. Journal of College Science Teaching. 2014;43(4):12-7. 3. Park HY, Berkowitz O, Symes K, Dasgupta S. The art and science of selecting graduate students in the biomedical sciences: Performance in doctoral study of the foundational sciences. PLoS One. 2018;13(4):e0193901. Epub 20180403. doi: 10.1371/journal.pone.0193901. PubMed PMID: 29614110; PubMed Central PMCID: PMC5882097. 4. Postbaccalaureate Research Education Program (PREP) (R25): National Institutes of Health; 2024. Available from: https://www.nigms.nih.gov/training/PREP/Pages/default.aspx. 5. Gazley JL, Remich R, Naffziger-Hirsch ME, Keller J, Campbell PB, McGee R. Beyond Preparation: Identity, Cultural Capital, and Readiness for Graduate School in the Biomedical Sciences. J Res Sci Teach. 2014;51(8):1021-48. doi: 10.1002/tea.21164. PubMed PMID: 26366013; PubMed Central PMCID: PMC4564061. 6. Hardy TM, Hansen MJ, Bahamonde RE, Kimble-Hill AC. Insights Gained into the Use of Individual Development Plans as a Framework for Mentoring NIH Postbaccalaureate Research Education Program (PREP) Trainees. J Chem Educ. 2022;99(1):417-27. Epub 20211124. doi: 10.1021/acs.jchemed.1c00503. PubMed PMID: 36186731; PubMed Central PMCID: PMC9521764. 7. Hall A, Mann J, Bender M. Analysis of Scholar Outcomes for the NIGMS Postbaccalaureate Research Education Program 2015. Available from: https://loop.nigms.nih.gov/2015/09/outcomes-analysis-of-the-nigms-postbaccalaureateresearch-education-program-prep/. 8. Schwartz NB, Risner LE, Domowicz M, Freedman VH. Comparisons and Approaches of PREP Programs at Different Stages of Maturity: Challenges, Best Practices and Benefits. Ethn Dis. 2020;30(1):55-64. Epub 20200116. doi: 10.18865/ed.30.1.55. PubMed PMID: 31969784; PubMed Central PMCID: PMC6970524. 9. Research and Mentoring for Postbaccalaureates in Biological Sciences (RaMP): National Science Foundation; 2024. Available from: https://new.nsf.gov/funding/opportunities/research-mentoring-postbaccalaureates-biological. 10. ACS Center for Diversity in Cancer Research (DICR) Training: American Cancer Society; 2024. Available from: https://www.cancer.org/research/acs-center-for-diversity-in-cancer-research-training.html. 11. Shadding CR, Whittington D, Wallace LE, Wandu WS, Wilson RK. Cost-Effective Recruitment Strategies That Attract Underrepresented Minority Undergraduates Who Persist to STEM Doctorates. SAGE Open. 2016;6(3):2158244016657143. doi: 10.1177/2158244016657143. 12. Gray M, Daugherty MK. Factors that Influence Students to Enroll in Technology Education Programs. Journal of Technology Education. 2004;15. doi: 10.21061/jte.v15i2.a.1. 13. Fischbach R. Assessing the impact of university open house activities. College Student Journal. 2006;40:227+. 14. Pew Research Center. STEM Jobs See Uneven Progress in Increasing Gender, Racial and Ethnic Diversity. 2021 April 2021. Report No. 15. Jensen-Ryan D, Murren CJ, Rutter MT, Thompson JJ. Advancing Science while Training Undergraduates: Recommendations from a Collaborative Biology Research Network. CBE Life Sci Educ. 2020;19(4):es13. doi: 10.1187/cbe.20-05-0090. PubMed PMID: 33215973; PubMed Central PMCID: PMC8693944. 16. Notice of NIH's Interest in Diversity: National Institutes of Health; 2019. Available from: https://grants.nih.gov/grants/guide/notice-files/NOT-OD-20-031.html. 17. Diversity and STEM: National Science Foundation; 2024. Available from: https://ncses.nsf.gov/pubs/nsf23315/faqs#:~:text=Underrepresented%20minorities%20include%20individuals%20of,Ameri can%20Indians%20or%20Alaska%20Natives. 18. Americans With Disabilities Act of 1990, (1990). 19. (NCSES) NCfSaES. Diversity and STEM: Women, Minorities, and Persons with Disabilities 2023. Alexandria, VA: National Science Foundation, 2023. 20. Stassun KG, Burger A, Lange SE. The Fisk-Vanderbilt Masters-to-PhD Bridge Program: A Model for Broadening Participation of Underrepresented Groups in the Physical Sciences through Effective Partnerships with Minority-Serving Institutions. Journal of Geoscience Education. 2010;58(3):135-44. doi: 10.5408/1.3559648. 21. Robinson L, Rousseau J, Mapp D, Morris V, Laster M. An Educational Partnership Program with Minority Serving Institutions: A Framework for Producing Minority Scientists in NOAA-Related Disciplines. Journal of Geoscience Education. 2007;55(6):486-92. doi: 10.5408/1089-9995-55.6.486. 22. Ghee M, Collins D, Wilson V, Pearson Jr W. The Leadership Alliance: Twenty Years of Developing a Diverse Research Workforce. Peabody Journal of Education. 2014;89(3):347-67. doi: 10.1080/0161956X.2014.913448. 23. Estrada M, Hernandez PR, Schultz PW. A Longitudinal Study of How Quality Mentorship and Research Experience Integrate Underrepresented Minorities into STEM Careers. CBELife Sciences Education. 2018;17(1):ar9. doi: 10.1187/cbe.17-04-0066. bioRxiv preprint doi: https://doi.org/10.1101/2024.01.08.574670; this version posted January 9, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. Figure 1 A Familiarity with the post-baccalaureate training program? Response (%) 80 Pre-Survey Post-Survey 60 40 20 Ve r y U U nf am ila r nf am ili ar N ei th er Fa m Ve ila ry r Fa m ila r 0 B Interest in participating in a post-baccalaureate training program? Response (%) 80 Pre-Survey Post-Survey 60 40 20 St ro St ng ro ng ly A gr ee ee gr A th er e ei re N ag is D ly D is ag re e 0 C Open House make it more likely that I apply to post-baccalaureate training program Response (%) 80 60 40 20 St ro ng ly D is a D gre is e ag N ree ei St th ro e ng Ag r ly re A e gr ee 0 Post-Survey bioRxiv preprint doi: https://doi.org/10.1101/2024.01.08.574670; this version posted January 9, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. Figure S1. Post-Baccalaureate Open House Agenda Day 1: Primary Undergraduate Institution (PUI) Time Activity 12:00p 1:00p Lunch, pre-survey taking 1:00p 1:10p Welcome and introductions, PUI 1:10p 1:25p What to do with a science degree: A brief overview of graduate school and careers in STEM 1:25p 1:35p Undergraduate vs. Graduate School and how Post-bac programs can bridge the gap 1:35p 1:55p Life as a post-bac, previous post-bac turned graduate student 1:55p 2:05p Break 2:05p 2:20p Overview of a post-bac program: goals, design, student timeline 2:20p 2:35p Example 1: PUI faculty research 2:35p 2:50p Example 2: PUI faculty research 2:50p 3:00p Break 3:00p 3:30p Butler Campus tour - Labs/science area 3:30p 4:00p Diversity resources 4:00p 6:00p Break, check in to hotel 6:00p 8:00p Dinner and Networking Reception *All PUI, MSI, RI, faculty & students invited Day 2: R1 Research Institution (R1) Time Activity 8:30a 9:00a Coffee and baked goods 9:00a 9:10a Welcome and introductions, IUSM 9:10a 9:40a Getting from here to there: Benefits to a post-bac program and life as a post-bac in Indianapolis (faculty post-bac expert) 9:40a 10:10a 10:10a 10:25a Graduate Student Panel, Society for the Advancement of Chicanos/Hispanics and Native Americans in Science (SACNAS): Trainee life in Indianapolis and benefits to post-bac program Break 10:25a 10:40a Example 1: R1 faculty research 10:40a 10:55a Example 2: R1 faculty research 10:55a 11:05a Break 11:05a 12:00p R1 Campus tour Centers for Electron Microscopy and Proteomics 12:00p 1:00p Thank you, Lunch, survey  ...

... Received: 16 May 2023 | Accepted: 1 December 2023 DOI: 10.1111/2041-210X.14279 RESEARCH ARTICLE A conceptual framework for host-associated microbiomes of hybrid organisms Benjamin T. Camper1 | Zachary Laughlin1 Robert Denton2 | Sharon Bewick1 1 Department of Biological Sciences, Clemson University, Clemson, South Carolina, USA 2 Biology Department, Marian University, Indianapolis, Indiana, USA Correspondence Benjamin T. Camper Email: btcampers@gmail.com Funding information NSF: Division of Integrative Organismal Systems, Grant/Award Number: 2105604; Clemson University Support for Early Exploration and Development (CUSEED) Grant; Clemson University Creative Inquiry (CI) Program Handling Editor: Antonino Malacrin | Daniel Malagon1 | Abstract 1. Hybridization between organisms from evolutionarily distinct lineages can have profound consequences on organismal ecology, with cascading effects on fitness and evolution. Most studies of hybrid organisms have focused on organismal traits, for example, various aspects of morphology and physiology. However, with the recent emergence of holobiont theory, there has been growing interest in understanding how hybridization impacts and is impacted by host-associated microbiomes. Better understanding of the interplay between host hybridization and host-associated microbiomes has the potential to provide insight into both the roles of host-associated microbiomes as dictators of host performance as well as the fundamental rules governing host-associated microbiome assembly. Unfortunately, there is a current lack of frameworks for understanding the structure of host-associated microbiomes of hybrid organisms. 2. In this paper, we develop four conceptual models describing possible relationships between the host-associated microbiomes of hybrids and their progenitor or parent taxa. We then integrate these models into a quantitative 4H index and present a new R package for calculation, visualization and analysis of this index. 3. We demonstrate how the 4H index can be used to compare hybrid microbiomes across disparate plant and animal systems. Our analyses of these data sets show variation in the 4H index across systems based on host taxonomy, host site and microbial taxonomic group. 4. Our four conceptual models, paired with our 4H index and associated visualization tools, facilitate comparison across hybrid systems. This, in turn, allows for systematic exploration of how different aspects of host hybridization impact the host-associated microbiomes of hybrid organisms. KEYWORDS Aitchison simplex, holobiont, host-associated microbiome, hybrid, R package This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. 2024 The Authors. Methods in Ecology and Evolution published by John Wiley & Sons Ltd on behalf of British Ecological Society. Methods Ecol Evol. 2024;15:511529.  wileyonlinelibrary.com/journal/mee3 | 511 | 1 | CAMPER et al. I NTRO D U C TI O N be important determinants of dietary niche (Blyton et al., 2019; Greene et al., 2020; Heys et al., 2021; Kohl et al., 2014; Moeller & Hybridization is increasingly recognized as an important compo- Sanders, 2020), either by provisioning hosts with key nutrients (Hu nent of ecological and evolutionary processes. Consequences of et al., 2018; Jing et al., 2020; Ju et al., 2020) or by detoxifying de- hybridization span the fitness spectrum ranging from infertility and fensive compounds found in host food sources (Zheng et al., 2016). death (Brucker & Bordenstein, 2013; Zhang et al., 2014) to innova- Beyond diet and metabolism, HA microbiomes influence a range of tion and adaptation (Abbott et al., 2013; Dowling & Secor, 1997; other host traits as well (Archie & Theis, 2011; Bravo et al., 2011; Patton et al., 2020; Seehausen, 2004). Ultimately, these fitness Davidson et al., 2018; Ezenwa et al., 2012; Gaona et al., 2016; consequences dictate the role that hybridization plays in the suc- Grinberg et al., 2022; Jia et al., 2021; Kirchoff et al., 2019; Neufeld cess or failure of different genetic lineages (Seehausen, 2004; Pala et al., 2011; Sampson & Mazmanian, 2015; Sharon et al., 2010). & Coelho, 2005; Larouche et al., 2020; Todesco et al., 2016). If, for Healthy gut (Chen et al., 2018; Kamada et al., 2013), skin (Chen example, hybridization produces sterile offspring, then it can drive et al., 2018; Harris et al., 2006; Kueneman et al., 2014) and vagi- the emergence of genetic sinks and evolutionary dead ends (Tripp & nal microbiomes (Brotman et al., 2010), for example, provide patho- Manos, 2008) and thus serve as a brake for evolution. Alternatively, gen resistance across a broad spectrum of animal species (Buffie & if hybridization facilitates ecological release and/or sexual isola- Pamer, 2013; Ubeda et al., 2017; Woodhams et al., 2016). Indeed, tion (either directly through mating barriers or indirectly through amphibian skin microbiomes have been extensively studied as altered temporal or spatial proximity), then it can promote lineage a means of defending hosts from devastating fungal pathogen diversification and thus serve as a motor for evolution (Heard & (Batrachochytrium dendrobatidis and B. salamandrivorans) epidemics Hauser, 1995). (Bates et al., 2018, 2022; Rebollar et al., 2016, 2020). In humans, Most early research on hybrid organisms focused on under- disruptions to healthy HA microbiomes also underly a range of non- standing how hybridization impacts host fitness through effects infectious diseases (Ahn et al., 2013; Zackular et al., 2013) such as on host traits, for example, fecundity (Campbell et al., 2006; rheumatoid arthritis (Bergot et al., 2019; Scher & Abramson, 2011) Dobzhansky, 1934; Forejt, 1996; Hovick & Whitney, 2014; Reed and irritable bowel syndrome (Chong et al., 2019; Pimentel & & Sites Jr, 1995), physiology (Brown & Bouton, 1993; Cooper & Lembo, 2020). Ultimately, the cascading effects of HA microbiomes Shaffer, 2021; Lafarga-De la Cruz et al., 2013; Martins et al., 2019; on host traits and processesranging from host energy balance and Pereira et al., 2014), morphology (Capblancq et al., 2020; Carreira dietary niche through disease risk and immune dysfunctionhave et al., 2008; Jackson, 1973; Mrot et al., 2020) and behaviour (Robbins strong consequences on host ecological success (Abbott et al., 2021) et al., 2010, 2014). Recently, however, there has been growing recog- and, by extension, host evolution (Kolodny et al., 2020; Opstal, & nition that macroorganisms are not autonomous units. Rather, they Bordenstein, 2015; Zilber-Rosenberg & Rosenberg, 2008). are collectives or holobionts compromised of both a host and all of Although there has been substantial literature document- its host-associated (HA) microbes (Baedke et al., 2020; Bordenstein ing both coevolutionary (Ehrlich & Raven, 1964; Janz, 2011; & Theis, 2015; Bosch & Miller, 2016; Margulis & Fester, 1991). Thus, Janzen, 1980; Thompson, 1994, 2005) processes and codiversi- just as it is important to understand how hybridization impacts fication patterns (Janz, 2011; Nishida & Ochman, 2021; Suzuki the traits of the host, it is equally important to understand how et al., 2022; Thompson, 1989) between hosts and their HA micro- hybridization impacts the traits of the holobiont, including charac- biomes (Apprill et al., 2020; Chiarello et al., 2018; Ley et al., 2008; teristics of the HA microbiome (Miller et al., 2021). Indeed, the eco- Meadows, 2022; Moran & Sloan, 2015; Ochman et al., 2010; Phillips evolutionary basis for holobionts has led to entirely new branches of et al., 2012; Sanders et al., 2014; Scheelings et al., 2020; Walker research in areas as diverse as human health (Postler & Ghosh, 2017; et al., 2019), the study of how HA microbiomes respond when di- Walter et al., 2013), conservation (Bahrndorff et al., 2016; Banerjee vergent host lineages reunite, or admix, through hybridization is et al., 2020; Carthey et al., 2020; Jimnez & Sommer, 2017; Jin Song relatively new (Malukiewicz et al., 2019). One of the earliest inves- et al., 2019; Maebe et al., 2021; Redford et al., 2012; Trevelline tigations into hybrid microbiomes was in Nasonia wasps (Brucker & et al., 2019; West et al., 2019; Zhu et al., 2021) and biotechnology Bordenstein, 2013). In this system, up to 90% lethality is observed in (Bredon et al., 2020; Ren et al., 2022), and it is currently poised to do F2 males of N. vitripennis/N. giraulti crosses. However, rearing wasps so within the field of hybridization research as well. under germ-free conditions results in near complete rescue of the The importance of the holobiont concept stems from the many same F2 males. This suggests a microbial basis to hybrid lethality. host traits and processes that are either partially or fully depen- Interestingly, the 10% of hybrid N. vitripennis/N. giraulti males that dent on host-associated microbes (Fontaine & Kohl, 2020; Friesen survive under natural conditions exhibit highly transgressive micro- et al., 2011; Nobs et al., 2019; Walters et al., 2020). As an example, bial phenotypes. This includes both the appearance of novel micro- gut microbiomes are strong regulators of host metabolic phenotype bial taxa in hybrid microbiomes as well as shifts in the abundances of (Claus et al., 2008; Li et al., 2008; Mayneris-Perxachs et al., 2016). microbial taxa that are shared among parents and hybrids. This, in turn, impacts host energy balance (Corbin et al., 2020; More recent studies on hybrid vertebrates paint a similar picture. Nieuwdorp et al., 2014; Xifra et al., 2019), including both energy For example, hybrid house mice (Mus musculus musculus and Mus m. intake as well as use and expenditure. Gut microbiomes can also domesticus) in central Europe (Wang et al., 2015) exhibit widespread 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 512 transgressive microbiomes. Furthermore, like the Nasonia wasp 513 host genetics and the environment impact HA microbiomes across system, there is evidence that the altered microbial pheno- a Neotoma woodrat hybrid zone, Nielsen et al. (2023) demonstrated types of hybrid individuals at least partially explain their poor fit- that HA microbial composition was predominately driven by host ness outcomes (Baird et al., 2012; Britton-Davidian et al., 2005; genetics (genotypic classes), while HA microbial richness was pre- Forejt & Ivnyi, 1974; Good et al., 2008; Sage et al., 1986; Turner dominately driven by the environment (core diet + vegetation com- et al., 2012). In particular, there is an interaction between inflamma- munities). Applying similar approaches to other hybrid systems may tion, immune gene expression and the gut microbiome that appears be a fruitful avenue for disentangling the long-standing nature ver- to cause hybrid mice to exhibit defects in immunoregulation. This sus nurture paradigm as it applies to HA microbiomes and HA micro- may be one reason why hybrid individuals are restricted to a nar- biome assembly. row tension zone where the two parent subspecies co-occur (Balard Despite the increasing recognition that HA microbiomes are & Heitlinger, 2022; Barton & Hewitt, 1985). A range of additional an important facet of hybridization and that hybrid organisms are studies, including hybridization of sika deer (Cervus nippon) and elk valuable systems for understanding HA microbiome structure and (Cervus elaphus) (Li et al., 2016), lake whitefish lineages (Coregonus function, there is a lack of frameworks for describing and compar- clupeaformis) (Sevellec et al., 2019), blunt snout bream (Megalobrama ing hybrid HA microbiomes across the tree of life. In this paper, we amblycephala) and topmouth culter (Culter alburnus) (Li et al., 2018) develop four conceptual models delineating potential relationships and desert (Neotoma lepida) and Bryant's (Neotoma bryanti) woodrats between hybrid microbiomes and the microbiomes of their progen- (Nielsen et al., 2023) have reiterated the finding that hybrid animals itors. We discuss the underlying implications of each model, how often exhibit altered microbiomes relative to their progenitors (i.e. each model might arise based on fundamental host mechanisms and parent lineages or parent taxa). Indeed, even beyond the animal how each model could impact host fitness. We then integrate these kingdom, hybrid macroorganisms are commonly associated with four models into a quantitative 4H index that can be used to assess perturbations to the HA microbiome (Cregger et al., 2018; O'Brien the relative importance of each model across widely disparate hybrid et al., 2019; Wagner et al., 2020). systems. Finally, we introduce an R package, HybridMicrobiomes As suggested above, the study of hybridization and its impact (https://c ran.r-p rojec t.o rg/w eb/p ackag es/H ybrid Micro biome s/ on HA microbiomes is important for understanding host fitness and index.html), containing a series of functions that allow researchers evolution (Baeckens, 2019; Muoz & Bodensteiner, 2019). However, to apply the 4H index to their own hybrid microbiome data sets. even beyond host success, hybrid systems are of interest because they facilitate an understanding of genotypephenotype interactions (Kearney, 2005; Kratochwil & Meyer, 2015). Many hybrid zones (Cooper & Shaffer, 2021; Robbins et al., 2010; Walls, 2009), especially systems where F2 individuals readily admix with their 2 | M ATE R I A L S A N D M E TH O DS 2.1 | Conceptual models progenitors, provide variable genetic combinations (Lee et al., 2017; Pfennig, 2021) and degrees of heterozygosity across hybrid individ- We propose four conceptual modelsthe Union Model, the uals. Consequently, these systems serve as natural laboratories for Intersection Model, the Gain Model and the Loss Modelto describe understanding how host genetics and environmental characteris- the potential relationships between the HA microbiomes of hybrid tics influence host traits. For example, in a study investigating how individuals and those of their progenitors (see Figure 1). These four F I G U R E 1 The limiting scenarios for each of our four conceptual models describing the host-associated microbiomes of hybrid organisms. Hybrid organisms can (a) host all of the microbial taxa found on either progenitor (Union), (b) host only those microbial taxa found on both progenitors (Intersection), (c) host only novel microbial taxa found on neither progenitor (Gain) or (d) be missing all microbial taxa found on one or both progenitors (Loss). Note that, in the final scenario, assuming that there is no gain of microbial taxa, the hybrid has no microbiome at all. 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | CAMPER et al. | CAMPER et al. models represent extreme or limiting scenarios, each portraying an invasion (McCann, 2000). If this is true for HA microbial communi- idealized relationship between hybrid and progenitor HA microbi- ties, then hybrids following the Union Model may gain the advantage omes. Realistic hybrid microbiomes can then be described as differ- of having a more resilient microbiome that is more resistant to colo- ing combinations of these four idealized models. In what follows, we nization by pathogens (Harrison et al., 2019). delineate the four models, discuss possible underlying mechanisms However, there are likely costs to the Union Model as well. and outline their potential for impacting hybridization outcomes. For Most obvious are the challenges of bringing together large num- the sake of simplicity, we introduce each of the models within the bers of distinct microbial taxa from different progenitors. Consider framework of microbiome membership (taxon incidence) rather than BatesonDobzhanskyMuller (BDM) incompatibilities (Muller, 1942; composition (taxon abundance). However, comparable arguments Orr, 1995; Orr & Turelli, 2001), which emerge in hybrid organisms can be made for abundance relationships between progenitor and due to mismatches between the genes from their two progenitors. hybrid microbiomes as well. If BDM incompatibilities are a common outcome of combining different progenitor genomes, then analogous mismatches that result from combining different microbial metagenomes should also 2.1.1 | The Union Model be possible. Furthermore, there could be mismatches between the microbial metagenome from one progenitor and the host genome In its most extreme form, this model implies that hybrid microbiomes from the other, as outlined in the microbial-assisted BDM model are comprised of all microbial taxa present on at least one progenitor by Brucker and Bordenstein (2012). Indeed, if BDM incompatibilities and nothing else (see Figure 1a). This could occur if carrying a particu- scale with genome size (a tenuous assumption), hostmicrobe and lar host genome fosters colonization by associated microbial taxa. microbemicrobe incompatibilities may be more likely than tradi- Notably, such fostering could emerge either directly through host tional BDM incompatibilities simply because the microbial metag- interactions with the microbe (e.g. if specific hybrid and/or progeni- enome is typically much larger than the genome of the host itself. tor morphologies provide housing for symbiotic microbes (Belcaid Whether or not this is the case, the Mus musculus and Nasonia sys- et al., 2019; Delaux & Schornack, 2021; Fronk & Sachs, 2022)) or tems suggest that microbial-assisted BDMs are certainly a possibility indirectly through effects on host behaviour or ecology (e.g. if hy- among hybrid HA microbiome systems. brids colonize a progenitor's environment and subsequently acquire environmental microbes). To the extent that hybrid individuals share genetic material from both progenitors (note that this may vary de- 2.1.2 | The Intersection Model pending on the extent of back-crossing), hybrids should support all microbes present on either progenitor. Said differently, in the Union In its most extreme form, this model suggests that hybrid microbi- Model, the host genome acts as a ticket for acquiring a particu- omes are comprised of all microbial taxa simultaneously present on lar microbiome. Having two tickets (i.e. each representing a unique both progenitors and nothing else (see Figure 1b). Note that, within genomic component) results in the acquisition of two microbiomes, the framework of our conceptual models, the Intersection Model is one from each progenitor. a subset of the Union Model (i.e. the Union Model comprises all mi- Hybrids characterized exclusively by the Union Model (see crobial taxa found on one or both progenitors, while the Intersection Figure 1a) should have more taxonomically diverse microbiomes than Model comprises only those microbial taxa found on both progeni- either progenitor. Importantly, greater taxonomic diversity could re- tors). We describe our conceptual models in this way because it best sult in greater functional diversity as well (Petchey & Gaston, 2002), reflects potential mechanisms by which Union and Intersection of with important consequences for host health and ecological perfor- HA microbiomes may emerge. However, a slightly different defini- mance. Consider a thought experiment wherein two different insect tion of the Union model (not including microbial taxa found on both species are each limited to a distinct set of host plants based on the progenitors) is used for the 4H index. This is done to avoid dou- need for gut microbial detoxification of plant defensive compounds. ble counting components of microbial diversity (see Section 2.2). If the hybrid offspring of these two insect species harbour the gut Further note that, unlike the Union Model, a hybrid cannot be exclu- microbiomes of both progenitors, then hybrid microbiomes should sively characterized by the Intersection Model unless there are no be able to detoxify both sets of host plants, allowing hybrids to uti- microbial taxa unique to one of the two progenitors. This is because lize all resources open to either progenitor. More broadly, greater a hybrid that only harbours microbial taxa found on both progenitors functional capacity of hybrid microbiomes could enable hybrids must, in addition, have lost all microbial taxa found on only one of to persist in habitats that are intermediate to their progenitors or the two progenitors. The inter-relatedness of the Intersection and across all habitats colonized by either progenitor. Beyond expanded Loss models is discussed more below. function, a more diverse hybrid microbiome may have other bene- The Intersection Model could occur if a particular host genome fits as well. Although contentious, both in microbiome (Deng, 2012; hinders or prevents colonization by unassociated microbial taxa. He et al., 2013; Wagg et al., 2018) and general ecology literature, Again, the underlying mechanism could be direct (e.g. changes in the diversity (Ives & Carpenter, 2007; McCann, 2000) has long been as- host immune system) or indirect (e.g. changes in host behaviour that sociated with lower temporal variability and increased resistance to alter exposure to environmental microbes). In either case, hybrids 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 514 515 that carry genetic material from both progenitors will be more re- potentially different detoxification properties as compared to their fractory to, or isolated from, a wider range of microbial taxa. Said progenitors. Thus, rather than being able to use both of their pro- differently, in the Intersection Model, each host genome acts as a genitor's host plants (Union Model) or neither of their progenitor's gate. Having two gates blocks a wider range of microbes, leaving host plants (Intersection Model), these saltational hybrids could only those taxa that are permitted access by both progenitors. Thus, potentially colonize an entirely novel set of host plants not used by hybrids characterized by the most extreme form of the Intersection either progenitor. Model (see Figure 1b) should have less taxonomically diverse micro- More so than the Union, Intersection or Loss models, the Gain biomes which could have consequences for functional diversity as Model is responsible for phenotypic novelty and thus, provides well. For instance, in our previous insect example (see The Union the building blocks for evolutionary innovation. This could result in Model), the Intersection Model could leave hybrids without the rapid adaptation, escape from competition with their progenitors or ability to detoxify either set of progenitor host plants, placing a sub- even reproductive isolation. Indeed, in some cases, the Gain Model stantial limitation on feeding opportunities. This, in turn, could have may actually accelerate speciation (Mallet, 2007). However, the impacts on fitness, leading to higher rates of starvation, underper- Gain Model may have non- or maladaptive consequences as well. formance due to toxin build-up or even poisoning directly. Similar Notably, there is no a priori reason to believe that the acquisition negative effects on survival could be possible due to more general of large numbers of novel microbial taxa will be generally beneficial mechanisms associated with microbial diversity as well, for example, to a host. In fact, there are many reasons to believe the opposite. the loss of microbiome stability and pathogen resistance. The bene- In particular, the Gain Model describes a scenario of rapid evolu- fit of the Intersection Model, of course, is that it virtually eliminates tionary change (i.e. the introduction of novel microbial metagenomic opportunities for microbehost or microbemicrobe incompatibili- content to the hybrid) that occurs far outside the confines of more ties. This is because, in the Intersection Model, all microbehost and typical hostmicrobe coevolutionary relationships forged over gen- microbemicrobe interactions that occur on the hybrid are already erations of symbiosis. As a result, the Gain Model exemplifies a high present on both progenitors. risk, high reward scenario, and novel microbes acquired by the hybrid could just as easily enhance or reduce host fitness. Thus, like Goldschmidt's hopeful monsters, the Gain Model relies on happy 2.1.3 | The Gain Model accidents (Ross, 1983-1994) meaning that many hybrid individuals are likely to fail for each ecological success. In its most extreme form, this model suggests that hybrid microbiomes only include microbial taxa not present on either progenitor (see Figure 1c). Like the Intersection Model, a hybrid cannot be 2.1.4 | The Loss Model exclusively characterized by the Gain Model unless the progenitor microbiomes are fully devoid of microbial taxa. Again, this is because In its most extreme form, this model suggests that hybrid microbi- a hybrid that only harbours novel microbial taxa must, in addition, omes are missing all microbial taxa that are present on one or both have lost all microbial taxa found on the two progenitors. The Gain progenitors and have gained no new microbial taxa (see Figure 1d). Model is possible if HA microbiomes are idiosyncratically sensitive In other words, the most extreme form of the Loss Model implies to specific gene combinations that arise from merging progenitor that hybrids have no microbiome at all. Again, this is unrealistic. genomes. Broadly speaking, the Gain Model is the microbial equiva- Thus, just as the Intersection and the Gain models cannot occur re- lent of Bateson's saltational evolution (Bateson, 1984, 2002) or alistically independent of the Loss Model, nor can the Loss Model Goldschmidt's hopeful monsters (Goldschmidt, 1933, 1940). Like occur realistically independent of at least one or more of the other Bateson's and Goldshmidt's models, the Gain Model posits that models. The non-independence of the various models reflects the hybridization can yield profound (saltational) changes in pheno- fact that, except in very special and typically non-realistic scenarios type (Theien, 2006, 2009), and that these phenotypic changes (e.g. when the progenitors or hybrids have no HA microbes), real- may enable hybrids to establish an entirely novel ecological niche ized systems will always be combinations of the idealized models. relative to their progenitors (Dittrich-Reed & Fitzpatrick, 2013; The idealized models, however, serve as limits that emerge out of Goldschmidt, 1933; Mallet, 2007). However, unlike Bateson and various scenarios by which host genomes, and in particular hybrid Goldschmidt, who focused on host genes, the Gain Model assumes host genomes, could feasibly impact HA microbiome assembly. As that there are underlying microbial dimensions to the saltational in the Gain Model, the loss of microbes present on both progenitors change. Arguably, adding microbial dimensions provides even more describes a saltational scenario that is possible if HA microbiomes opportunity for saltational change, again because of the vast size are idiosyncratically sensitive to gene combinations of the progeni- and diversity of functions encompassed by the microbial metage- tors. In contrast to the Gain Model, however, the saltational change nome relative to the host genome itself. Once more, consider our invoked by the Loss Model is the deletion, rather than the addition, hypothetical insect example (see The Union Model). Hybrid insects of microbial taxa. characterized by the most extreme form of the Gain Model should The Loss Model gives rise to hybrid microbiomes with lower harbour an entirely new set of gut bacteria with novel taxa and overall diversity and potentially lower functional capacity as well. 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | CAMPER et al. | CAMPER et al. Indeed, as suggested above, in the limit of a hybrid organism ex- like to consider abundance thresholds have the option to do so. clusively characterized by the Loss Model, there would be no host- Thus, both the 4H index and the HybridMicrobiomes R package pro- associated microbiome at all. Returning, for the last time, to our vide flexibility that can be decided within the context of a particular hypothetical insect complex (see The Union Model), the Loss Model hybrid system (though a common set of parameters should be used predicts that hybrid insects should lack microbes present on one or for any comparison between hybrid systems). both progenitors. In the case of the latter, hybrids would lose the For the 4H index, , and are the same across all host classes. ability to detoxify host plants that are usable by both of their progen- However, each host class (i.e. each progenitor and the hybrid) is sep- itors. Like the Intersection Model, this could limit opportunities for arately assigned its own core microbiome. Thus, a microbial taxon feeding, cause toxin build-up or result in poisoning of hybrid insects. is part of a host's core microbiome provided it is found on at least Lower microbiome diversity could also lead to a suite of additional N hosts of that host class at a minimum average abundance of challenges like greater microbiome instability and lower pathogen and a minimum abundance on at least one host of , where N is the resistance. Again, however, the costs of low diversity microbiomes number of hosts of each class and should be the same across all host may be balanced out by the benefits of reduced opportunities for classes (i.e. a balanced design with equal numbers of each progenitor hostmicrobe or microbemicrobe incompatibilities. and the hybrid; note that the HybridMicrobiomes package includes bootstrapping steps that will downsample data sets such that a bal- 2.2 | The 4H index and quaternary plots anced design is achieved). In what follows, we describe four versions of the 4H index, two based on incidence of microbial taxa and two based on abundance of microbial taxa. While the four conceptual models in Section 2.1 present limiting, extreme or idealized scenarios, any realistic hybrid system will almost certainly exhibit mixed support across two or more conceptual 2.2.1 | Incidence-based analyses models. To examine the importance of each of the four conceptual models to any given hybrid system, we introduce the 4H index, along Our two incidence-based methods are inspired by the Jaccard index with R package HybridMicrobiomes (https://cran.r-projec t.org/ (Jaccard, 1908) and the Sorensen index, respectively (Dice, 1945; web/packages/HybridMicrobiomes/index.html), which can be used Sorensen, 1948). For any given , and , we define P1, P2 and H as to calculate and graph the 4H index for any hybrid system. The 4H the set of core microbial taxa present on the first progenitor, the index uses the core microbiomes of each host class (where we use second progenitor and hybrids. We then determine the number of host class to refer to any one of the three types of hoststhe first microbial taxa shared by different combinations of hybrid and pro- progenitor, the second progenitor or the hybridin a hybrid com- genitor classes. Specifically, we define: plex) to determine which microbial taxa are lost and gained on hybrid organisms relative to their progenitors. To define the core microbiome, we use a tunable parameter, , which can range from = 1 (i.e. microbial taxa are only considered if they are present on every host of a particular host class) to = 0 (i.e. the full microbiome; all microbial taxa are considered regardless of the number of hosts they are found on). Consistent with the common definition of a core microbiome, we typically select higher values of . This is based on the assumption that microbial taxa with strong consequences for host ecology and/or evolution should be detectable on the majority of hosts within a population. However, researchers who have reason to suspect otherwise can use a lower value of or can compare the 4H index across a range of values (see Figures S2.1S2.4; Tables S2.1 and S2.2). While core microbial taxa are usually defined as those present on a threshold number of hosts, alternate definitions exist that incorporate microbial abundances as well (Shade & Stopnisek, 2019). To ) | |( a = | P1 P2 H|, | | ) | |( b = | P1 P2 H| a, | | (1a) (1b) b1 = ||P1 H|| a, b2 = ||P2 H|| a, c = |H| a b, (1c) d = ||P1 P2 H|| a b c, (1d) d1 = ||P1 || ||P1 P2 || ||P1 H|| + a, allow for this, we include a second threshold, , based on the average d2 = ||P2 || ||P1 P2 || ||P2 H|| + a, relative abundance (across all hosts within a class) that a microbial d12 = d d1 d2 . taxon must reach to be considered part of the core. In addition, we include a third threshold, , based on the minimum relative abundance that a microbial taxon must reach on at least one host to be In Equation (1), |S| denotes the cardinality of set S, where S is any set. Accordingly, a is the number of microbial taxa shared by both progen- considered part of the core. By default, the HybridMicrobiomes R itors and the hybrid, b is the number of microbial taxa shared by one package sets both = 0 and = 0. However, researchers who would progenitor (but not both) and the hybrid, b1 is the number of microbial 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 516 taxa shared by the first progenitor and the hybrid, b2 is the number of c , 3a + 2b + c + d1 + d2 + 2d12 (3c) d1 + d2 + 2d12 = 1 . 3a + 2b + c + d1 + d2 + 2d12 (3d) = microbial taxa shared by the second progenitor and the hybrid, c is the number of microbial taxa found only on the hybrid, d is the number of microbial taxa found only on one or both progenitors, d1 is the number of microbial taxa found only on the first progenitor, d2 is the number of = 517 microbial taxa found only on the second progenitor and d12 is the num- Notice that the difference between the Jaccard-and Sorensen-inspired ber of microbial taxa found only on both progenitors. For the Jaccard- methods is whether taxa found on multiple host classes are weighted inspired method, we define the four dimensions of the 4H index (three according to the number of host classes that they occur on. Similar independent dimensions) as: to the Jaccard index for beta diversity, the Jaccard-inspired 4H index only counts the unique microbial taxa shared by each combination of = b , a+b+c+d (2a) host classes. By contrast, like the Sorensen index for beta diversity, the Sorensen-inspired 4H index triple weights microbial taxa shared by all three host classes and double weights microbial taxa shared by two of b1 1 = , a+b+c+d the three host classes. Function FourHbootstrap in HybridMicrobiomes takes a phyloseq object (McMurdie & Holmes, 2013) and a vector specifying 2 = b2 , a+b+c+d = a , a+b+c+d (2b) c , a+b+c+d (2c) d = 1 , a+b+c+d (2d) progenitor and hybrid classifications. It then calculates the 4H index over bootstrapped samples of hybrid organisms and their progenitors. FourHbootstrap outputs a data frame with the percentage of microbial taxa that fall into each of the four models (see Equation 1). The data frame also includes the fraction of progenitor microbial taxa that are found on both progenitors. Finally, the data frame breaks = = the Union Model into separate components attributable to the first progenitor and the second progenitor, respectively ( 1 + 2 = ). 2.2.2 | Abundance-based analyses where , , and reflect the extent of Union, Intersection, Gain and Loss models, respectively. Briefly, is the fraction of microbial Like incidence, we include two different abundance-based methods taxa found on hybrids and on one (but not both) progenitor (note that for calculating the 4H index, with the first inspired by the Ruzicka this is a slight deviation from the conceptual Union Model, which does index (Legendre, 2014) and the second inspired by the BrayCurtis not distinguish between taxa found on one or both progenitors. This index (Bray & Curtis, 1957). Similar to incidence-based methods, the deviation is necessary to avoid double counting microbial taxa in the two abundance-based methods also focus on core microbial taxa as 4H index. Also note that can be divided into a component that the defined by , and . However, abundance-based methods require hybrid shares with the first progenitor, 1, and a component that the an additional pre-step to find representative microbial abundances hybrid shares with the second progenitor, 2). is the fraction of mi- for each host class. This step is performed on the full microbiome crobial taxa found on hybrids and on both progenitors, is the fraction (i.e. core and non-core microbial taxa) based on either the mean or of microbial taxa only found on hybrids, and is the fraction of micro- median relative abundance of each microbial taxon on each host bial taxa only found on progenitors. class. These representative abundances can then be used raw or can Similarly, for the Sorensen-inspired method, we define the four dimensions of the 4H index (three independent dimensions) as: be renormalized based only on microbial taxa that comprise the core of each host class. Renormalization results in a metric that is density invariant (i.e. does not vary with the number of reads attributed to = 2b , 3a + 2b + c + d1 + d2 + 2d12 (3a) the core microbiome of each host class) (Jost et al., 2011). However, a downside of renormalization is that it constrains the 4H index to a two-dimensional plane (the same is true when using = 0 since the 2b1 1 = , 3a + 2b + c + d1 + d2 + 2d12 FourHbootstrapA function rarefies microbiome data sets, thereby forcing all hybrid classes to have equivalent numbers of reads). By contrast, using raw reads allows the total number of reads to differ 2b2 2 = , 3a + 2b + c + d1 + d2 + 2d12 between host classes. While this results in a metric that is not den- 3a = , 3a + 2b + c + d1 + d2 + 2d12 we view raw reads as the preferred option. This is because full micro- sity invariant (i.e. it changes with the number of reads attributed to the core of a particular host class, see Supplemental Information S1), (3b) biomes are rarefied to the same number of reads prior to selection 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | CAMPER et al. | CAMPER et al. of the core. Thus, differences in read numbers attributed to the core C , A+B+C+D (5c) D = 1 . A+B+C+D (5d) = reflect biologically meaningful differences in composition. To generalize our incidence-based 4H index to account for abundance, we follow the method in Tams et al. (2001) and define: A= B= S { i=1 S i=1 = ( ( ) ) S min min xiP1 , xiP2 , xiH = i . i=1 (4a) Similarly, for the BrayCurtis-inspired method, we define the four dimensions of the 4H index (three independent dimensions) as: ( ) ( )} S { } min xiP1 i , xiH i + min xiP2 i , xiH i = 1i + 2i , i=1 (4b) B1 = S { ( )} S min xiP1 i , xiH i = 1i , B2 = S { ( )} S min xiP2 i , xiH i = 2i , i=1 i=1 C= D= i=1 = 2B , 3A + 2B + C + D1 + D2 + 2D12 1 = 2B1 , 3A + 2B + C + D1 + D2 + 2D12 2 = 2B2 , 3A + 2B + C + D1 + D2 + 2D12 = 3A , 3A + 2B + C + D1 + D2 + 2D12 (6b) = C , 3A + 2B + C + D1 + D2 + 2D12 (6c) D1 + D2 + 2D12 = 1 . 3A + 2B + C + D1 + D2 + 2D12 (6d) i=1 S { } xiH i 1i 2i , (4c) i=1 S { ( )} xiP1 + xiP2 i 1i 2i min xiP1 , xiP2 , i=1 D1 = S { D2 = S { i=1 ( )} ( )} xiP1 1i min xiP1 , xiP2 (4d) , = i=1 D12 = xiP2 2i min xiP1 , xiP2 (6a) , Again, the difference between the Ruzicka-and BrayCurtis-inspired methods is whether reads/fractions of reads found on multiple host S { ( ) } min xiP1 , xiP2 i , classes are weighted according to the number of host classes that i=1 they occur on. Similar to the Ruzicka index for beta diversity, the where xiP1 , xiP2 and xiH are the number/fraction of reads of microbial Ruzicka-inspired 4H index only counts shared microbial reads once taxon i on the first progenitor, the second progenitor and the hybrid re- regardless of the number of host classes that they occur on. By con- spectively, and S is the total number of microbial taxa in the system. A is trast, like the BrayCurtis index for beta diversity, the BrayCurtis- then the number/fraction of reads shared by both progenitors and the hy- inspired 4H index triple weights reads/fractions of reads shared by brid, B is the number/fraction of reads shared by one progenitor (but not all three host classes and double weights reads/fractions of reads both) and the hybrid, B1 is the number/fraction of reads shared by the first shared by two of the three host classes. Function FourHbootstrapA progenitor and the hybrid, B2 is the number/fraction of reads shared by in HybridMicrobiomes performs abundance-based bootstraps of the second progenitor and the hybrid, C is the number/fraction of reads the 4H index with input and output as described for the function found only on the hybrid, D is the number/fraction of reads found only on FourHbootstrap (see above). one or both progenitors, D1 is the number/fraction of reads found only on the first progenitor, D2 is the number/fraction of reads found only on the second progenitor and D12 is the number/fraction of reads found only on 2.2.3 | Bootstrap analysis both progenitors. For the Ruzicka-inspired method, we define the four dimensions of the 4H index (three independent dimensions) as: Function FourHcentroid takes the output from FourHbootstrap or FourHbootstrapA and calculates the centroid of the boot- = B , A+B+C+D (5a) FourHbootstrap or FourHbootstrapA on multiple hybrid systems and uses a PERMANOVA test (Anderson, 2014) on the isometric log- B1 1 = , A+B+C+D ratio transformed (Egozcue et al., 2003) 4H indices (with the option to use a centred log-ratio transformation, an additive log-ratio trans- B2 2 = , A+B+C+D = A , A+B+C+D strapped samples. Function FourHcompare takes the outputs from formation or untransformed data instead (Filzmoser et al., 2010; Quinn et al., 2019)) to determine whether different hybrid systems vary with respect to the importance of the Union, Intersection, Gain (5b) and Loss models, respectively. 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 518 2.2.4 | Quaternary plots 2.2.5 | Null planes 519 To visualize the 4H index (see Figures 2 and 3), which can be par- As suggested above (see The Loss Model), both our conceptual ticularly helpful for comparison between systems, we introduce models and the four dimensions of the 4H index conflate microbial a quaternary plotting technique (i.e. a four-d imensional barycen- taxon loss due to the intersection of progenitor microbiomes (i.e. tric plot or an Aitchison Simplex (Aitchison, 1982)). This positions loss of microbial taxa only present on one progenitor) with broader each of our four index dimensions (, , and ) at a vertex microbial taxon loss (i.e. including loss of microbial taxa present on of a triangular prism, with one edge of the prism connecting the both progenitors). Thus, the 4H index does not indicate whether the Gain and Loss models (henceforth termed the transgressive axis) microbial taxa that are lost versus retained by hybrid organisms rep- and the opposite edge connecting the Union and Intersection resent microbial taxa that are shared by both progenitors or taxa that models (henceforth termed the parental axis). Function are only found on one progenitor. Unfortunately, conflation of these FourHquaternary takes the output from FourHbootstrap or different types of loss is necessary to double-counting microbial FourHbootstrapA and generates an interactive and rotatable taxa ( + + + = 1) while still using a maximum of four (ben- quaternary plot of the bootstrapped samples with the option to eficial for visualization) dimensions. To offset this constraint, and include the centroid. Function FourHquaternarycentroid takes better identify the particular microbial taxa that are lost by hybrid the output from FourHbootstrap or FourHbootstrapA and gener- organisms, we develop null planes. Specifically, we assume a null ates an interactive quaternary plot of only the centroids over the model wherein all microbial taxa (or reads in the case of abundance- bootstrapped samples. based methods) present on progenitors are equally likely to be lost F I G U R E 2 Quaternary plots showing 500 bootstrapped genus-level microbial samples (small circles) and the bootstrap centroid (large circles) of the Jaccard-inspired 4H index for (a) gut microbiomes from hybrid Kikihia cicadas (black), Neotoma woodrats (brown) and Aspidoscelis neomexicanus whiptail lizards (green); (b, c) woodrat and lizard systems individually along with the system null planes; (d) leaf (green) and rhizosphere (brown) bacterial/archaeal (16S rRNA, light) and fungal (ITS, dark) microbiomes from B73 line Mo17 line maize hybrids; (e) B73 line Mo17 line maize hybrid leaf and rhizosphere bacterial/archaeal systems along with system null planes; (f) leaf bacterial/archaeal microbiomes from B73 line Mo17 line (red), B73 line CML103 line (green) and B73 line Mo18W line (blue) maize hybrids. For systems in (ac), bootstraps consisted of seven hybrid individuals and seven of each progenitor. For systems in (df), bootstraps consisted of 10 hybrid individuals and 10 of each progenitor. A microbial genus was defined as being part of the core microbiome if at least 50% of hosts from a particular class carried that microbial genus. 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | CAMPER et al. | CAMPER et al. F I G U R E 3 Quaternary plots showing 500 bootstrapped genus-level microbial samples (small circles) and the bootstrap centroid (large circles) of the BrayCurtis-inspired 4H index for (a) gut microbiomes from hybrid Kikihia cicadas (black), Neotoma woodrats (brown) and Aspidoscelis neomexicanus whiptail lizards (green); (b, c) woodrat and lizard systems individually along with the system null planes; (d) leaf (green) and rhizosphere (brown) bacterial/archaeal (16S rRNA, light) and fungal (ITS, dark) microbiomes from B73 line Mo17 line maize hybrids; (e) B73 line Mo17 line maize hybrid leaf and rhizosphere bacterial/archaeal systems along with system null planes; (f) leaf bacterial/archaeal microbiomes from B73 line Mo17 line (red), B73 line CML103 line (green) and B73 line Mo18W line (blue) maize hybrids. For systems in (ac), bootstraps consisted of seven hybrid individuals and seven of each progenitor. For systems in (df), bootstraps consisted of 10 hybrid individuals and 10 of each progenitor. A microbial genus was defined as being part of the core microbiome if at least 50% of hosts from a particular class carried that microbial genus. by hybrids. This allows us to define a plane bisecting the quaternary concentrated among microbial taxa only found on one of the two plot at the expected fraction of hybrid microbial taxa that should be progenitors as in the Intersection Model). 4H indices that lie more shared with one versus both progenitors, assuming that there is no towards the vertex relative to the null plane suggest that hybrids preferential loss of one over the other. For any value of = + are disproportionately likely to retain microbes only found on one (i.e. the summed fractions of microbial taxa following the Gain and of the two progenitors (i.e. loss is concentrated among microbial Loss models), the null plane is given by: taxa found on both progenitors and is saltational). By using the null where = = a + d12 a + d12 + d1 + d2 2(a + d12 ) 2(a + d12 ) + d1 + d2 null = (1 )(1 ), (7a) null = (1 ), (7b) for the Jaccard-inspired for the Sorensen-inspired 4H index, = for the Ruzicka-inspired 4H index and = 4H plane as a reference, it is possible to assess the degree to which the loss occurs due to the intersection of progenitor microbiomes versus the broader saltational loss of microbes present on both progenitors. Function FourHnullplane takes the output from index, FourHbootstrap or FourHbootstrapA and graphs the (average) A + D12 A + D12 + D1 + D2 null plane for a particular hybrid system onto a quaternary plot. for the Function FourHplaneD takes the output from FourHbootstrap or BrayCurtis-inspired 4H index. In general, is the fraction of pa- FourHbootstrapA and reports both the average distance between 2(A + D12 ) 2(A + D12 ) + D1 + D2 rental microbial taxa that are found on both progenitors. null and the expected, null, and observed, , value of the intersection di- null are thus the expected fractions of microbial taxa that should mension, as well as the fraction, p, of bootstrap samples that lie be found on only one parent versus both parents under null model further from the vertex than expected (this is useful for testing assumptions. 4H indices that lie more towards the vertex rela- the hypothesis that microbes shared by both progenitors are more tive to the null plane indicate that hybrids are disproportionately likely to be retained by the hybrid than microbes only found on one likely to retain microbes shared by both progenitors (i.e. loss is progenitor). 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 520 3 | R E S U LT S 521 across all systems. Hybrid woodrats, for example, are dominated by the Intersection Model ( = 0.4927), while this is less important for To illustrate the usefulness of the 4H index, we apply functions hybrid cicadas ( = 0.1337) and hybrid lizards ( = 0.1160). Instead, from our HybridMicrobiomes package to a range of plant and animal lizards and cicadas feature a mix of the Gain and Loss models, with hybrid microbiota data sets including results from both 16S rRNA Gain being more important for lizards ( = 0.3438) and Loss being gene and ITS sequencing (see Supplemental Information S2 for ad- more important for cicadas ( = 0.3758). Figure 2b,c illustrates null ditional system pre-analysis steps and Supplemental Information S3 planes for the woodrat and lizard systems. From the null planes and for more information on the organismal systems). These analyses Table 2, we see that the two progenitor woodrat species share a demonstrate how any given hybrid system shows mixed support larger fraction of their core microbes as compared to the two pro- for each of our conceptual models and how the degree of support genitor lizard species. Furthermore, we see that hybrid woodrats are for any particular conceptual model varies from one system to an- biased towards the Intersection Model as compared to the null plane other. First, we consider applying the Jaccard-inspired 4H index ( null = 0.1358, p = 0.002). This means that hybrid woodrats are (incidence-based) to our data sets (see Figure 2; Tables 1 and 2; more likely to retain microbial taxa that are shared by both progeni- Supplemental Video files Figures 2a.mp42f.mp4). Figure 2a shows tors than they are to retain microbial taxa found on only one of the quaternary plots (R version 4.2.1, phyloseq 1.41.1) comparing gut two progenitors. By contrast, hybrid lizards do not show any obvious microbiota from F1 crosses of Neotoma woodrats (brown) (Nielsen bias, and thus, are equally likely to retain (or lose) microbial taxa that et al., 2023), Kikihia cicadas with evidence of mitochondrial intro- are shared by both progenitors or only present on one of the two gression (black) (Haji et al., 2022) and a parthenogenetic Aspidoscelis progenitors. lizard of hybrid origin (green, our own data). Table 1 shows the av- Figure 2d shows quaternary plots of both the phyllosphere erage values of the 4H indices for each system. Table 2 shows the (green) and rhizosphere (brown) of hybrid maize (B73 line Mo17 average values of (the fraction of the overall parental microbiome line) for both bacterial/archaeal (16S rRNA gene, light shade) and found on both progenitors), the mean distance between the pre- fungal (ITS1 gene, dark shade) microbiotas (Wagner et al., 2020). dicted and observed value of and the proportion of bootstrap Figure 2e shows the same bacterial/archaeal microbiotas but in- samples that lie further from the vertex than expected based on cludes their respective null planes, and Figure 2f compares the the null model. Despite the variation in life history (vertebrate vs. in- bacterial/archaeal phyllosphere microbiotas across three different vertebrate, ectotherm vs. endotherm, herbivore vs. insectivore) and maize hybrids: B73 line Mo17 line (red; stiff stalk crossed with mode of hybridization (F1 crosses, mitochondrial introgression, hy- non-stiff stalk varieties (Wagner et al., 2020)), B73 line CML103 brid speciation/parthenogenesis), the 4H index enables comparison line (yellow; temperate crossed with tropical varieties (Woodhouse TA B L E 1 Centroid values of the Jaccard-inspired 4H index as calculated by the FourHcentroid function. Parental axis Neotoma woodrat Transgressive axis + + 0.07296 0.4927 0.5657 0.0811 0.3532 0.4343 Aspidoscelis lizard 0.2941 0.1160 0.4101 0.3438 0.2461 0.5899 Kikihia cicada 0.1999 0.1337 0.3336 0.2906 0.3758 0.6664 Maize B73 Mo17 (leaf, bacteria) 0.1272 0.4949 0.6221 0.0575 0.3205 0.3780 Maize B73 Mo17 (rhizosphere, bacteria) 0.1892 0.5000 0.6892 0.1624 0.1484 0.3108 Maize B73 Mo17 (leaf, fungi) 0.1029 0.5238 0.6267 0.0902 0.2832 0.3734 Maize B73 Mo17 (rhizosphere, fungi) 0.1936 0.4226 0.6162 0.1524 0.2314 0.3838 Maize B73 CML103 (leaf, bacteria) 0.1070 0.5044 0.6114 0.0798 0.3087 0.3885 Maize B73 Mo18W (leaf, bacteria) 0.1601 0.4784 0.6385 0.1632 0.1983 0.3615 Maize B73 CML103 (rhizosphere, fungi) 0.2152 0.4468 0.6620 0.1123 0.2256 0.3379 Maize B73 Mo18W (rhizosphere, fungi) 0.1433 0.4635 0.6068 0.1146 0.2785 0.3931 Note: Summing the values + and + gives totals along the parental axis and the transgressive axis, respectively, and can be used as a broader scale comparison between systems. 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | CAMPER et al. | CAMPER et al. TA B L E 2 Fraction of shared microbial taxa among progenitors, , as calculated by the FourHcentroid function for the Jaccard- inspired 4H index. Displacement from the null plane ( null)a and the proportion of bootstrap samples (p) falling further from the vertex than the null plane are both calculated by the FourHnullplaneD function. Next, we consider applying the BrayCurtis-inspired 4H index (abundance-based) to our data sets (see Figure 3; Tables 3 and 4; Supplemental Video files Figures 3a.mp43f.mp4). Similar to the Jaccard-inspired 4H index (see Figure 2; Table 1), hybrid woodrats are dominated by the Intersection Model ( = 0.7745), while this is less important for hybrid cicadas ( = 0.1645) and hybrid lizards null p ( = 0.07921). Instead, lizards and cicadas feature a mix of the Gain Neotoma woodrat 0.6273 0.1358 0.002 and Loss models. For the BrayCurtis-inspired index, however, the Aspidoscelis lizard 0.2490 0.0110 0.364 Gain Model is almost equivalent for lizards ( = 0.1411) and cicadas Kikihia cicada 0.2827 0.0373 0.178 ( = 0.1420). Meanwhile, the Loss Model is slightly more import- Maize B73 Mo17 (leaf, bacteria) 0.5984 0.1208 0 ant for lizards ( = 0.5705) as compared to cicadas ( = 0.4993), Maize B73 Mo17 (rhizosphere, bacteria) 0.6291 0.0663 0 Maize B73 Mo17 (leaf, fungi) 0.6483 0.1167 0 Maize B73 Mo17 (rhizosphere, fungi) 0.5664 0.0742 0 Maize B73 CML103 (leaf, bacteria) 0.6475 0.1077 0 Maize B73 Mo18W (leaf, bacteria) 0.6121 0.0872 0 Maize B73 CML103 (rhizosphere, fungi) 0.5565 0.0781 0 Maize B73 Mo18W (rhizosphere, fungi) 0.6145 0.0910 0 which is the reverse of findings for the Jaccard-inspired 4H index. Subtle differences in results based on the chosen metric are consistent with the different interpretations of the metrics. In this case, a Positive values of null indicate that points lie closer to the intersection vertex than expected by chance, suggesting that hybrids are more likely to retain taxa shared by both progenitors than they are to retain taxa shared by only one of the two progenitors. Negative values indicate the opposite, namely that hybrids are more likely to retain taxa only found on one of the two progenitors than they are to retain taxa found on both progenitors. for instance, community membership differences suggest that lizard hybrids feature more Gain and less Loss, but that these differences are insignificant or reversed when considering abundance changes. Such discrepancies are expected when membership changes occur primarily in rare taxa and thus contribute little to abundance change, which may instead be dominated by shifts in abundance of microbes shared by hybrids and progenitors. From the null planes and Table 4, we see that, consistent with the Jaccard-inspired 4H index, the Bray Curtis-inspired 4H index suggests that the two progenitor woodrat species share a larger fraction of their core microbes as compared to the two progenitor lizard species. Furthermore, hybrid woodrats are biased towards retaining microbes shared by both progenitors ( null = 0.0476518, p = 0.006), whereas hybrid lizards do not preferentially retain microbes based on whether they are shared by one or both progenitors ( null = 0.030863, p = 0.764). Like our animal examples, abundance- and incidence-based 4H indices for maize hybrids exhibit a similar pattern. In particular, with our abundance-based analysis, we again find that the entire maize system is dominated by the Intersection Model. Indeed, like et al., 2021)) and B73 line Mo18W line (blue; flooding sensitive woodrats, the dominance of the Intersection Model is even more crossed with flooding insensitive varieties (Campbell et al., 2015)). apparent for the BrayCurtis-inspired 4H index than it is for the Tables 1 and 2 show the corresponding values of the 4H indices, as Jaccard-inspired 4H index with >60% of model support across all well as relationships of the hybrid microbiotas to their respective comparisons (see Table 3). However, differences between the hybrid null planes. As with our animal examples, our analysis of maize hy- rhizosphere (brown) and hybrid phyllosphere (green) are not as obvi- brids demonstrates the versatility of the 4H index and how the 4H ous and/or are reversed when changes in abundance are accounted index can be used to compare not only between microbiotas from for. Again, this suggests that microbiota membership changes on different host species but also between microbiotas from different the hybrid are sometimes but not always consistent with abundance parts of a single organism (roots vs. leaf) or different microbial tax- changes. onomic groups (bacteria vs. fungi). Notably, the entire maize sys- One of the benefits of the 4H index is the fact that it can be ap- tem is dominated by the Intersection Model as seen by the nearly plied to any hybrid system, regardless of the type of host, the type of 50% or more of model support across all comparisons (see Table 1). microbiome, microbiome composition or even microbiome diversity. However, the hybrid rhizosphere (brown) is more prone to Union and This flexibility follows from the fact that the 4H index is monotonic Gain. This is apparent from its relatively greater clustering nearest to with respect to each vertex/dimension (, , and ), option- the Union and Gain vertices, as well as relatively greater support for ally density invariant and replication invariant (see Supplemental these two models (see Figure 2d; Table 1; Supplemental Video files Information S1) (Magurran & McGill, 2010). Despite this, some Figure 2d.mp4). By contrast, the hybrid phyllosphere is more prone standardization of data sets from different systems is necessary to Loss. This is apparent from its relatively greater clustering near for fair comparison. For example, the 4H index can be applied to the Loss vertex and relatively greater support for the Loss Model microbiomes at any taxonomic scale. As expected, however, higher (see Figure 2d; Table 1; Supplemental Video files Figure 2d.mp4). taxonomic scales predict a greater importance of the Intersection 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 522 TA B L E 3 Centroid values of the Bray Curtis-inspired 4H index as calculated by the FourHcentroid function. Parental axis 523 Transgressive axis + + Neotoma woodrat 0.1011 0.7745 0.8756 0.02852 0.09587 0.12439 Aspidoscelis lizard 0.2092 0.07921 0.28841 0.1411 0.5705 0.7116 Kikihia cicada 0.1942 0.1645 0.3587 0.1420 0.4993 0.6413 Maize B73 Mo17 (leaf, bacteria) 0.2326 0.5790 0.8116 0.02496 0.1635 0.1884 Maize B73 Mo17 (rhizosphere, bacteria) 0.1199 0.7311 0.8510 0.02944 0.1196 0.1490 Maize B73 Mo17 (leaf, fungi) 0.08865 0.7709 0.8596 0.03248 0.1079 0.1404 Maize B73 Mo17 (rhizosphere, fungi) 0.1157 0.6092 0.7248 0.06660 0.2086 0.2752 Maize B73 CML103 (leaf, bacteria) 0.1094 0.7372 0.8467 0.03673 0.1166 0.1533 Maize B73 Mo18W (leaf, bacteria) 0.1611 0.6719 0.8330 0.03272 0.1343 0.1670 Maize B73 CML103 (rhizosphere, fungi) 0.1542 0.6257 0.7800 0.05275 0.1673 0.2200 Maize B73 Mo18W (rhizosphere, fungi) 0.1357 0.6090 0.7447 0.06074 0.1945 0.2553 Note: Summing the values + and + gives totals along the parental axis and the transgressive axis, respectively, and can be used as a broader scale comparison between systems. Model because hybrids are more likely to share distantly related FourHbootstrap and FourHbootstrapA do have the option to rarefy microbial taxa with progenitors than they are to share identical or microbiome samples to a lower read depth than the minimal number near-identical microbial taxa (see Figures S1.1S1.4; Tables S1.1 and of reads of the lowest sample. This allows for standardization of read S1.2). Importantly, because taxonomic scale can have considerable depth across systems. effects, systems should always be compared using the same taxonomic scale, and interpretation of the index should always be within the context of the taxonomic scale chosen. Likewise, defining the 4 | DISCUSSION core microbiome based on a lower fraction of hosts also favours Intersection (at least some hybrids and some of each parental spe- The advent of low-cost sequencing has greatly contributed to our cies are likely to have a particular microbial taxon, even if it is just a understanding of the importance of both hybridization and HA mi- transient acquisition from the environment; see Figures S2.1S2.4; crobiomes on host ecological traits and evolutionary consequences. Tables S2.1 and S2.2). Again, then, it is important to use the same These two fields come together in the study of HA microbiomes of value of for all systems that are being compared. Host sample size hybrid organismsa newly emerging area of research across disci- has a smaller, but still detectable, effect resulting in somewhat dif- plines ranging from agricultural science to ecology and conserva- ferent trends across systems but generally shifting the 4H index to- tion. In this paper, we integrate four conceptual models to develop wards the parental axis and away from the transgressive axis (see a framework for understanding the relationship between hybrid Figures S3.1S3.4; Tables S3.1 and S3.2). Although the effect of microbiomes and the microbiomes of their progenitors. We then host sample size is relatively small, particularly for larger host sam- use these models to develop a four-dimensional (three independ- ple sizes, it is still best to compare systems by subsampling to the ent dimensions) metricthe 4H indexto describe where a par- smallest number of hosts available for any host class across all sys- ticular hybrid complex falls among our four models. Our 4H index tems (e.g. see Figure 2 where we were limited to 7 individuals based borrows inspiration from beta diversity metrics, and thus takes on on the number of available cicada microbiotas). Finally, sequencing four different forms; two are incidence-based (Jaccard-inspired and depth has almost no impact on predictions, at least for >1000 reads Sorensen-inspired), and two are abundance-based (BrayCurtis- or more. This last feature of the 4H index is a benefit of focusing inspired and Ruzicka-inspired). Importantly, the 4H index facilitates on core microbiomes since low abundance microbial taxa that are comparisons across widely disparate systems, ultimately making it likely to be missed at low read depths are unlikely to be part of the possible to identify patterns that emerge across hybrid microbiomes core of any given species. For this reason, it is largely unnecessary to from different organisms. For example, the 4H index could be used standardize for read depth across systems. Nevertheless, functions to determine whether there are systematic differences between 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | CAMPER et al. | CAMPER et al. TA B L E 4 Fraction of shared microbial taxa among progenitors, , as calculated by the FourHcentroid function for the BrayCurtis-inspired 4H index. Displacement from the null plane ( null)a and the proportion of bootstrap samples (p) falling further from the vertex than the null plane are both calculated by the FourHnullplaneD function. null 0.0476518 For this reason, we envision the 4H index as a tool that can be used for exploring and comparing patterns and formulating hypotheses about underlying eco-evolutionary processes of microbiome restructuring after host hybridization. While not explicitly explored in this manuscript, the 4H index could easily be extended to examine HA microbiome functional p change between hybrids and their progenitors. Indeed, when applied 0.006 to function, the 4H index could be useful for forming hypotheses re- 0.764 lated to hybrid ecology and/or ecological success. However, even in Neotoma woodrat 0.830025 Aspidoscelis lizard 0.392932 Kikihia cicada 0.3025042 0.05380129 0.19 this context the 4H index should be interpreted as a tool for charac- Maize B73 Mo17 (leaf, bacteria) 0.6684608 0.03651511 0.132 terizing patterns of change, rather than mechanisms. This is because Maize B73 Mo17 (rhizosphere, bacteria) 0.7801631 0.06678432 0 Maize B73 Mo17 (leaf, fungi) 0.8294745 0.05757388 0 Maize B73 Mo17 (rhizosphere, fungi) 0.7474605 0.06723603 0 Maize B73 CML103 (leaf, bacteria) 0.8038933 0.05631366 0.006 Maize B73 Mo18W (leaf, bacteria) 0.7243933 0.06803136 0.016 Maize B73 CML103 (rhizosphere, fungi) 0.7003835 0.07931675 0 Maize B73 Mo18W (rhizosphere, fungi) 0.7082141 0.0815329 0 0.030863 microbiome function and host ecology can have bidirectional impacts, and thus, it can be challenging to delineate cause and effect. As a result, though function may provide better insight into potential pattern a Positive values of null indicate that points lie closer to the intersection vertex than expected by chance, suggesting that hybrids are more likely to retain taxa shared by both progenitors than they are to retain taxa shared by only one of the two progenitors. Negative values indicate the opposite, namely that hybrids are more likely to retain taxa only found on one of the two progenitors than they are to retain taxa found on both progenitors. generating mechanisms, the 4H index is not a test for causality, but rather an exploratory tool for hypothesis generation. Outside the context of hybridization, it is worth noting that this same framework can be applied to any triplet of host species, where one of the three host species is in some way intermediate to the other two. Thus, for example, a 4H index could be calculated for the microbiomes of organisms from an ecotonal habitat, and then compared to the microbiomes of organisms from the two pure habitat types on either end of the ecotone (O'Brien et al., 2022), even if it is the same host taxon across the entire zone. Likewise, a 4H index could be calculated for species (e.g. swordtail males, Xiphophorus nigrensis) that exhibit three discrete size classes, with one size class being intermediate to the other two (Morris et al., 1992). Similarly, a 4H index could be calculated for captive animals fed two different pure diets as compared to captive animals fed a mixed diet. In these scenarios, the interpretation of our four conceptual models would change. However, because the 4H metric is defined solely based on distributions of microbial presence/absence or abundance across non-overlapping sets hybrid plant versus hybrid animal microbiomes, or between hybrid of host classes, it is valid for any analysis where there is ecological, vertebrate versus hybrid invertebrate microbiomes. Likewise, the evolutionary, morphological or physiological reason to believe that 4H index could be used to determine how phylogenetic and/or phe- one host class falls between the other two host classes. notypic distances between progenitors or ploidy level impact the hybrid microbiome. AU T H O R C O N T R I B U T I O N S Importantly, the intent of each of the four conceptual models Benjamin T. Camper and Sharon Bewick wrote the first draft of the and, indeed, the 4H index in general is to highlight hybridprogeni- manuscript. Sharon Bewick developed the code for the R package. All tor microbiome relationships. Thus, like beta diversity, the 4H index authors contributed substantially to the content of the manuscript. should be taken as a measure of pattern, not process. Just as beta diversity cannot be used to explain why turnover differs among com- AC K N OW L E D G E M E N T S munities, the 4H index should not be used to discriminate among We thank Andrew Kanes, Thomas Dempster, August Spencer microbiome reassembly mechanisms responsible for microbiome and Eva Purcell for their field assistance in New Mexico as well restructuring after host hybridization. In the woodrat system, for as Eva Purcell, Lily Margeson, Simon Dunn, Georgianna Bellinger, example, the 4H index cannot be used to explain why Intersection Henry Egloff, Kaila Hodges, Camryn Lachica and Savannah Utz dominates. It may be that the hybrid woodrat immune system is re- for their assistance assembling drift fence trapping arrays. We fractory to all microbes not found on both progenitors. Alternatively, thank Daniel Nielsen for generously providing additional woodrat it could be that hybrid woodrats are restricted to habitats where data and insight into the woodrat system. All research was ap- both progenitors overlap and the hybrid microbiome reflects micro- proved by Clemson University under IACUC protocol numbers bial exposure patterns of hybrid animals. Regardless, experimental #2020-015 and #2021-0 47. We completed this work under the work will always be needed to understand what drives hybridpro- Sevilleta National Wildlife Refuge Special Use Permit #SEV_ genitor HA microbiome relationships observed using the 4H index. Bewick_Camper_2022_59, the USDA-A RS Jornada study permit 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 524 #592 and the New Mexico Department of Game and Fish permit authorization #3772. This study was funded by the NSF award #2105604, a Clemson University Support for Early Exploration and Development (CUSEED) Grant and the Clemson University Creative Inquiry (CI) Program. C O N FL I C T O F I N T E R E S T S TAT E M E N T No conflicts of interest have been declared. PEER REVIEW The peer review history for this article is available at https://w ww. webofs cience.com/a pi/g ateway/wos/p eer-review/10.1111/2041- 210X.14279. DATA AVA I L A B I L I T Y S TAT E M E N T Cleaned data for cicadas, woodrats, lizards and maize and all scripts to generate the figures and tables for this manuscript are available on GitHub (https://github.com/bewicklab/HybridMicrobiom eFramework) and Zenodo (https://zenodo.org/records/10358091) (Camper et al., 2023). The HybridMicrobiomes R package is available from CRAN (https://cran.r-projec t.org/web/packages/HybridMicr obiomes/index.html). ORCID Benjamin T. Camper Zachary Laughlin Daniel Malagon https://orcid.org/0000-0002-7861-485X https://orcid.org/0009-0003-7931-7130 https://orcid.org/0000-0003-2831-4370 Robert Denton https://orcid.org/0000-0002-8629-1376 Sharon Bewick https://orcid.org/0000-0002-2563-5761 REFERENCES Abbott, K. C., Eppinga, M. B., Umbanhowar, J., Baudena, M., & Bever, J. D. (2021). Microbiome influence on host community dynamics: Conceptual integration of microbiome feedback with classical hostmicrobe theory. Ecology Letters, 24, 27962811. Abbott, R., Albach, D., Ansell, S., Arntzen, J. W., Baird, S. J., Bierne, N., Boughman, J., Brelsford, A., Buerkle, C. A., & Buggs, R. (2013). Hybridization and speciation. Journal of Evolutionary Biology, 26, 229246. Ahn, J., Sinha, R., Pei, Z., Dominianni, C., Wu, J., Shi, J., Goedert, J. J., Hayes, R. B., & Yang, L. (2013). Human gut microbiome and risk for colorectal cancer. Journal of the National Cancer Institute, 105, 19071911. Aitchison, J. (1982). The statistical analysis of compositional data. Journal of the Royal Statistical Society: Series B (Methodological), 44, 139160. Anderson, M. J. (2014). Permutational multivariate analysis of variance (PERMANOVA). Wiley Statsref: Statistics Reference Online, 115. https://doi.org/10.1002/9781118445112.stat07841 Apprill, A., Miller, C. A., Van Cise, A. M., U'Ren, J. M., Leslie, M. S., Weber, L., Baird, R. W., Robbins, J., Landry, S., & Bogomolni, A. (2020). Marine mammal skin microbiotas are influenced by host phylogeny. Royal Society Open Science, 7, 192046. Archie, E. A., & Theis, K. R. (2011). Animal behaviour meets microbial ecology. Animal Behaviour, 82, 425436. Baeckens, S. (2019). Evolution of animal chemical communication: Insights from non-model species and phylogenetic comparative methods. Belgian Journal of Zoology, 149, 6393. 525 Baedke, J., Fbregas-Tejeda, A., & Nieves Delgado, A. (2020). The holobiont concept before Margulis. Journal of Experimental Zoology Part B: Molecular and Developmental Evolution, 334, 149155. Bahrndorff, S., Alemu, T., Alemneh, T., & Lund Nielsen, J. (2016). The microbiome of animals: Implications for conservation biology. International Journal of Genomics, 2016, 17. Baird, S. J., Ribas, A., Macholn, M., Albrecht, T., Pilek, J., & Goy de Bellocq, J. (2012). Where are the wormy mice? A reexamination of hybrid parasitism in the European house mouse hybrid zone. Evolution: International Journal of Organic Evolution, 66, 27572772. Balard, A., & Heitlinger, E. (2022). Shifting focus from resistance to disease tolerance: A review on hybrid house mice. Ecology and Evolution, 12, e8889. Banerjee, A., Cornejo, J., & Bandopadhyay, R. (2020). Emergent climate change impact throughout the world: Call for microbiome conservation before it's too late. Biodiversity and Conservation, 29, 345348. Barton, N. H., & Hewitt, G. M. (1985). Analysis of hybrid zones. Annual Review of Ecology and Systematics, 16, 113148. Bates, K. A., Clare, F. C., O'Hanlon, S., Bosch, J., Brookes, L., Hopkins, K., McLaughlin, E. J., Daniel, O., Garner, T. W., & Fisher, M. C. (2018). Amphibian chytridiomycosis outbreak dynamics are linked with host skin bacterial community structure. Nature Communications, 9, 693. Bates, K. A., Sommer, U., Hopkins, K. P., Shelton, J. M., Wierzbicki, C., Sergeant, C., Tapley, B., Michaels, C. J., Schmeller, D. S., & Loyau, A. (2022). Microbiome function predicts amphibian chytridiomycosis disease dynamics. Microbiome, 10, 116. Bateson, P. (2002). William Bateson: A biologist ahead of his time. Journal of Genetics, 81, 4958. Bateson, P. (1984). Behavioral evolution and integrative levels: The Tc Schneirla conferences series, Volume 1. Psychology Press. Belcaid, M., Casaburi, G., McAnulty, S. J., Schmidbaur, H., Suria, A. M., Moriano-Gutierrez, S., Pankey, M. S., Oakley, T. H., Kremer, N., & Koch, E. J. (2019). Symbiotic organs shaped by distinct modes of genome evolution in cephalopods. Proceedings of the National Academy of Sciences of the United States of America, 116, 30303035. Bergot, A.-S., Giri, R., & Thomas, R. (2019). The microbiome and rheumatoid arthritis. Best Practice & Research. Clinical Rheumatology, 33, 101497. Blyton, M. D., Soo, R. M., Whisson, D., Marsh, K. J., Pascoe, J., Le Pla, M., Foley, W., Hugenholtz, P., & Moore, B. D. (2019). Faecal inoculations alter the gastrointestinal microbiome and allow dietary expansion in a wild specialist herbivore, the koala. Animal Microbiome, 1, 118. Bordenstein, S. R., & Theis, K. R. (2015). Host biology in light of the microbiome: Ten principles of holobionts and hologenomes. PLoS Biology, 13, e1002226. Bosch, T. C., & Miller, D. J. (2016). The holobiont imperative. Springer. Bravo, J. A., Forsythe, P., Chew, M. V., Escaravage, E., Savignac, H. M., Dinan, T. G., Bienenstock, J., & Cryan, J. F. (2011). Ingestion of lactobacillus strain regulates emotional behavior and central GABA receptor expression in a mouse via the vagus nerve. Proceedings of the National Academy of Sciences of the United States of America, 108, 1605016055. Bray, J. R., & Curtis, J. T. (1957). An ordination of the upland forest communities of southern Wisconsin. Ecological Monographs, 27, 326349. Bredon, M., Herran, B., Bertaux, J., Grve, P., Moumen, B., & Bouchon, D. (2020). Isopod holobionts as promising models for lignocellulose degradation. Biotechnology for Biofuels, 13, 114. Britton-Davidian, J., Fel-Clair, F., Lopez, J., Alibert, P., & Boursot, P. (2005). Postzygotic isolation between the two European subspecies of the house mouse: Estimates from fertility patterns in wild and laboratory-bred hybrids. Biological Journal of the Linnean Society, 84, 379393. 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | CAMPER et al. | Brotman, R. M., Klebanoff, M. A., Nansel, T. R., Yu, K. F., Andrews, W. W., Zhang, J., & Schwebke, J. R. (2010). Bacterial vaginosis assessed by gram stain and diminished colonization resistance to incident gonococcal, chlamydial, and trichomonal genital infection. Journal of Infectious Diseases, 202, 19071915. Brown, H., & Bouton, J. H. (1993). Physiology and genetics of interspecific hybrids between photosynthetic types. Annual Review of Plant Biology, 44, 435456. Brucker, R. M., & Bordenstein, S. R. (2012). Speciation by symbiosis. Trends in Ecology & Evolution, 27, 443451. Brucker, R. M., & Bordenstein, S. R. (2013). The hologenomic basis of speciation: Gut bacteria cause hybrid lethality in the genus Nasonia. Science, 341, 667669. Buffie, C. G., & Pamer, E. G. (2013). Microbiota-mediated colonization resistance against intestinal pathogens. Nature Reviews Immunology, 13, 790801. Campbell, L. G., Snow, A. A., & Ridley, C. E. (2006). Weed evolution after crop gene introgression: Greater survival and fecundity of hybrids in a new environment. Ecology Letters, 9, 11981209. Campbell, M. T., Proctor, C. A., Dou, Y., Schmitz, A. J., Phansak, P., Kruger, G. R., Zhang, C., & Walia, H. (2015). Genetic and molecular characterization of submergence response identifies Subtol6 as a major submergence tolerance locus in maize. PLoS ONE, 10, e0120385. Camper, B., Laughlin, Z., Malagon, D., Denton, R., & Bewick, S. (2023). Zenodo HybridMicrobiomesFrameworkv1.1. https://zenodo.org/ records/10358091 Capblancq, T., Desprs, L., & Mavrez, J. (2020). Genetic, morphological and ecological variation across a sharp hybrid zone between two alpine butterfly species. Evolutionary Applications, 13, 14351450. Carreira, V. P., Soto, I. M., Fanara, J. J., & Hasson, E. (2008). A study of wing morphology and fluctuating asymmetry in interspecific hybrids between Drosophila buzzatii and D. koepferae. Genetica, 133, 111. Carthey, A. J., Blumstein, D. T., Gallagher, R. V., Tetu, S. G., & Gillings, M. R. (2020). Conserving the holobiont. Functional Ecology, 34, 764776. Chen, Y. E., Fischbach, M. A., & Belkaid, Y. (2018). Skin microbiotahost interactions. Nature, 553, 427436. Chiarello, M., Auguet, J.-C ., Bettarel, Y., Bouvier, C., Claverie, T., Graham, N. A., Rieuvilleneuve, F., Sucr, E., Bouvier, T., & Villger, S. (2018). Skin microbiome of coral reef fish is highly variable and driven by host phylogeny and diet. Microbiome, 6, 114. Chong, P. P., Chin, V. K., Looi, C. Y., Wong, W. F., Madhavan, P., & Yong, V. C. (2019). The microbiome and irritable bowel syndromeA review on the pathophysiology, current research and future therapy. Frontiers in Microbiology, 10, 1136. Claus, S. P., Tsang, T. M., Wang, Y., Cloarec, O., Skordi, E., Martin, F. P., Rezzi, S., Ross, A., Kochhar, S., & Holmes, E. (2008). Systemic multicompartmental effects of the gut microbiome on mouse metabolic phenotypes. Molecular Systems Biology, 4, 219. Cooper, R. D., & Shaffer, H. B. (2021). Allele-specific expression and gene regulation help explain transgressive thermal tolerance in non-native hybrids of the endangered California tiger salamander (Ambystoma californiense). Molecular Ecology, 30, 9871004. Corbin, K. D., Krajmalnik-Brown, R., Carnero, E. A., Bock, C., Emerson, R., Rittmann, B. E., Marcus, A. K., Davis, T., Dirks, B., & Ilhan, Z. E. (2020). Integrative and quantitative bioenergetics: Design of a study to assess the impact of the gut microbiome on host energy balance. Contemporary Clinical Trials Communications, 19, 100646. Cregger, M., Veach, A., Yang, Z., Crouch, M., Vilgalys, R., Tuskan, G., & Schadt, C. (2018). The Populus holobiont: Dissecting the effects of plant niches and genotype on the microbiome. Microbiome, 6, 114. Davidson, G. L., Cooke, A. C., Johnson, C. N., & Quinn, J. L. (2018). The gut microbiome as a driver of individual variation in cognition and functional behaviour. Philosophical Transactions of the Royal Society B: Biological Sciences, 373, 20170286. CAMPER et al. Delaux, P.-M., & Schornack, S. (2021). Plant evolution driven by interactions with symbiotic and pathogenic microbes. Science, 371, eaba6605. Deng, H. (2012). A review of diversity-stability relationship of soil microbial community: What do we not know? Journal of Environmental Sciences, 24, 10271035. Dice, L. R. (1945). Measures of the amount of ecologic association between species. Ecology, 26, 297302. Dittrich-Reed, D. R., & Fitzpatrick, B. M. (2013). Transgressive hybrids as hopeful monsters. Evolutionary Biology, 40, 310315. Dobzhansky, T. (1934). Studies on hybrid sterility. Zeitschrift fr Zellforschung und Mikroskopische Anatomie, 21, 169223. Dowling, T. E., & Secor, C. L. (1997). The role of hybridization and introgression in the diversification of animals. Annual Review of Ecology and Systematics, 28, 593619. Egozcue, J. J., Pawlowsky-Glahn, V., Mateu-Figueras, G., & Barcelo-V idal, C. (2003). Isometric logratio transformations for compositional data analysis. Mathematical Geology, 35, 279300. Ehrlich, P. R., & Raven, P. H. (1964). Butterflies and plants: A study in coevolution. Evolution, 18, 586608. Ezenwa, V. O., Gerardo, N. M., Inouye, D. W., Medina, M., & Xavier, J. B. (2012). Animal behavior and the microbiome. Science, 338, 198199. Filzmoser, P., Hron, K., & Reimann, C. (2010). The bivariate statistical analysis of environmental (compositional) data. Science of the Total Environment, 408, 42304238. Fontaine, S. S., & Kohl, K. D. (2020). Optimal integration between host physiology and functions of the gut microbiome. Philosophical Transactions of the Royal Society B: Biological Sciences, 375, 20190594. Forejt, J. (1996). Hybrid sterility in the mouse. Trends in Genetics, 12, 412417. Forejt, J., & Ivnyi, P. (1974). Genetic studies on male sterility of hybrids between laboratory and wild mice (Mus musculus L.). Genetics Research, 24, 189206. Friesen, M. L., Porter, S. S., Stark, S. C., von Wettberg, E. J., Sachs, J. L., & Martinez-Romero, E. (2011). Microbially mediated plant functional traits. Annual Review of Ecology, Evolution, and Systematics, 42, 2346. Fronk, D. C., & Sachs, J. L. (2022). Symbiotic organs: The nexus of host microbe evolution. Trends in Ecology & Evolution, 37, 599610. Gaona, O., Cerqueda-Garca, D., Martnez-Martnez, D., Moya, A., & Falcn, L. I. (2016). Microbiome of the sexual scent organ of Leptonycteris yerbabuenae. Report no. 2167-9843 (PeerJ Preprints). Goldschmidt, R. (1933). Some aspects of evolution. Science, 78, 539547. Goldschmidt, R. (1940). The material basis of evolution. Yale University Press. Good, J. M., Handel, M. A., & Nachman, M. W. (2008). Asymmetry and polymorphism of hybrid male sterility during the early stages of speciation in house mice. Evolution: International Journal of Organic Evolution, 62, 5065. Greene, L. K., Williams, C. V., Junge, R. E., Mahefarisoa, K. L., Rajaonarivelo, T., Rakotondrainibe, H., O'Connell, T. M., & Drea, C. M. (2020). A role for gut microbiota in host niche differentiation. The ISME Journal, 14, 16751687. Grinberg, M., Levin, R., Neuman, H., Ziv, O., Turjeman, S., Gamliel, G., Nosenko, R., & Koren, O. (2022). Antibiotics increase aggression behavior and aggression-related pheromones and receptors in Drosophila melanogaster. Iscience, 25, 104371. Haji, D., Vailionis, J., Stukel, M., Gordon, E., Lemmon, E. M., Lemmon, A. R., & Simon, C. (2022). Lack of host phylogenetic structure in the gut bacterial communities of New Zealand cicadas and their interspecific hybrids. Scientific Reports, 12, 20559. Harris, R. N., James, T. Y., Lauer, A., Simon, M. A., & Patel, A. (2006). Amphibian pathogen Batrachochytrium dendrobatidis is inhibited by the cutaneous bacteria of amphibian species. EcoHealth, 3, 5356. 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 526 Harrison, X. A., Price, S. J., Hopkins, K., Leung, W. T., Sergeant, C., & Garner, T. W. (2019). Diversity-stability dynamics of the amphibian skin microbiome and susceptibility to a lethal viral pathogen. Frontiers in Microbiology, 10, 2883. He, J., Li, J., & Zheng, Y. (2013). Thoughts on the microbial diversity- stability relationship in soil ecosystems. Biodiversity Science, 21, 411420. Heard, S. B., & Hauser, D. L. (1995). Key evolutionary innovations and their ecological mechanisms. Historical Biology, 10, 151173. Heys, C., Fisher, A., Dewhurst, A., Lewis, Z., & Liz, A. (2021). Exposure to foreign gut microbiota can facilitate rapid dietary shifts. Scientific Reports, 11, 16791. Hovick, S. M., & Whitney, K. D. (2014). Hybridisation is associated with increased fecundity and size in invasive taxa: Meta-analytic support for the hybridisation-invasion hypothesis. Ecology Letters, 17, 14641477. Hu, Y., Sanders, J. G., ukasik, P., D'Amelio, C. L., Millar, J. S., Vann, D. R., Lan, Y., Newton, J. A., Schotanus, M., & Kronauer, D. J. (2018). Herbivorous turtle ants obtain essential nutrients from a conserved nitrogen-recycling gut microbiome. Nature Communications, 9, 964. Ives, A. R., & Carpenter, S. R. (2007). Stability and diversity of ecosystems. Science, 317, 5862. Jaccard, P. (1908). Nouvelles recherches sur la distribution florale. Bulletin de la Societe Vaudoise des Sciences Naturelles, 44, 223270. Jackson, J. F. (1973). The phenetics and ecology of a narrow hybrid zone. Evolution, 27, 5868. Janz, N. (2011). Ehrlich and Raven revisited: Mechanisms underlying codiversification of plants and enemies. Annual Review of Ecology, Evolution, and Systematics, 42, 7189. Janzen, D. H. (1980). When is it coevolution? Evolution, 34(3), 611612. https://doi.org/10.1111/j.1558-5646.1980.tb04849.x Jia, Y., Jin, S., Hu, K., Geng, L., Han, C., Kang, R., Pang, Y., Ling, E., Tan, E. K., & Pan, Y. (2021). Gut microbiome modulates drosophila aggression through octopamine signaling. Nature Communications, 12, 2698. Jimnez, R. R., & Sommer, S. (2017). The amphibian microbiome: Natural range of variation, pathogenic dysbiosis, and role in conservation. Biodiversity and Conservation, 26, 763786. Jin Song, S., Woodhams, D. C., Martino, C., Allaband, C., Mu, A., Javorschi-Miller-Montgomery, S., Suchodolski, J. S., & Knight, R. (2019). Engineering the microbiome for animal health and conservation. Experimental Biology and Medicine, 244, 494504. Jing, T.-Z., Qi, F.-H., & Wang, Z.-Y. (2020). Most dominant roles of insect gut bacteria: Digestion, detoxification, or essential nutrient provision? Microbiome, 8, 120. Jost, L., Choa, A., & Chazdon, R. L. (2011). Compositional similarity and (beta) diversity. In A. E. Magurran & B. J. Mc Gill (Eds.), Biological diversity: Frontiers in measurement and assessment (pp. 6684). Oxford University Press. Ju, J.-F., Bing, X.-L ., Zhao, D.-S., Guo, Y., Xi, Z., Hoffmann, A. A., Zhang, K.-J., Huang, H.-J., Gong, J.-T., & Zhang, X. (2020). Wolbachia supplement biotin and riboflavin to enhance reproduction in planthoppers. The ISME Journal, 14, 676687. Kamada, N., Chen, G. Y., Inohara, N., & Nez, G. (2013). Control of pathogens and pathobionts by the gut microbiota. Nature Immunology, 14, 685690. Kearney, M. (2005). Hybridization, glaciation and geographical parthenogenesis. Trends in Ecology & Evolution, 20, 495502. Kirchoff, N. S., Udell, M. A., & Sharpton, T. J. (2019). The gut microbiome correlates with conspecific aggression in a small population of rescued dogs (Canis familiaris). PeerJ, 7, e6103. Kohl, K. D., Weiss, R. B., Cox, J., Dale, C., & Denise Dearing, M. (2014). Gut microbes of mammalian herbivores facilitate intake of plant toxins. Ecology Letters, 17, 12381246. Kolodny, O., Callahan, B. J., & Douglas, A. E. (2020). The role of the microbiome in host evolution. Philosophical Transactions of the Royal Society B: Biological Sciences, 375, 20190588. | 527 Kratochwil, C. F., & Meyer, A. (2015). Closing the genotypephenotype gap: Emerging technologies for evolutionary genetics in ecological model vertebrate systems. BioEssays, 37, 213226. Kueneman, J. G., Parfrey, L. W., Woodhams, D. C., Archer, H. M., Knight, R., & McKenzie, V. J. (2014). The amphibian skin-associated microbiome across species, space and life history stages. Molecular Ecology, 23, 12381250. Lafarga-De la Cruz, F., Nez-Acua, G., & Gallardo-Escrate, C. (2013). Hybridization between Haliotis rufescens and Haliotis discus hannai: Evaluation of fertilization, larval development, growth and thermal tolerance. Aquaculture Research, 44, 12061220. Larouche, O., Hodge, J. R., Alencar, L. R., Camper, B., Adams, D. S., Zapfe, K., Friedman, S. T., Wainwright, P. C., & Price, S. A. (2020). Do key innovations unlock diversification? A case-study on the morphological and ecological impact of pharyngognathy in acanthomorph fishes. Current Zoology, 66, 575588. Lee, C.-R ., Wang, B., Mojica, J. P., Mandkov, T., Prasad, K. V., Goicoechea, J. L., Perera, N., Hellsten, U., Hundley, H. N., & Johnson, J. (2017). Young inversion with multiple linked QTLs under selection in a hybrid zone. Nature Ecology & Evolution, 1, 0119. Legendre, P. (2014). Interpreting the replacement and richness difference components of beta diversity. Global Ecology and Biogeography, 23, 13241334. Ley, R. E., Hamady, M., Lozupone, C., Turnbaugh, P. J., Ramey, R. R., Bircher, J. S., Schlegel, M. L., Tucker, T. A., Schrenzel, M. D., & Knight, R. (2008). Evolution of mammals and their gut microbes. Science, 320, 16471651. Li, M., Wang, B., Zhang, M., Rantalainen, M., Wang, S., Zhou, H., Zhang, Y., Shen, J., Pang, X., & Zhang, M. (2008). Symbiotic gut microbes modulate human metabolic phenotypes. Proceedings of the National Academy of Sciences of the United States of America, 105, 21172122. Li, W., Liu, J., Tan, H., Yang, C., Ren, L., Liu, Q., Wang, S., Hu, F., Xiao, J., & Zhao, R. (2018). Genetic effects on the gut microbiota assemblages of hybrid fish from parents with different feeding habits. Frontiers in Microbiology, 9, 2972. Li, Z., Wright, A. D. G., Si, H., Wang, X., Qian, W., Zhang, Z., & Li, G. (2016). Changes in the rumen microbiome and metabolites reveal the effect of host genetics on hybrid crosses. Environmental Microbiology Reports, 8, 10161023. Maebe, K., Vereecken, N. J., Piot, N., Revert, S., Cejas, D., Michez, D., Vandamme, P., & Smagghe, G. (2021). The holobiont as a key to the adaptation and conservation of wild bees in the anthropocene. Frontiers in Ecology and Evolution, 9, 781470. Magurran, A. E., & McGill, B. J. (2010). Biological diversity: Frontiers in measurement and assessment. OUP Oxford. Mallet, J. (2007). Hybrid speciation. Nature, 446, 279283. Malukiewicz, J., Cartwright, R. A., Dergam, J. A., Igayara, C. S., Kessler, S., Moreira, S. B., Nash, L. T., Nicola, P. A., Pereira, L. C., & Pissinati, A. (2019). The effects of host taxon, hybridization, and environment on the gut microbiome of callithrix marmosets. BioRxiv, 708255. https://doi.org/10.1101/708255 Margulis, L., & Fester, R. (1991). Symbiosis as a source of evolutionary innovation: Speciation and morphogenesis. MIT Press. Martins, N., Pearson, G. A., Gouveia, L., Tavares, A. I., Serrao, E. A., & Bartsch, I. (2019). Hybrid vigour for thermal tolerance in hybrids between the allopatric kelps Laminaria digitata and L. pallida (Laminariales, Phaeophyceae) with contrasting thermal affinities. European Journal of Phycology, 54, 548561. Mayneris-Perxachs, J., Bolick, D. T., Leng, J., Medlock, G. L., Kolling, G. L., Papin, J. A., Swann, J. R., & Guerrant, R. L. (2016). Protein-and zinc-deficient diets modulate the murine microbiome and metabolic phenotype. The American Journal of Clinical Nutrition, 104, 12531262. McCann, K. S. (2000). The diversitystability debate. Nature, 405, 228233. 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License CAMPER et al. | McMurdie, P. J., & Holmes, S. (2013). phyloseq: An R package for reproducible interactive analysis and graphics of microbiome census data. PLoS ONE, 8, e61217. Meadows, B. A. (2022). Detecting signatures of discordant population structure and local adaptation in a codiversifying host-obligate endosymbiont mutualism. Texas Tech University. Mrot, C., Debat, V., Le Poul, Y., Merrill, R. M., Naisbit, R. E., Tholance, A., Jiggins, C. D., & Joron, M. (2020). Hybridization and transgressive exploration of colour pattern and wing morphology in Heliconius butterflies. Journal of Evolutionary Biology, 33, 942956. Miller, A. K., Westlake, C. S., Cross, K. L., Leigh, B. A., & Bordenstein, S. R. (2021). The microbiome impacts host hybridization and speciation. PLoS Biology, 19, e3001417. Moeller, A. H., & Sanders, J. G. (2020). Roles of the gut microbiota in the adaptive evolution of mammalian species. Philosophical Transactions of the Royal Society B: Biological Sciences, 375, 20190597. Moran, N. A., & Sloan, D. B. (2015). The hologenome concept: Helpful or hollow? PLoS Biology, 13, e1002311. Morris, M. R., Batra, P., & Ryan, M. J. (1992). Male-male competition and access to females in the swordtail Xiphophorus nigrensis. Copeia, 1992, 980986. Muller, H. J. (1942). Isolating mechanisms, evolution, and temperature. Biology Symposium, 6, 71125. Muoz, M., & Bodensteiner, B. (2019). Janzen's hypothesis meets the Bogert effect: Connecting climate variation, thermoregulatory behavior, and rates of physiological evolution. Integrative Organismal Biology, 1, oby002. Neufeld, K., Kang, N., Bienenstock, J., & Foster, J. A. (2011). Reduced anxiety-like behavior and central neurochemical change in germ- free mice. Neurogastroenterology and Motility, 23, 255264; e119. Nielsen, D. P., Harrison, J. G., Byer, N. W., Faske, T. M., Parchman, T. L., Simison, W. B., & Matocq, M. D. (2023). The gut microbiome reflects ancestry despite dietary shifts across a hybrid zone. Ecology Letters, 26, 6375. Nieuwdorp, M., Gilijamse, P. W., Pai, N., & Kaplan, L. M. (2014). Role of the microbiome in energy regulation and metabolism. Gastroenterology, 146, 15251533. Nishida, A. H., & Ochman, H. (2021). Captivity and the co-diversification of great ape microbiomes. Nature Communications, 12, 5632. Nobs, S. P., Tuganbaev, T., & Elinav, E. (2019). Microbiome diurnal rhythmicity and its impact on host physiology and disease risk. EMBO Reports, 20, e47129. O'Brien, A. M., Laurich, J., & Frederickson, M. E. (2022). Having the right microbiome matters for host trait expression and the strength of mutualism between duckweeds and microbes. bioRxiv, 2022.2002. 2010.479958. O'Brien, A. M., Sawers, R. J., Strauss, S. Y., & Ross-Ibarra, J. (2019). Adaptive phenotypic divergence in an annual grass differs across biotic contexts. Evolution, 73, 22302246. Ochman, H., Worobey, M., Kuo, C.-H., Ndjango, J.-B. N., Peeters, M., Hahn, B. H., & Hugenholtz, P. (2010). Evolutionary relationships of wild hominids recapitulated by gut microbial communities. PLoS Biology, 8, e1000546. Opstal, E. J. V., & Bordenstein, S. R. (2015). Rethinking heritability of the microbiome. Science, 349, 11721173. Orr, H. A. (1995). The population genetics of speciation: The evolution of hybrid incompatibilities. Genetics, 139, 18051813. Orr, H. A., & Turelli, M. (2001). The evolution of postzygotic isolation: Accumulating Dobzhansky-Muller incompatibilities. Evolution, 55, 10851094. Pala, I., & Coelho, M. M. (2005). Contrasting views over a hybrid complex: Between speciation and evolutionary dead-end. Gene, 347, 283294. Patton, A. H., Margres, M. J., Epstein, B., Eastman, J., Harmon, L. J., & Storfer, A. (2020). Hybridizing salamanders experience accelerated diversification. Scientific Reports, 10, 112. CAMPER et al. Pereira, R. J., Barreto, F. S., & Burton, R. S. (2014). Ecological novelty by hybridization: Experimental evidence for increased thermal tolerance by transgressive segregation in Tigriopus californicus. Evolution, 68, 204215. Petchey, O. L., & Gaston, K. J. (2002). Functional diversity (FD), species richness and community composition. Ecology Letters, 5, 402411. Pfennig, K. S. (2021). Biased hybridization and its impact on adaptive introgression. Trends in Ecology & Evolution, 36, 488497. Phillips, C. D., Phelan, G., Dowd, S. E., McDONOUGH, M. M., Ferguson, A. W., Delton Hanson, J., Siles, L., Ordez-Garza, N., San Francisco, M., & Baker, R. J. (2012). Microbiome analysis among bats describes influences of host phylogeny, life history, physiology and geography. Molecular Ecology, 21, 26172627. Pimentel, M., & Lembo, A. (2020). Microbiome and its role in irritable bowel syndrome. Digestive Diseases and Sciences, 65, 829839. Postler, T. S., & Ghosh, S. (2017). Understanding the holobiont: How microbial metabolites affect human health and shape the immune system. Cell Metabolism, 26, 110130. Quinn, T. P., Erb, I., Gloor, G., Notredame, C., Richardson, M. F., & Crowley, T. M. (2019). A field guide for the compositional analysis of any-omics data. GigaScience, 8, giz107. Rebollar, E. A., Antwis, R. E., Becker, M. H., Belden, L. K., Bletz, M. C., Brucker, R. M., Harrison, X. A., Hughey, M. C., Kueneman, J. G., & Loudon, A. H. (2016). Using omics and integrated multi-omics approaches to guide probiotic selection to mitigate chytridiomycosis and other emerging infectious diseases. Frontiers in Microbiology, 7, 68. Rebollar, E. A., Martnez-Ugalde, E., & Orta, A. H. (2020). The amphibian skin microbiome and its protective role against chytridiomycosis. Herpetologica, 76, 167177. Redford, K. H., Segre, J. A., Salafsky, N., del Rio, C. M., & McAloose, D. (2012). Conservation and the microbiome. Conservation Biology: The Journal of the Society for Conservation Biology, 26, 195197. Reed, K. M., & Sites, J. W., Jr. (1995). Female fecundity in a hybrid zone between two chromosome races of the Sceloporus grammicus complex (Sauria, Phrynosomatidae). Evolution, 49, 6169. Ren, C. G., Liu, Z. Y., Wang, X. L., & Qin, S. (2022). The seaweed holobiont: From microecology to biotechnological applications. Microbial Biotechnology, 15, 738754. Robbins, T. R., Pruitt, J. N., Straub, L. E., Mccoy, E. D., & Mushinsky, H. R. (2010). Transgressive aggression in Sceloporus hybrids confers fitness through advantages in male agonistic encounters. Journal of Animal Ecology, 79, 137147. Robbins, T. R., Walker, L. E., Gorospe, K. D., Karl, S. A., Schrey, A. W., McCoy, E. D., & Mushinsky, H. R. (2014). Rise and fall of a hybrid zone: Implications for the roles of aggression, mate choice, and secondary succession. Journal of Heredity, 105, 226236. Ross, B. (19831994). The joy of painting. Sage, R. D., Heyneman, D., Lim, K.-C ., & Wilson, A. C. (1986). Wormy mice in a hybrid zone. Nature, 324, 6063. Sampson, T. R., & Mazmanian, S. K. (2015). Control of brain development, function, and behavior by the microbiome. Cell Host & Microbe, 17, 565576. Sanders, J. G., Powell, S., Kronauer, D. J., Vasconcelos, H. L., Frederickson, M. E., & Pierce, N. E. (2014). Stability and phylogenetic correlation in gut microbiota: Lessons from ants and apes. Molecular Ecology, 23, 12681283. Scheelings, T. F., Moore, R. J., Van, T. T. H., Klaassen, M., & Reina, R. D. (2020). Microbial symbiosis and coevolution of an entire clade of ancient vertebrates: The gut microbiota of sea turtles and its relationship to their phylogenetic history. Animal Microbiome, 2, 112. Scher, J. U., & Abramson, S. B. (2011). The microbiome and rheumatoid arthritis. Nature Reviews Rheumatology, 7, 569578. Seehausen, O. (2004). Hybridization and adaptive radiation. Trends in Ecology & Evolution, 19, 198207. 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 528 Sevellec, M., Laporte, M., Bernatchez, A., Derome, N., & Bernatchez, L. (2019). Evidence for host effect on the intestinal microbiota of whitefish (Coregonus sp.) species pairs and their hybrids. Ecology and Evolution, 9, 1176211774. Shade, A., & Stopnisek, N. (2019). Abundance-occupancy distributions to prioritize plant core microbiome membership. Current Opinion in Microbiology, 49, 5058. Sharon, G., Segal, D., Ringo, J. M., Hefetz, A., Zilber-Rosenberg, I., & Rosenberg, E. (2010). Commensal bacteria play a role in mating preference of Drosophila melanogaster. Proceedings of the National Academy of Sciences of the United States of America, 107, 2005120056. Sorensen, T. (1948). A method of establishing groups of equal amplitude in plant sociology based on similarity of species content and its application to analyses of the vegetation on Danish commons. Biologiske Skrifter, 5, 134. Suzuki, T. A., Fitzstevens, J. L., Schmidt, V. T., Enav, H., Huus, K. E., Mbong Ngwese, M., Griehammer, A., Pfleiderer, A., Adegbite, B. R., & Zinsou, J. F. (2022). Codiversification of gut microbiota with humans. Science, 377, 13281332. Tams, J., Podani, J., & Csontos, P. (2001). An extension of presence/ absence coefficients to abundance data: A new look at absence. Journal of Vegetation Science, 12, 401410. Theien, G. (2006). The proper place of hopeful monsters in evolutionary biology. Theory in Biosciences, 124, 349369. Theien, G. (2009). Saltational evolution: Hopeful monsters are here to stay. Theory in Biosciences, 128, 4351. Thompson, J. N. (1989). Concepts of coevolution. Trends in Ecology & Evolution, 4, 179183. Thompson, J. N. (1994). The coevolutionary process. University of Chicago Press. Thompson, J. N. (2005). The geographic mosaic of coevolution. University of Chicago Press. Todesco, M., Pascual, M. A., Owens, G. L., Ostevik, K. L., Moyers, B. T., Hbner, S., Heredia, S. M., Hahn, M. A., Caseys, C., & Bock, D. G. (2016). Hybridization and extinction. Evolutionary Applications, 9, 892908. Trevelline, B. K., Fontaine, S. S., Hartup, B. K., & Kohl, K. D. (2019). Conservation biology needs a microbial renaissance: A call for the consideration of host-associated microbiota in wildlife management practices. Proceedings of the Royal Society B: Biological Sciences, 286, 20182448. Tripp, E. A., & Manos, P. S. (2008). Is floral specialization an evolutionary dead-end? Pollination system transitions in Ruellia (Acanthaceae). Evolution, 62, 17121737. Turner, L. M., Schwahn, D. J., & Harr, B. (2012). Reduced male fertility is common but highly variable in form and severity in a natural house mouse hybrid zone. Evolution: International Journal of Organic Evolution, 66, 443458. Ubeda, C., Djukovic, A., & Isaac, S. (2017). Roles of the intestinal microbiota in pathogen protection. Clinical & Translational Immunology, 6, e128. Wagg, C., Dudenhffer, J. H., Widmer, F., & Van Der Heijden, M. G. (2018). Linking diversity, synchrony and stability in soil microbial communities. Functional Ecology, 32, 12801292. Wagner, M. R., Roberts, J. H., Balint-Kurti, P., & Holland, J. B. (2020). Heterosis of leaf and rhizosphere microbiomes in field-grown maize. New Phytologist, 228, 10551069. Walker, D. M., Hill, A. J., Albecker, M. A., McCoy, M. W., Grisnik, M., Romer, A., Grajal-Puche, A., Camp, C., Kelehear, C., & Wooten, J. (2019). Variation in the slimy salamander (Plethodon spp.) skin and gut-microbial assemblages is explained by geographic distance and host affinity. Microbial Ecology, 79, 113. Walls, S. C. (2009). The role of climate in the dynamics of a hybrid zone in Appalachian salamanders. Global Change Biology, 15, 19031910. 529 Walter, J., Martnez, I., & Rose, D. J. (2013). Holobiont nutrition: Considering the role of the gastrointestinal microbiota in the health benefits of whole grains. Gut Microbes, 4, 340346. Walters, A. W., Hughes, R. C., Call, T. B., Walker, C. J., Wilcox, H., Petersen, S. C., Rudman, S. M., Newell, P. D., Douglas, A. E., & Schmidt, P. S. (2020). The microbiota influences the Drosophila melanogaster life history strategy. Molecular Ecology, 29, 639653. Wang, J., Kalyan, S., Steck, N., Turner, L. M., Harr, B., Knzel, S., Vallier, M., Hsler, R., Franke, A., & Oberg, H.-H . (2015). Analysis of intestinal microbiota in hybrid house mice reveals evolutionary divergence in a vertebrate hologenome. Nature Communications, 6, 110. West, A. G., Waite, D. W., Deines, P., Bourne, D. G., Digby, A., McKenzie, V. J., & Taylor, M. W. (2019). The microbiome in threatened species conservation. Biological Conservation, 229, 8598. Woodhams, D. C., Bletz, M., Kueneman, J., & McKenzie, V. (2016). Managing amphibian disease with skin microbiota. Trends in Microbiology, 24, 161164. Woodhouse, M. R., Cannon, E. K., Portwood, J. L., Harper, L. C., Gardiner, J. M., Schaeffer, M. L., & Andorf, C. M. (2021). A pan-genomic approach to genome databases using maize as a model system. BMC Plant Biology, 21, 110. Xifra, G., Moreno-Navarrete, J. M., & Fernndez-Real, J. M. (2019). The microbiota and energy balance. In P. Sbraccia & N. Finer (Eds.), Obesity. Endocrinology (pp. 109126). Springer. Zackular, J. P., Baxter, N. T., Iverson, K. D., Sadler, W. D., Petrosino, J. F., Chen, G. Y., & Schloss, P. D. (2013). The gut microbiome modulates colon tumorigenesis. MBio, 4, e00692-13. Zhang, Z., Chen, J., Li, L., Tao, M., Zhang, C., Qin, Q., Xiao, J., Liu, Y., & Liu, S. (2014). Research advances in animal distant hybridization. Science China Life Sciences, 57, 889902. Zheng, H., Nishida, A., Kwong, W. K., Koch, H., Engel, P., Steele, M. I., & Moran, N. A. (2016). Metabolism of toxic sugars by strains of the bee gut symbiont Gilliamella apicola. MBio, 7, e01326-16. Zhu, L., Wang, J., & Bahrndorff, S. (2021). The wildlife gut microbiome and its implication for conservation biology. Frontiers in Microbiology, 12, 1617. Zilber-Rosenberg, I., & Rosenberg, E. (2008). Role of microorganisms in the evolution of animals and plants: The hologenome theory of evolution. FEMS Microbiology Reviews, 32, 723735. S U P P O R T I N G I N FO R M AT I O N Additional supporting information can be found online in the Supporting Information section at the end of this article. Appendix S1: Monotonicity, density invariance and replication invariance. Appendix S2: System pre-analysis. Appendix S3: Example datasets. Appendix S4: Additional figures and tables. How to cite this article: Camper, B. T., Laughlin, Z., Malagon, D., Denton, R., & Bewick, S. (2024). A conceptual framework for host-associated microbiomes of hybrid organisms. Methods in Ecology and Evolution, 15, 511529. https://doi. org/10.1111/2041-210X.14279 2041210x, 2024, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14279 by Marian University, Wiley Online Library on [29/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | CAMPER et al.  ...

... A Historical View of How One Specialized, Alternative Day School Program Moved Away from the Practices of Restraint and Seclusion ____________________________ A Capstone Project Presented to The Faculty of the Fred S. Klipsch Educator College Marian University _________________________ In Partial Fulfillment of the Requirements for the Degree Doctor of Education in Organizational Leadership _______________________ by Taryn Richard May, 2024  Copyright by Taryn Richard All Rights Reserved May 2024 ABSTRACT This retrospective causal-comparative study looks at a specialized day school program to identify themes that led to a change in philosophy and practice that reduced the use of restraint and seclusion. The team reviewed data from the 2011 -2012 school year through the start of the intervention in the 2016-2017 school year to the 2022-2023 school year. After reviewing relevant data, the team utilized the ATALS tool for group reflection. Finally, staff provided individual feedback through a staff survey that collected no identifiable information. The themes endorsed included Collaboration with teams, ProACT Crisis Intervention Training, Restoration Practices, Building Connections, Having a Positive Impact, Regulation Strategies, and Administrative Trust in Staff. Each theme identified had training that the group engaged with to create the system that decreased the use of restraint and seclusion for the scholars served in the specialized day school program. Fred S. Klipsch Educators College Marian University Indianapolis, Indiana APPROVAL OF THE CAPSTONE PROJECT This capstone project, A Historical View of How One Specialized, Alternative Day School Program Moved Away from the Practices of Restraint and Seclusion, has been approved by the Graduate Faculty of the Fred S. Klipsch Educators College in partial fulfillment of the requirements of the degree of Doctor of Education. ______________________________________________ Committee Chair, Dr. Latonya Turner ______________________________________________ Committee Member, Dr. Jacob Tandy ______________________________________________ Committee Member, Dr. Richard VanAker _______________________________ INSERT DATE DEDICATIONS To my loving husband, Colin Richard, who keeps me going. For my son, who is my light. Our sons and daughters deserve a society better than the one they were born into. This journey is dedicated to all who have felt that something is off and entertains the idea that we can move differently. v ACKNOWLEDGEMENTS Thank you to Jaime Wright, Behavior Specialist and thought partner throughout this journey. I met Jaime before working at this specialized day program, where she supported students' movement in and out of the program featured in this research project. In my darkest moments, she was there to help me return to being objective and present. She has traveled into the rabbit hole and back again more times than I can count. Thank you to Trenton Worden, who joined this team when we needed him most. He has been a thought partner and advocate for this work, bringing balance to us all. He has been the mad scientist in the lab as we created something new. Thank you to the leadership team of the specialized day program. Without your push and pull, we would not be where we are today. Lisa Braden, Jaymie Popcheff, William Jones, Adam Honeycutt, Sarah Augustine, Louis Kersh, and Lauren Beckham. You made this vision a reality. Support for this work came from the schools Education Center and outside consultants from ProACT Crisis Intervention, including Dr. Rick Van Acker, who served on this committee. Thank you to Dr. Jacob Tandy, who has partnered with this program and brought amazing scholars into the teaching profession. Thank you to Dr. Latonya Turner, who guided me in pulling the story together. vi TABLE OF CONTENTS DEDICATION...iv ACKNOWLEDGEMENTS...vi LIST OF TABLES...viii LIST OF FIGURES..ix CHAPTERS I. STUDY DESCRIPTION...1 II. LITERATURE REVIEW.....14 III. METHODOLOGY...67 IV. INTERVETNION PROPOSAL..81 V. INTERVENTION EVALUATION..85 VI. IMPLICATIONS AND REFLECTION112 VII. EXECUTIVE SUMMARY114 REFERENCES116 APPENDICES...129 vii LIST OF TABLES Table 1: Steps and Timeline for Study85 Table 2: Incidents of Restraint & Seclusion Compared with Similar Programs94 Table 3: Multinomial Test for Use of Seclusion.,..95 Table 4: Multinomial Test for Use of Restraint......96 Table 5: Multinomial Test for Suspension Incidents.98 Table 6: Multinomial Test for Days of Suspension Issued..99 Table 7: Incidents Totaled Including Reported Restorative Activities .101 Table 8: Staff and Student Body by School Year...103 Table 9: Professional Learning by School Year....................................................104 Table 10: Structural Changes by School Year........................................106 Table 11: ATLAS Participants107 Table 12: Participants in Survey by Year Hired......110 Table 13: Themes from Survey..110 Table 14: Top Themes Identified from Survey Results..119 viii LIST OF FIGURES Figure 1: Enrollment Percentages by Race...88 Figure 2: Enrollment Percentages by Gender...89 Figure 3: Percentages of Grade Levels Served by Year.90 Figure 4: Percentages of Disability Categories Served by Year.91 Figure 5: Average Daily Attendance by Year.92 Figure 6: Restraint and Seclusion Fall 2011-Spring 2023...93 Figure 7: Out of School Suspension98 Figure 8: OSHA Report...101 Figure 9: Percentage of Credits Earned Grades 9-12102 ix 1 STUDY DESCRIPTION Introduction Many studies are available when researching support for scholars with emotional and behavioral challenges. There are studies on the disproportionality of different groups, the impact on teachers, and evidence of various strategies and interventions to support the needs. Recently, there has been a strong movement to focus on positive approaches to behavior change and a movement away from more punitive efforts. Public school institutions focus on Positive Behavior Interventions and Supports (PBIS) and Multi-Tiered Systems of Support (MTSS). Policy and legislative efforts have been made to remove or seriously limit the use of behavioral practices that employ physical restraint and seclusion of children and adolescents in schools. Centennial School, in partnership with Leigh University, provides an example of a school setting where scholars with emotional and behavioral challenges could attend without the fear of restraint and seclusion as these were uncommon practices (Kern, et al., 2011). Centennial School is funded through the Department of Education of Pennsylvania. In 2016, a public school in Indiana restructured its support for students with emotional and behavioral challenges, shifting to a philosophy grounded in the work of Kern and George. Restraint and seclusion have been reported to be overused for issues that do not pose a serious threat of danger or injury. Yet, the potential for injury, trauma, and even death has been reported due to the use of these practices (Alliance Against Restraint & Seclusion, 2022). This school was motivated to find better outcomes for those with emotional and behavioral challenges. They 2 worked to create a safer environment for responding staff and scholars and to increase achievement for all they serve. Unfortunately, many schools continue to address the behavior of schools with emotional and behavioral challenges, employing punitive and exclusionary practices (Sprick & Borgmeier, 2010). These public schools are responding with reactionary practices (Oliver & Reschly, 2010), for example, by increasing timeout spaces, school resource officers, and crisis management training in their respective public school institutions in response to the increase in aggressive outbursts (Sprick & Borgmeier, 2010). There exists considerable concern that these punitive approaches are being employed more frequently with scholars who display special needs. Alliance Against Restraint & Seclusion, 2022, CRDC, 2020). While scholars identified as having a disability, according to the procedures outlined by the Individuals with Disabilities Education Act (IDEA, 1997), represented 13% of all scholars in K-12 education, those who experienced the use of physical restraint represented 80% of all incidents self-reported by the Civil Rights Data Collection (CRDC, 2020). In addition, 41% of self-reported incidents of mechanical restraint were scholars identified with a disability (CRDC, 2020). 77% of self-reported incidents of the use of seclusion were scholars identified as having a disability (CRDC, 2020). When broken down by race, the data continues to be disturbing, with trends for black, indigenous, and persons of color (BIPOC). African American scholars who have an identified disability experience the use of physical restraint in 48% of self-reported incidents. The lived outcomes suggest 3 that the systematic approach within our institutions lacks the ability to address emotional and behavioral challenges. Educational settings are required to provide services through the Individuals with Disabilities Educational Improvement Act IDEIA (2004). Outcomes for scholars with emotional and behavioral challenges supported under IDEIA are dismal (Sprick & Borgmeier, 2010). This leaves the question, what does it take for a public school agency to create lasting change that increases safety and improves outcomes for the emotionally and behaviorally challenged scholars they serve? With so much literature on what to do, why then are there no notable collective improvements within those expressed outcomes? How does one create a system that is good for all? Background Prior to the 2016-2017 school year, a specialized, alternative school day program in Indiana had a high rate of incidents of restraint and seclusion (see data in Methodology). Student academic achievement was low. Restraint and seclusion sometimes led to injuries that resulted in a need for outside medical attention. In the 2016-2017 academic year, 53 initial reports were made that identified a staff member as injured. Eleven of these incidents need additional outside medical treatment. Like many similar programs, the outcomes for scholars attending this program were dismal. Researchers are attempting to help school systems move away from zero-tolerance policies, including the use of restraint and seclusion as a response to maladaptive behaviors (Skiba & Patterson, 2000; Rambo, 2020). Both scholars and responding staff who work 4 and learn in these institutions report varying levels of trauma, and some report that they were responded to in a manner that was considered disrespectful and undignified (Bonner et al., 2002). Many school systems are adding seclusion rooms in the surrounding area of the specialized day program in Indiana. These spaces are often coded as "time out" rooms; however, they lock from the outside and have padded walls. School systems keep the addition of these rooms secret, as evidenced by families who are beginning to discover this and express concern (Schuyler, 2023). A specialized, alternative day school program, where this researcher is the principal of a school in Indiana, experienced a major remodel during the 20202021 school year. This remodel dismantled the seclusion rooms. The school now operates without a designated space for this response. The staff has yet to reduce incidents of seclusion to zero; however, they are steadily progressing in this direction and have been since beginning this work in 2016. Programming was partly modeled through an understanding of how Centennial School operates in partnership with Lehigh University. The major difference between the two schools is that Centennial is considered a private school setting and has layers of staff and scholars from their partnership with Lehigh University supporting their work. The school where the researcher works is a publicly operated and funded school program, which, like so many others, was significantly understaffed in the 2021-2022 school year, all while doing this same work. Dismantling the seclusion spaces in this public school institution did not just happen on a whim. Much work, structure, processing, and reflection led to 5 this monumental moment. Support for this work came from the schools Education Center and outside consultants, including Dr. Van Aker and members of the Indiana Department of Education (IDOE) Wilburn and McGinley. Problem Statement Many factors come together that impact the effectiveness of responding staff when working to support a scholar with emotional and behavioral challenges. Some of these include internal bias, choices in the application of professional learning and training, knowledge of the neurosciences, and the ability to see objectively what has happened to create a conflict. The efforts and need to integrate research and practice are timely and more important than ever before because of the stakes involved in academic and clinical training, research, practice, and health care in general (Kazdin, 2008, p. 157). Research has provided empirically validated positive interventions. Access to this information and choice in implementation often hinder the successful use of those positive interventions. This results in the public school setting is a system that perpetuates the challenges both school staff and scholars face. Application of best practices when working with scholars with emotional and behavioral challenges results from the staffs internal work. Institutions do not know where to turn to create a system that supports those best practices, as so few examples exist. This research intends to uncover factors that lead to lasting change in a specialized, alternative day school program. The purpose of this retrospective causalcomparative study will be to uncover factors that led one specialized, alternative 6 day school program to move from a school that heavily used restraint and seclusion to one that seldom does. This same school increased academic demands, resulting in increased credit attainment for secondary scholars. This occurred in a K-12, public school setting in Indiana over seven years, at the time of this research design, and continues today. This study will focus on those scholars with a broader diagnosis of emotional and behavioral disabilities within the public school system. This can include those with mental health challenges or those with challenges within their environment that make learning difficult. Society continues to explore interactions with one another and the impact of those interactions on self and community. Within the community of differing abilities, there are variations in physical ability, intellectual ability, and emotional and behavioral health. The fight to be heard, to be accepted, and to be recognized has been an ongoing challenge for each of these groups respectively. Responses "through a lens of dignity and respect" (Lendzion, 2022) appear to be the common theme within each group. The community of those with emotional and behavioral challenges has provided society with a way to converse about their differences. Sociologist Judy Signer (2017) coined the distinction between neurotypical and neurodivergent individuals in 1997 to help frame practitioners thinking around this work. While these terms originated with those who are diagnosed with autism, this conversation has opened the idea that society must work with the neurotypical as much as the neurodivergent communities in order to build an understanding of one another (Signer, 2017). There is a history of the neurotypical community 7 member coming in to "fix" the neurodivergent, or in this case, the emotionally or behaviorally challenged individual. There are intervention plans, best practices, and therapies ready to help tackle the challenge of the person with a challenge, opening debates about symptoms and correctly identifying with labels (Yergeau, 2018, p. 139). The neurotypical community member is in danger of being a savior or martyr without having accomplished much to support the individual in making sense of their world. Teachers over-apply consequences and punishments (Sprick & Borgmeier, 2010) using rationales for their responses. These mindsets have evolved from a long and dark history, which will be explored in the literature review (Yergeau, 2018; Meldon, 2017). Recently, attempts for neurodivergent and neurotypical individuals to come together and learn from one another are happening in public schools through more restorative practices and social-emotional learning (Peace Learning Center, Stutzman-Amstutz & Mullett, 2005; Costello, Wachtel, & Wachtel, 2010). This growing effort focuses on how to provide communities with the skills needed to restore safety and care to ensure that challenges do not grow to a place that resorts to aggression. In a restorative conversation, the goal is to move away from blame and shame and toward a mutual understanding of one another. All parties involved in an incident will learn how they contributed to the incident occurring in the hopes that they will move forward from this increased understanding (Costello, Wachtel, Wachtel, 2010). This push for increased social and emotional learning now has a stigma where some feel this is a version of indoctrination. 8 As a society, it appears that the collective is grappling with the idea of safety and deciding which approach is needed to keep everyone safe. Desautels (2012) questioned, "Is life to be creative and pleasurable, or are we to feel increasing suffering and stifling gloom of circumstances as we mature and leave our childhood?" (p. 54). The Alliance Against the Use of Restraint and Seclusion (2022) argued the need to ban the use of restraint and seclusion to teach. Located in Maine, this organization has a strong social media presence, organizes protests, and works to raise awareness of these practices in schools and institutions that are designed to support those with mental health challenges and neurodivergent thinking. There is a movement happening now to continue the conversation that explores what dignity and respect for both looks and feels like (Davis & Oompa, 2016, Yergeau, 2018; Lendzion, 2022). This conversation is gaining popularity in our post-COVID era. In the post-COVID-19 evolving society, many employers are struggling to find people to do the work. Those in social services are seeing a large staff turnover because of the uncertainty and volatile nature of the work potential (Dima, et al., 2021). The physicians who work to support individuals with these challenges have restrictions on how they work that can hinder progress (Redmond, 1998). Educationally, a case conference committee can increase services, which can result in a change in teaching staff midyear to provide those services when a change in placement is decided (Indiana Department of Education Office of Special Education, 2019). 9 These factors and more unite to create a system that provides struggling scholars with the outcomes that society experiences today. Scholars with emotional and behavioral challenges are passed around from staff to staff and supported in a way that is often punitive in nature, from increased interaction with School Resource Officers to the use of restraint and seclusion (Alliance Against Restraint & Seclusion, 2022). The overarching question is, can those working in social services do better? Is there another way to approach this work? What would it look like to create and live in a supportive system that sees results where there are no victims to gain access to the support needed to navigate these realities? What does this mean for the neurotypical community regarding how staff show up and respond to those with emotional and behavioral challenges? Conceptual Framework Adaptive leadership was the conceptual framework used in order to create the change in practices at the specialized, alternative day school program (Heifetz, 1994). The point here is to provide a guide to goal formation and strategy. In selecting adaptive work as a guide, one considers not only the values that the goal represents, but also the goal's ability to mobilize people to face, rather than avoid, tough realities and conflicts. The hardest and most valuable task of leadership may be advancing goals and designing strategies that promote adaptive work (Heifetz, 1994, p. 23). Changing the systematic approach to find better outcomes and responses toward scholars with emotional and behavioral challenges requires responding staff to 10 conduct internal work to enable approaching the work differently. This adaptive leadership approach was the foundation for how to support an evolution in the work conducted in the specialized, alternative day school program. This work requires that the trusted authority figure mobilizes the group and actively seeks to gain perspectives within that group to support the effort when facing challenges. The work of competing values and ideas, especially when navigating fears, is the heart of this endeavor; "Different values shed light on the different opportunities and facets of a situation. The implication is important: the inclusion of competing value perspectives may be essential to adaptive success" (Heifetz, 1994, p. 23). This approach from the trusted authority figure, "places emphasis on the act of giving clarity and articulation to a community's guiding values" (Heifetz, 1994, p. 23). Through ongoing diagnostic work and navigating the holding environment (Heifetz, 1994), the specialized, alternative day school program continues to evolve. Definition of Terms aversive practices - Responses toward behavior that negatively impact a scholar's access to the learning environment and include removals such as outof-class suspensions, out-of-school suspensions, restraint, and seclusion. healthy re-engagement - The desired outcome for both scholar and staff working with the scholar after an escalated behavior which includes positive relationship building and activities that restore. specialized, alternative day school program: This is broadly known as an alternative education program. Specialized indicates that staff are uniquely 11 trained to work with scholars who require a specialized program, as determined by the scholar's case conference committee defined in Article 7 (2019). neurotypical: Neurotypical is a descriptor that refers to someone who has the brain functions, behaviors, and processing considered standard or typical (Resnick, 2021). neurodivergent: Neurodivergence is the term for people whose brains function differently in one or more ways than is considered standard or typical (Resnick, 2021). restraint: Refers to a personal restriction that immobilizes or reduces the ability of a student to move his or her torso, arms, legs, or head freely (CRDC, 2021). scholars: A term some use for students in PK-12, public education who often are working harder than their neurodivergent counterparts. The intent is to acknowledge the work they are putting into their growth, with respect to their individual challenges and long-term goals. seclusion: Refers to the involuntary confinement of a student alone in a room or area from which the student is physically prevented from leaving (CRDC, 2021). Missing from the literature There is very little discussion about the internal work educators need to successfully engage with in order to reduce the use of aversive practices. Much of the work is how to teach skills and target behaviors of concern to the scholar so that responding staff may meet the need of the scholar in a more socially 12 appropriate way. Missing from the literature is the internal journey that responding staff members will experience. It feels simplistic to learn the work of teaching skills, learning more about how the brain functions and how to engage with emotions and mental health. Without this internal work and commitment from the supporting adult, progress may be slowed or halted. Successfully navigating the challenges faced is a journey and requires both sides of the conflict to put in the work. Research Question How did a specialized, alternative day school program in a public school system in Indiana utilize professional learning and feedback gained from diagnostic work to restructure, thereby reducing the use of restraint, seclusion, and injuries that required outside medical attention while increasing academic attainment over the course of seven years? Dependent variable: Incidents of restraint, seclusion, and injuries that required outside medical attention and academic attainment. Independent variables: Professional learning and a collaborative restructuring. Population: Staff and scholars participating in the designated, specialized, alternative day school program located in a public school system in the state of Indiana. Measurement Each data piece is collected through the data management system and supports the specialized, alternative day program in making decisions. Data will 13 be presented as described in the Methods section to help capture the progression of this program. High School credit attainment Attendance rate Occupational Safety and Health Administration (OSHA) reports for injury Number of incidents of level three and level four behaviors reported through incident report review Intensity as defined by ProACT Crisis Intervention (Fox, 2022) of incidents reported through incident report review Interviews from staff who were hired at four identified intervals Prior to Spring 2016 Fall 2016 - Spring 2019 Fall 2019 to Spring 2022 Fall 2022 to Spring 2024 Null Hypothesis Over the course of seven years, professional learning, feedback gained from diagnostic work, and a restructure to reduce the use of restraint, seclusion, and injuries that required outside medical attention or academic attainment have had no impact. Hypothesis There has been a statistically significant, positive improvement from professional learning, feedback gained from diagnostic work, and a restructure to 14 reduce the use of restraint, seclusion, and injuries that required outside medical attention or academic attainment over the course of seven years. 15 LITERATURE REVIEW This review intentionally moves away from the focus on strategies and studies that demonstrate how to do the work of supporting those with emotional and behavioral challenges. Instead, the focus is on the areas of internal work needed to help staff and families move away from a system of control, reward, and punishment and toward a system of regulation, advocacy, and acceptance. The argument presented is that the body of research and systems that exist already provides many of the answers needed to design a program to support these ideals (Kern, et al., 2011; Sprick, 2021). Instead, the failure of responding staff to engage in that system perpetuates the challenges public school institutions find today. The argument proposed is that the work lies internally with responding staff, not with the scholar who is educationally diagnosed as having emotional and behavioral challenges. This literature review will focus on how we got where we are and current research practices that are moving away from that paradigm. When approaching a restructuring of a school program and conducting diagnostic work, the following questions began to emerge. What are the conditions that allow for responding staff to engage in making changes? What internal work could be fostered to implement the structures and practices known to be effective? The framing that was used to explore these topics was adaptive leadership (Heifetz, 1994). Through diagnostic work at the systematic and individual levels, the school staff were able to uncover several conflicting ideas on how to engage. Then, through managing the holding environment, the school 16 staff were able to continuously push the conversations around the work of facing the fears and concerns expressed by staff, scholars, and families. This review will begin by acknowledging the history of treatment for mental health. This foundation allows for a better understanding of the approach to the work of supporting those with emotional and behavioral needs, as these lines of code run in the background of thinking. Then, the literature review will explore the evolution of service models that have been attempted to meet the needs of those with emotional and behavioral challenges within the public school settings. From here, the review will move through the school model intervening from concerns of safety to the demands placed on teachers for academic excellence. The concerns presented here will allow future administrators to understand some of the arguments that will surface during a restructuring of a school program that reframes the work. Finally, the journey will work through questions posed by the community during diagnostic work that enabled the specialized, alternative day school program to transform and what research says regarding those questions and concerns. The intent of this literature review is to provide an understanding of how the system was developed to create the outcomes that public schools experience today and what it takes to confront those concerns. This foundation will serve as the place to begin the internal journey of engaging with this work. History: Where We've Come From It is important to know where society has come from when considering the treatment of those with emotional and behavioral challenges to know how to begin to change. The history of mistreatment of those with challenges provides a 17 foundation that continues to shape how the restructuring of a specialized, alternative day program is approached when these challenges inevitably arise. Mental health services in the United States and the development of programming begin with a lack of understanding and the mistreatment of those born outside of what was considered the norm. These darker times saw those with disabilities treated through restraint, seclusion, instruments of torture, and even death Department of Administrations, Governor's Council on Developmental Disabilities, 2022). The 1700s explored ideas around containment to support those with disabilities. The 1800s normalized pushing those with disabilities out of their communities. Through the 1900s, a shift took place, though these ideas around containment and moving differences away from our communities continue to be pervasive today, as described further below. Legislation was introduced in 1975, the Education for All Handicapped Children Act (PL 94-142), where those impacted begin to take shape, allowing for more voices to prevail within the conversation. This legislation protected the rights of scholars with disabilities and granted access. Later reauthorized in 1990 as the Individuals with Disabilities Education Act and again in 2004 as the Individuals with Disabilities Education Improvement Act (US Department of Education, 2024). The subjects of this research will have qualified, according to the Individuals with Disabilities in Education Improvement Act (IDEIA), in the disability category of emotional disabilities and other health impairments. While this educational designation does not mean that the scholar is labeled with a 18 mental health diagnosis, there can be an overlap. One in five children is eligible for mental health services (CDC, 2022), yet there is no data available for how many scholars are identified with both a mental health diagnosis through a mental health service agency or physician and an education designation of emotional disability. This is largely in part due to the potential violation of the Health Insurance Portability and Privacy Act (HIPPA) (1996) as mental health services are up to the person receiving the services to disclose. It is up to the family to gain access to a mental health diagnosis which will be explored later. Those with mental health challenges have experienced a range of treatment options and possibilities throughout history from state-funded hospitals to outpatient care, acute hospitalization, and residential treatment today as summarized next. It was in the 1700s that professionals began to consider the treatment of these individuals and what possibilities may exist: Beginning in the late 1700s, European hospitals introduced what they called "moral treatment." Doctors, particularly in France and England, discouraged physical restraints, such as shackles or straitjackets. They focused instead on emotional well-being, believing this approach would cure patients more effectively. This intention to treat patients humanely inspired Dr. Thomas Story Kirkbride. Kirkbride directed the Institute of the Pennsylvania Hospital in Philadelphia starting in the 1840s. Using moral treatment ideas, Kirkbride created what he thought was the ideal architectural model for a humane hospital. The "Kirkbride" building model 19 featured several different wings for patients, nurses, and doctors. Gardens and farmland surrounded the buildings. (Meldon, 2017, para. 4) Various religious hospitals with similar views on the treatment of those with mental health needs were created in the United States through the 1800s. Throughout the 1800s, communities "warned out" and "passed on" those with disabilities to keep their communities from including those with differing abilities (Department of Administrations, Governor's Council on Developmental Disabilities, 2022). Families would keep their loved ones hidden away. For those who did not have a family with the means to care for them, living conditions were harsh. Dorothea Dix, a nurse in the 1800s, advocated for individuals with disabilities. She visited asylums and treatment facilities and stated that: More than nine-thousand idiots, epileptics, and insane in these United States, destitute of appropriate care and protection. Bound with galling chains, bowed beneath fetters and heavy iron balls, attached to dragchains, lacerated with ropes, scourged with rods, and terrified beneath storms of profane execrations and cruel blows; now subject to jibes, and scorn, and torturing tricks, now abandoned to the most loathsome necessities or subject to the vilest and most outrageous violations. (Department of Administrations, Governor's Council on Developmental Disabilities, 2022). Institutions and hospitalizations then authorized the use of restraint, seclusion, electrotherapy, and lobotomies, many of which were eventually shut down. 20 Professionals then pursued eugenics and various medicinal treatments for a "more morally guided" answer to the challenges faced (Meldon, 2017). At the height of institutionalized mental health care, society began to question what insanity was. The idea that insanity arose from the body and brain and not from bestial tendencies or an inherent lack of reason gradually led to insanity being a medical problem, allowing it to be viewed in terms of symptoms rather than a degradation of man to beast" (Rojas-Velasquez, 2017, pp. 19-20). Rojas-Velasquez (2017) went on to explain that asylums promised a cure for mental illnesses; however, over time, the idea of treatment returned to confinement. From the 1960s to today institutions that provide for those who cannot care for themselves, due to significant mental health challenges, are closed (Bassuk, Gerson,1978). Bassuk and Gerson (1978) concluded by saying, One must accept the fact that psychiatry is not now able to cure some forms of severe emotional disability and that psychiatry alone cannot assume the broad responsibilities of a society to care for its helpless fellows (1978, p. 53). The history of ostracizing those with emotional and behavioral concerns is wellestablished in societal thinking. It was in the 1900s that professionals collectively moved away from the "moral treatment" of those with mental health needs (Meldon, 2017). It was during this time that doctors pursued hydrotherapy, electric shock therapy, and eugenics. In the 1960s the shift to close state-funded hospitals began. The 1990 Americans with Disabilities Act and later 2008 amendment pushed the 21 conversations of accessibility and that all decisions made would include those for whom the programming is being built. Individuals with Disabilities Education Act (IDEA, 1997) further solidified the conversation that decisions are to be made alongside those with varying challenges, rather than for them. It is through these acts that the pressure for schools to find support and services to help those with emotional and behavioral challenges was born. In the school setting, staff finds evidence that this undercurrent of history resides in the way schools are structured. Most of the rules are arbitrary and often do not apply to society. Examples include when and for how long to use a restroom, access to water, communicating with peers, and how to walk down a hallway. One explanation of this comes from Ashley Davis and Oompa's 2016 National Poetry Slam Finals House Slam titled "Simon Says." Responding staff give directions forgetting humanity and demonstrating that they are unable to put into practice what is needed to teach a new behavior or skill set. Historically, society has treated those with disabilities, especially black, indigenous, and people of color (BIPOC), and individuals with disabilities very poorly (CRDC, 2020). Arguably, there has been progress made with the acceptance of individuals with many disabilities. It is not uncommon to see scholars invited into general education classrooms or schools who are supported for their disability. Dr. Desautels (2016) wrote: We accept and empathize with physical challenges, such as immobility and compromised physical senses. We don't think twice when we observe 22 hearing aids, eyeglasses, wheelchairs, and apparatuses that assist in living, moving, and experiencing the world in "normal" ways. We modify and accommodate instruction for children who learn differently, providing audio books, assistive technology, and continuous emotional support. As teachers, our frustration with how the brain learns rarely reaches the level where we become angry or aggressive toward a student. Yet, with children who experience pain and shame, we struggle to understand. (p. 74). The idea that staff are struggling to understand and accept certain disabilities permeates our current system. Schools have worked to navigate how to include scholars with emotional and behavioral challenges in these spaces. Two foundational challenges around services include one for the teacher in public schools and the other for the mental health community. Understanding these challenges supports changes around the systematic approach for scholars with emotional and behavioral challenges. Teachers in educator preparation schools are exiting these programs without a foundation in the work needed to find success with scholars who have emotional and behavioral challenges (Henderson et al., 2005). Disability categories have been aggregated with no real understanding of what their characteristics and nuances truly mean for a scholar. The focus tends to be on defining the categories with little focus on how an individual interprets the world around them (Indiana Department of Education, 2022). 23 For example, teacher preparation programs teach that a specific learning disability means that a scholar has a discrepancy between their crystalized intelligence score and their achievement, despite attempts at supporting them to close this gap (IDEA, 1997; IDEIA, 2004). What would be helpful is knowing that a scholar has a processing disorder that causes information to have a delay in moving through the varying levels of the brain, back to an external output. The scholar can do it; however, needs extra time to process before responding. This also impacts, for example, the ability to move information from the board to their paper for notetaking. Describing the impact of disability supports a teacher in knowing how to support (McDowell, 2018). Confirming a state-issued definition does not. In 2006, teacher colleges and licensing boards made the change to move disability categories into cross-categorical licensing in an effort to increase licensed practitioners in the field (Oliver & Reschly, 2010). Those licensed before 2002 had some level of training for varying disability categories (Indiana Department of Education, 2022). For example, one could earn a master's degree working with scholars with emotional and behavioral disturbances. After 2002, this disability category was aggregated with all mild disabilities and no longer had specialized training. The focus shifted from knowledge of disability categories to what settings a teacher could practice. This created a learning curve that is contributing to teachers who now run from the field. For those who stay, the resulting effect is that "The field of education is replete with practices that lack research support" (Kern et al., 2011, p. 594). Teachers are directed to figure out 24 what interventions the acting-out scholar needs without formal training to meet this need. Mohr, et al. (1998) offered a description of the lived experiences of our scholars with emotional and behavioral challenges. Because of their unmet dependency needs, emotionally wounded children often perceive the world as hostile. If they come from abusive, neglectful, or chaotic environments these children may have been disappointed, abandoned, or violated by the adults on whom they should have been able to depend. They are often angry, slow to trust, and, as a self-protective measure, they become aggressive and may attack without provocation. The anger serves the purpose of distancing others. This can be an attempt at self-protection. It is also a way to express their intense pain. Fueled by past hurts and stunted by their emotionally impoverished histories, they are more preoccupied with survival than with growth. (p. 96) A foundational description helps the teacher to find an entry point to support the scholar. The system is designed for teachers to feel uncertain about what to do to support the needs they observe. In addition to the complexity of licensing and teacher training, there is a current teacher shortage in the post-COVID world. A quick internet search for the state of Indiana will show that every major news media has a story about the teacher shortage. These include: WFYI, Inside Indiana Business, Tribune-Star, Fox 59, WTHI, WTHE, IndyStar, WBIC, and more! Representatives of the Indiana State Teachers Association and the American Federation of Teachers said the states disinvestment from public education has left schools struggling to 25 attract and retain teachers and support staff at the expense of students (Appleton, 2022). Budget cuts, uncertainty from training programs, and the complex needs of scholars have created the system we experience today. Educationally, service models will vary from local school agencies to local school agencies. School systems will often band together in a cooperative service model to combine their resources to meet the needs of their scholars. After all, it does not make sense for a school system to create a program for two or three scholars who have similar needs. Combining resources to provide services with neighboring school systems so that a program for ten to fifteen scholars is cost-effective and enables the school systems to create better programming options. Unfortunately, this model means that a scholar is moved from their community to a neighboring community in order to attempt to have their needs met. The added time in transportation and the rejection of their community often adds additional stress for that scholar. Little is understood about this impact, though much is cited on the lowered academic demand placed on scholars who have maladaptive behaviors (Oliver & Reschly, 2010; Sprick & Borgmeier, 2010). This system is a modern-day version of the "passed on" response (Department of Administrations, Governor's Council on Developmental Disabilities, 2022). The second service challenge is regarding mental health services. To support a family in accessing mental health services for the struggles they are experiencing, the school considers how to provide these opportunities. DeCarlosSantiago, et al., (2018) provided schools with considerations for implementing 26 mental health support teams within the school setting. These included: how to establish consent, privacy, and confidentiality, how to collaborate with teachers and families, be mindful of secondary trauma experienced by the responding staff, and how to sustain the programming. Schools work to have agreements between their agencies and mental health agencies to be able to provide services to families and scholars within their facility. These partnerships provide families with easier access as therapeutic services can be provided within the school day. The mental health agency is also able to act as a liaison between the school and home in order to meet the needs of that scholar. Careful consideration of the outlined intervention design will help the school agency and mental health service agency to create a partnership that will benefit scholars who struggle with emotional and social skill learning. When considering changes in services, a school system comes up against what mental health services to offer or connect with in their public school institutions. Despite what is offered, not all families will choose to engage in these services. The choice to engage in services is a personal journey for families to embark on and no two stories are the same. In the same way that teachers will undergo varying levels of engagement with learning to support different scholars so will families. Rockhill, et al., (2013) reviewed how caregivers distress on the effectiveness of the intervention. In their study, the authors suggested that additional research is needed in addressing the distress of caregivers for scholars who are diagnosed with multiple mental health diagnoses in underserved communities. Another interesting finding from this 27 study suggested that caregivers working with children who had multiple diagnoses supported by mental health saw an increase in stress and depression. This suggests that working with scholars who have a higher rate of maladaptive behaviors adds additional strain to those who are working to help them navigate their world. Families are engaging in these challenges with or without services. The prevalence of children seeking medical support for their mental health conditions is on the five (Olfson, et al., 2014). While one in five children could qualify for some level of service in mental health, only about 20% of those families are accessing services (CDC, 2022). Access to services is fraught with challenges that can include lack of funding, lack of qualified individuals to provide care, labeling, stereotyping, lack of a provider who will work with the type of insurance a family holds, lack of flexibility in accessing the services, and a denial of the need for services (Shah & Beinecke, 2009). Educationally, a school can work to provide a label for school services through IDEIA as previously described; however, this is not the same as providing mental health services. The disconnect between what a school can provide a scholar and what a family needs to support their scholar with mental health challenges is often disappointing. Schools can provide access to school-based services through specially designed instruction or in a partnership within the school; however, schools are unable to require either mental health or educational services. It is ultimately up to the family to connect with mental health services and allow responding staff to engage in specially designed instruction. Staff then design 28 and provide specially designed instruction. This, however, becomes complicated as teacher training programs are not required to provide this training. Law: Disability and not a challenging moment in time. When approaching the work of restructuring to support those with emotional and behavioral challenges, it is important to have a foundational knowledge of the laws that govern this work. One staff member experiencing a challenge with one scholar does not warrant the label of a disability or mental health challenge. One scholar having a challenging year due to events in their personal life does not warrant a disability or mental health label. These protections are implemented to prevent a scholar from unnecessarily gaining a life-long label. Any program modification must first have this foundational understanding as a protection for those it will serve. The Individuals with Disabilities Education Improvement Act (IDEIA) is where governance for how those with different abilities are to be supported in the school setting. Of the twenty-one categories, emotional and behavioral conditions are where this research will focus. The state of Indiana designates this category as an emotional disability and sometimes as other health impairments based on an educational evaluation and case conference committee decision. Being identified as having an emotional disability does not require an outside diagnosis of a mental health condition (Indiana Department of Education Office of Special Education, 2019). Instead, a multidisciplinary team must show that the scholar exhibits one or more of the following five criteria to a marked degree and over a long period of time: 29 Sec. 300 (c) (4) (a) an inability to learn that cannot be explained by intellectual, sensory, or health factors; (b) an inability to build or maintain satisfactory interpersonal relationships; (c) inappropriate types of behaviors or feelings under normal circumstances; (d) a general pervasive mood of unhappiness or depression; and (e) a tendency to develop physical symptoms or fears associated with school or personal problems. (IDEA, 1997; IDEIA, 2004) Outside of these educational laws to support learners identification of educational services there are opportunities for mental health services. These services are provided by outside institutions, often in partnership with school agencies. The prevalence of children who qualify as having a mental health condition is one in five (CDC, 2022), though not all with a mental health condition qualify as needing special education services as described above. This difference between mental health and educational services can cause a disconnect for scholars with complex needs and confusion for families who must navigate these systems. Educationally, the continuum of services under Article 7 of Indiana State Law indicates that scholars with greater intensity of need will have increased services from within the school agency. The federal code for this is indicated through the least restrictive environment. The least restrictive environment (LRE) is defined in Section 1412(a)(5) of IDEA (1997) is defined here: (5) Least restrictive environment 30 (A) In general to the maximum extent appropriate, children with disabilities, including children in public or private institutions or other care facilities, are educated with children who are not disabled, and special classes, separate schooling, or other removal of children with disabilities from the regular educational environment occurs only when the nature or severity of the disability of a child is such that education in regular classes with the use of supplementary aids and services cannot be achieved satisfactorily In addition to protections around the identification of an emotional disability and services that can be offered, there are protections around how discipline can be applied. Suspension leads to a disruption in services and needs careful consideration before applying. The primary focus of this research will be on a public school separate day program for scholars who are identified by a case conference committee as needing services that outstrip what can be offered in the public, general education school setting and its journey toward change. This is also known as a federal indication of LRE 53, separate day school. Again, this is a systematic approach, reminiscent of the "passed on" mentality of the 1800s (Department of Administrations, Governor's Council on Developmental Disabilities, 2022). School and Classroom: Attempts at services From a basic foundational understanding of society's history in the mistreatment of those with mental health challenges and the laws to protect them, a school system begins to consider what structures and services to offer 31 for those who are in need within their respective communities. Teacher preparatory courses (described in the History section) have shifted through the last twenty years to move away from the knowledge of disability categories in response to the licensing changes implemented by the Indiana Department of Education (2022). Upon completion of their courses, a teacher entering the field or new to working with scholars who have emotional and behavioral challenges does not need to go far to learn more about what to do in order to support them toward healthy engagement. Each school system will research and choose which curriculums and support systems they will engage with on behalf of their community. Moore (2016) compared bowling to teaching; she explained that professional bowlers never aim for the middle pins. Instead, professional bowlers aim for the two most difficult pins, the 7-10 split. Similarly, when teachers plan with an aim for the challenges learners will encounter, they are able to design a learning environment that is inclusive for all. In collaboration with Dr. Rick Van Aker and modeled after Dr. Michael George and Dr. Lee Kern's work with Centennial School, the specialized day program conducted diagnostic work to restructure with the aim to better support scholars with emotional and behavioral challenges. Five overarching areas in the design of that system include: structure, instructional practices, crisis response, connection, and restoration. These will be explored next. Structure 32 Multi-tiered systems of support (MTSS) are more than just tiering interventions and data-based decision-making. It is the structure for how day-today activities occur within a public school institution. Indiana Department of Education (IDOE) notes six areas of development, including: "Leadership, Capacity Building Infrastructure, Communication and Collaboration, Problem Solving Processes, Multiple Tiers of Instruction and Intervention, and Data Evaluation" (Larrison & Mansfield, 2018). Through careful consideration of each of these areas, procedures for the operations of the school can be demystified, which helps the responding staff have clear expectations and smooth daily routines. Once guidelines and procedures are made clear, responding staff can turn their attention to helping teach school procedures and expectations for the scholars they will serve. School personnel often turn to positive behavior support systems for this structure. Especially in secondary schools, responding staff make assumptions that scholars know what is expected of them within the learning environment (Sprick & Borgmeier, 2010). A curriculum for making expectations clear, available through CHAMPS, which stands for Conversation, Help, Activity, Movement, Participation, Success, will help with the flow of routine development and ensure that scholars know what is expected of them for each phase of learning (Sprick, 2021). Through the development of the environment and clear, predictable routines, the staff are empowered to teach how to find success in the learning environment. 33 Eric Jensen (1998) argued that the focus needs to be on the elimination of negative interaction within schools and that the environment of public school institutions is more important than other factors such as an IQ score. This is argued because the brain can be manipulated to create new neural pathways through environmental conditions, where previously, the mind was believed to be fixed. It is now understood that the brain is capable of more when nurtured in an environment that reduces threats. Jensen (1998) explained that "Once threats are gone, we can go to work on the enrichment process" (p. 30). Stress, including social stress and threats, can negatively impact the ability, to form complex neural pathways, which can lead to a reduced ability to process the environment and language (Jensen, 1998, pp. 53 - 55). Additionally, Jensen (1998) pointed out: [T]he old paradigm of behaviorism told us that to increase a behavior, we simply need to reinforce the positive. If there's a negative behavior exhibited, we ought to ignore or punish it. This is the "outside-in" point of view. ... This approach says that if demotivation is an established condition then there are causes and symptoms. This way of understanding classroom behavior seemed to make sense for many. But our understanding of motivation and behavior has changed. Tokens, gimmicks, and coupons no longer make sense when compared with more attractive alternatives. (p. 65) Yet, in classrooms today, this continues to be the practice. Teachers use charts that clip up and down to teach and manage behavior and mark a citizenship 34 grade. These often have consequences attached in line with a punishment system. Attaching a reward for staying in a certain space all day is in line with a reward system. Alternatively, Zones of Regulation (Kuypers, 2011) is a curriculum to teach scholars how to identify how they are feeling and then apply a strategy to help with that emotional state. The human experience will fluctuate throughout the day. The reward and punishment systems leave little space for basic human emotion. Careful consideration must be given to how feedback is provided to each individual scholar with a focus on teaching, not conditioning. Dr. Candice Pert was a neuroscientist who focused on the discovery of the science of emotions. Pert (1997) explained that "When emotions are repressed, denied, not allowed to be whatever they may be, our network pathways get blocked, stopping the flow of the vital feel-good, unifying chemicals that run both our biology and our behavior" (p. 273). Jensen (1998) provides guidance on how to engage emotions in a productive manner, including: Movement, Stakes, Novelty, Sharing, Apprenticeships, and Thinking Big (pp. 94-95). With each, the goal is to provide scholars with an opportunity to engage using their emotions in a productive, guided, and safe space. Desautels and McKnight (2019) explored this in both Eyes Are Never Quiet (2019) and Connections over compliance (2020), which provided lessons and a curriculum to support the classroom. Each text provides activities to engage in, which help regulate and teach about the emotions scholars and responding staff are experiencing. Another area that seems to be implied but not always directly spoken about in the literature is that the above research on what is known about how the 35 brain functions and the need for regulation apply directly to adults, too. Adults need a working knowledge of the brain and how it functions to identify how they enter a space and what they need to regulate and be productive. Adults need a productive way to process their thinking, especially when it comes from that reaction space of stress. Adults need novelty, autonomy, trusted working conditions, and the ability to do what they need to regulate themselves. Failure within school systems to support responding staff in providing these opportunities to teachers is one contributing factor to outcomes experienced today. Dysregulated adults are no good for dysregulated children (Desautels, 2020). Responding staff have an opportunity to counter this within school systems by providing regulation opportunities directly for staff. By doing so, teachers are able to fully embrace and understand how to teach this to scholars. INSTRUCTIONAL PRACTICES Once there is a foundation for the overall structure, including MTSS, and how to teach emotional health, the educator can turn their attention to lesson design (Marzano, 2007; Fisher & Frey, 2008), and structures for teaching how to work in small groups and maintain an academic conversation (Kagan, 2015). While many scholars who have behavior challenges also have struggled with performing in the classroom, this is often misinterpreted as a cognitive or learning challenge rather than a lack of connection and access due to the intense emotions the scholar is experiencing. In the state of Indiana, there were 12,202 scholars identified with an emotional disability as their primary disability in 2021 (Holsapple, 2021). While there is no information on the intellect of these scholars 36 because of how the data is collected and reported, many scholars with behavioral and emotional challenges have average intelligence and are often capable of working at or above grade-level performance standards in the classroom (Sec. 300 (c) (4) IDEA, 1997; IDEIA, 2004). Not all scholars are given an intelligence equivalence test as part of their educational assessment to determine eligibility, as this is not a part of the criteria for consideration for an emotional disability. It is, therefore, difficult to know what challenges and strengths a scholar with an educational designation of emotional disability brings into the learning environment. Scholars with emotional and behavioral disabilities often have learning challenges. Because of the lack of evaluative information, it is difficult to substantiate where the challenge comes from, as well as a lack of instruction from disruptive behaviors or intellectual abilities (Oliver & Reschly, 2010). Many staff and community members hold an association that being identified with special education services as having a low intelligence (Kauffman, & Bader, 2013). Without qualified educators to teach skills for the emotional and behavioral challenges a scholar experiences, a lack of engagement in the classroom will create a learning performance gap (Henderson, et al., 2005). Supporting scholars back to healthy engagement with the classroom is the goal that gets lost and is very difficult without the proper steps in connecting with the scholar, their family, and an understanding of how the brain learns new materials (Jensen 1998; Desautels, 2016). 37 How information is presented and structured will determine how a scholar, especially a struggling learner, is able to engage with it. Jensen (1998) and Marzano (2007) explored the importance of supporting scholars in making connections across content areas: In the classroom, it's the ability to see ideas in relation to others as well as how individual facts become meaningful in a larger field of information. Help students see how economics relates to geography, how mathematics links to art and music, and how ecology links to science and politics. Through discussion, arts, or visual thinking, students can make important, meaningful patterns. (Jensen, 1998, p. 96) Giving scholars time to explore the ideas that are interconnected within all that they learn supports the brain in solidifying the learning. This includes connection through emotions. Without this time to process and make sense of their world, learning remains compartmentalized and often disconnected from the experience. Dismissing a scholar from a lesson or a class to address their emotional and behavioral challenges takes away the opportunity for input. Marzano (2007) outlined the optimal way for delivery of input for the brain to receive information: Small chunks of content must still be actively processed by students. This requires the use of a set of interacting instructional strategies. In effect, no single instructional strategy will suffice to meet the demands of actively processing content during a critical-input experience. Rather what might be thought of as macrostrategies must be employed. (p. 34) 38 These strategies included: "(1) summarizing and note taking, (2) nonlinguistic representations, (3) questions, (4) reflection, and (5) cooperative learning" (Marzano, 2007, p. 35). With processing time and time to manipulate the information scholars, are best able to retain new learning. When a scholar is resistant to classroom activities or struggles to remain within the norms of behavior, the teacher has a responsibility in finding ways that they can process the learning that supports their needs. Teachers attempt interventions and support to make learning accessible. When the intervention is ineffective and maladaptive behaviors increase, the continuum of services is explored as a possible next step for support. Unfortunately, teachers lessen the demand and water down curriculum as a means for control of the classroom setting (Hammond, 2015). Anytime a teacher chooses to address behavior by dismissing the focus on academics, scholars lose. One must support and address both academics and behavior simultaneously. After all, learning creates disequilibrium. Disequilibrium creates stress (Jensen, 1998). Managing stress is a part of what a scholar with social and emotional learning challenges is navigating (Mohr, et al., 1998). CRISIS RESPONSE Safety is at the forefront of the minds of many at gathering places. In the same way that one is encouraged to know where emergency exits are in a theater or on an airplane, public school institutions practice various safety drills. One part of this is having a crisis response team who regularly practices how to respond if a scholar escalates to a point of needing additional support. Often, 39 part of this training includes how to use restraint and seclusion as a response toward someone who has become unsafe to the point of injury that could require outside medical attention. Careful consideration is needed when choosing a crisis response training system. The use of restraint and seclusion requires that institutions that work with children follow the fifteen principles outlined by the US Department of Education (Duncan, 2021). The practice requires institutions to self-report to the Office of Civil Rights Education Department as part of their oversight of this response. It is questionable how much underreporting occurs. The acknowledgment that responding staff are a contributor to aggression in scholars takes a lot of courage and vulnerability (Rawcliffe & Wellman, 2002). Centennial schools demonstrated this vulnerability when they made the shift away from restraint and seclusion (Lurye, 2020). There are scholars escalating with one another that are moving into a response of violence (Musu, et. al., 2019). When a scholar who has a history of struggling with social and emotional learning has conflict with a peer, this can lead to feelings within the teacher to protect both scholars when navigating an especially challenging situation. Roseth, et al. (2008) studied teacher intervention with preschoolers who were in peer conflict. They found that the teachers' intervention often disrupted the cycle of peer resolution that would evolve naturally from peers working out their challenges on their own. It is noted that when aggression is involved, teacher intervention is a necessary step. What happens in response to this aggression is crucial to how the scholar will move 40 forward from this experience. Keeping scholars isolated from one another prevents the opportunity to restore the harm and grow from that experience (Peace Learning Center, 2021). It is crucial that teachers intervene in a way that facilitates the conflict rather than direct what is needed (Roseth, et al., 2008). Cognition drops as individuals are escalated (Fox, et al., 2020). This can mean that while the scholar's age is one that would not necessarily see physical aggression as a response, their developmental age shifts in escalation, and suddenly a scholar reverts to using physical aggression as a means to meet a need. Roseth, et al., (2008) research suggested that how the teacher responded to this indicates whether the lesson of how to handle those escalated feelings is learned or will perpetuate. In response to violence, school systems are providing training to responding staff for crisis intervention at an increasing rate (Musu, et al., 2019, p. 114). Couvillon et al., (2019) and Wei et al., (2010) reviewed considerations for training provided to school staff when looking for a crisis management training curriculum. This training sets the tone for how one approaches this response. In worst-case scenarios, responding staff fear escalating a scholar to a point of conflict that leads to death. This can happen when a scholar is restrained and later loses their life as a direct result of the restraint. Death from restraint occurs immediately in the event, or a day after when oxygen levels dropped, and medical staff were not alerted in enough time to provide treatment (Fox, et al., 2022). This potential will be explored more later. 41 Staff response is arguably the most important aspect to an escalated scholar for a variety of reasons. A scholar who will escalate to a point of physical aggression is infrequently modeled appropriate ways to manage stressors in their lived experience. Teachers have an opportunity to both model and teach alternatives to aggression in moments of stress and disequilibrium (Sprick & Borgmeier, 2010). The scholar benefits from this modeling through the mirror neurons that are always in action (Ramachandran, 2009). Over the course of time and practice, new neural pathways are forged, and the scholar has additional ways to respond. This happens with great intentionality and often begins with the teachers, family, and mental health providers who work with that scholar. As Stephen Lendzion said during ProACT training sessions, No matter what comes at me from these residents (children), what comes out of me must go through a filter of dignity and respect (personal communication, August 22, 2022). Stating that responding staff response will move "through a filter of dignity and respect" (Lendzion, 2022) sounds easy and reasonable enough; however, it takes great discipline and internal work. Teacher and practitioner training seldom delve into this kind of work. It is assumed that experience and on-the-job-training will be enough; however, all one needs to do is a quick search on the internet to see incidents where teachers have lost their cool and done something regrettable in response to an escalation in a classroom. Thinking that responding staff will just pick it up in time is not enough. Adams, et al., (2019) studied teacher responses as it applied to the comorbidity of autism spectrum disorder and 42 anxiety in scholars. Often, teacher responses were coded as "anxiety provoking." The findings suggested that teachers respond differently to scholars who have this comorbidity versus the ones who only have a diagnosis of anxiety, for example. This is the reason training must focus on the responding staff's respective internal work. It is a lot to ask of a staff member to think that they can find ways to work with scholars who are operating from this space which can lead to aggressive behaviors. ProACT (2022) teaches that there are ways to disrupt the pattern of escalation so that this moment does not repeat. Much of this work comes from the practitioner's internal work and intentionality in setting up the environment, knowing the history of the scholar, knowing the dynamics of the family, and meeting needs. The move is away from a system designed on control, punishment, and a behaviorist approach and toward one of respect, dignity, autonomy, and creating an environment for all to thrive. RESTORATION Restoration covers a wide range of ideas from restoring self after a stressful encounter to allowing each participant in the conflict to come together and restore any harm in their relationship. Careful attention to all areas of restoration is a shift that groups such as the Peace Learning Center are pushing toward. In restoring harm, a professional may uncover some underdeveloped social skills where additional instruction and practice are needed to help that individual know how to be within a space (Goldstein, 1988; Green, 2016; Desautels, 2016; Desautels, & McKnight, 2019). One may find avenues to help 43 the individual apply teachings on how to identify regulation and their strategies for managing emotions through Zones of Regulation (Kuypers, 2011). Engaging in restorative practices requires movement away from reward and punishment. Restoration includes understanding relationship building and regulation. Facilitating conflict, restoring harm, and moving forward takes discipline and a community willing to try a different approach: Key goals of restorative discipline To understand the harm and develop empathy for both the harmed and the harmer. To listen and respond to the needs of the person harmed and the person who harmed. To encourage accountability and responsibility through personal reflection within a collaborative planning process. To reintegrate the harmer (and, if necessary, the harmed) into the community as valuable, contributing members. To change the system when it contributes to the harm. (StutzmanAmstutz & Mullett, 2005, p. 10) Relationship building and relationship maintenance are at the heart of restorative practices. It requires that a system be put into place that considers solutions and restoration as a part of the continuum for consideration when there is a misbehavior or infraction against the school community. The focus is on nonviolence and teaching rather than punitive responses and consequences. (Stutzman-Amstutz & Mullett, 2005). Ultimately, the goal is to bring the scholar 44 back into the community with a plan for support. This moves away from the "passed on" line of code embedded in the collective (Department of Administrations, Governor's Council on Developmental Disabilities, 2022). Holding structured conversations to restore harm falls on a continuum from affective statements through formal conferences (Costello, et al., 2010, p. 12). These conversations, or circles, can be proactive, responsive, involving only staff, or include families (Costello, et al., 2010, p.12-13). Costello, et al., (2010) continued to explain that the structure of a circle has agreed-upon norms to keep the conversation productive. There is an identified facilitator who is viewed as a neutral participant and generally hears from each member prior to entering the conversation. Holding circles is one option that some schools are exploring as an alternative to punitive and assigned consequences for escalations and infractions at school. Effective restoration requires self-regulation and connection. Long, et al., (2001) and Green (2014; 2016) discussed the importance of doing the work on behalf of a young person struggling with behavior, alongside the learner rather than to them. Long, et al. (2001), discussed the conflict cycle as: Thinking Patterns Trigger Stress, Stress Triggers Feelings and Anxiety, Feelings Drive Behavior, and Inappropriate Behaviors Incites Others Reactions (pp. 27-34). They go on to provide a framework for working with a scholar to support them in breaking this cycle and ultimately not fall into its patterns. Similarly, Green (2014; 2016) provides a framework for working with learners to address lagging skills and unsolved problems. This framework 45 supports staff in identifying and targeting where energy will be focused to support the scholar through a few challenges at a time. As improvements are made, additional lagging skills and unsolved problems are then identified and targeted. Using these frameworks as part of the work to restore and return to the community supports the school in breaking the cycles of maladaptive behaviors that can lead to aggression. CONNECTION Brains work in a way to support one another in times of escalation, stress, sadness, and overall discomfort. Siegel (2010) stated that "We are hardwired to connect with one another thanks to our mammalian heritage" (p. 17). It is how the human species has survived this long! The need for connection and the use of mirror neurons allowed individuals to observe one another and learn from observations and interactions (Ramachandran, 2009). The energy that is carried between one another speaks louder than words! Dr. Siegel explained how escalations look within lived experiences (2010, pp. 26-30). This included: 1) bodily regulation, 2) attuned communication, 3) emotional balance, 4) response flexibility, 5) fear modulation, 6) empathy, 7) insight, 8) moral awareness, and 9) intuition" (Siegel, 2010, p. 26). He argued that when these areas are working together, alive, and well, individuals are at their best. When any of these areas fall out of sync with one another, the feeling is of dysregulation and disconnection. When experiencing this state, individuals begin to disconnect from others and pull away from things that will help. Responses become short, with a 46 tone of disapproval and often are inappropriate. This keeps the brain from being integrated and able to productively grow (Jensen, 1998; Siegel, 2010). In Siegel's hand model of the brain (described later), the middle prefrontal area is what allows for connection with others. The "resonance circuits including mirror neurons enable [others] to resonate with [our] own reactions" (Siegel, 2010, p. 138). "The resonance circuitry not only allows us to feel felt and to connect with one another, but it also helps to regulate our internal state" (Siegel, 2010, p. 138). It is in this regulated internal state that responding staff are most effective in the support provided to an escalated scholar. Siegel went on to explain that it is in these moments that individuals can fully connect and support one another, creating lasting change through the brain's hardwiring (2010, p. 139). Christakis and Fowler (2009) discussed the importance of social networks and connections. Christakis and Fowler (2009) summarized how individuals "catch" the emotions of others (p. 35). This has served evolution well through connection and sharing information about the safety of the environment (p. 37). Educationally, responding staff work to support the large emotions of scholars to help them navigate their world and make sense of their options. ProACT (Fox, et al., 2022) stresses the importance of a team intervention at times of escalation with a scholar, as well. The idea of a community gathering to support the needs of its members is as old as the idea of tribes and civilization itself. Dr. Siegel (2010) explained that as a therapist it, is important to pay close attention to emotional contagions. It requires that he keep a clear distinction 47 between himself and the people he serves. Siegel explained that "When resonance literally becomes mirroring, when we confuse me with you, then objectivity is lost. Resonance requires that we remain differentiated - that we know who we are - while also becoming linked" (p. 63). ProACT (Fox, et al., 2022) teaches the importance of restoring after an escalation with a scholar, in part for this very reason. The goal is to help teachers to keep from being overly invested in, or drawing into the emotions of, the scholar who has experienced an escalation and later misinterprets a minor escalation as something more (Siegel, 2010. p. 62). Keeping that separation supports the scholar in regaining control and the teacher in actively maintaining objectivity to remain connected in a healthy way with that individual in need. Dr. Desautels (2012) stated that she does not understand the switch that happens as we grow from "children living carefree into "adolescents and adults who... float in pools of self-perceived stress and hopelessness" (p. 55). There is an overall sense from the educational system that with increased testing, demands on what a scholar should learn and preparation for the workforce is more about learning to live within a stifling system of control than to experience the arts, music, and play that brings forth so much joy and creativity to the human experience. Authors such as Goldstein (1988), Long (2001), Green (2014, 2016), and Desautels and McKnight (2016) have questioned how it is that staff came to a place where it is believed that punishment is the only means for supporting the growth of youth. Regarding internal work needed, Dr. Dan Seigel (2010) explained: 48 The internal state of others from joy and play to sadness and fear - directly affects our own state of mind. The contagion can even make us interpret unrelated events with an uncertain bias - so for example, after we've been around someone who is depressed, we interpret someone else's seriousness as sadness... Our awareness of another person's state of mind depends on how well we know our own - we notice the belly filled with laughter at a party or with sadness at the funeral home. This is the main reason that people who are more aware of their bodies have been found to be more empathetic. (pp. 62-63) A staff member who works with a scholar who has emotional and behavioral challenges is impacted by the feelings that they present throughout the day (Shanker, 2017, p. 21). The same responding staff member who intervenes with a scholar who has an escalation, if not very careful, will misinterpret a minor or low-level behavior as one that is equally escalated through this phenomenon (Siegel, 2010). This is also explained through the ProACT (Fox, et al., 2022) Crisis Management training where teachers are warned that they are in danger of re-escalating a scholar when scholars are coming down from an episode but are still engaged in acting out behaviors known as recovery. Christakis & Fowler (2009) discussed how "Behaviors also spread, and many of these behaviors have big effects on your health" (p. 105). If a teacher is going to carry the stress of a scholar that they are working to support into their personal lives, they, too, will have challenges to navigate. This can compound and impact mood and as the teacher enters the next workday from the following 49 days escalation (Fox, et al., 2022). It takes someone who is in tune with their bodies and how they are entering into and remaining in a space to effectively respond to this cycle. The discipline and internal work needed to be successful here is only brought up lightly by each of these authors. Desautels (2020) goes on to summarize how the brain works when it comes to stress and trauma: I liken this resistance to our innate negative brain bias, which has been a part of our earliest biology because of the survival instinct. The brain's foremost purpose in our lives is to help us survive. Unless we feel safe, we cannot move forward with connecting and developing. (p. 3) It is important to understand that the feelings and emotions that rise are just a part of the human experience. Knowing what to do about them and how they serve is a crucial part of success. There also needs to be an understanding of how this impacts self, colleagues, and the community. While there are similarities within human experience, there are also nuances and variables for how others experience those emotions and feelings. Taking time to connect and get to know one another is where success in intervening and supporting healthy engagement is found. REFLECTION What is missing from the frameworks presented are what it takes for responding staff to engage in them appropriately and successfully. Long, et al., (2001) provide prerequisite skills needed by the scholar to engage in their framework. A few of these include being aware of self, events and other people, 50 self-regulation, common language, and so on (p. 10). The internal work needed from adults who are navigating these conversations with the scholar is not discussed. What about the adults needing to unpack triggers that they find annoying but do not harm others? Or the implicit bias that is carried into an interaction? Hammond (2010) discussed reflective tools that can support reflection on implicit bias. These included: 1. Spending some time viewing the replay in your mind. 2. Make a list of your assumptions, reactions, and interpretations of behaviors as a scenario replays. 3. Try on alternative explanations. 4. Check your explanations. 5. Build your cross-cultural background knowledge. 6. Leverage technology and watch positive movies or television series that will allow you to virtually step into another cultural experience. (pp. 61-62) Without careful consideration into how responding staff show up in a conversation with a scholar processing an escalation, there is a potential to do more harm than good on behalf of that scholar. Responding staff must actively set aside judgements and opinions in order to be present. Being a reflective practitioner willing to do this internal work requires vulnerability and an openness to see the challenge from multiple perspectives. Staff experience trauma or have retriggered trauma responses after responding with the use of restraint (Mohr, et al., 1998, p. 99). Forcibly putting hands on another when not in a life-threatening situation but rather under pressure from 51 outside forces causes undue stress and puts everyone at risk. Bonner, et al., (2002) conducted a pilot study where they debriefed with individuals after an incident of restraint to see what insights could be gained. In this study, both victims of restraint and the responding staff members who utilized this response indicated that they were triggered in the moment by past trauma they had experienced. They further explored the factors that led up to the incident of restraint. All involved identified staff response as a contributing factor and included in that a difference between those staff who were considered regulars and those who were considered temporary. Individuals found the debriefing process to be helpful in supporting their understanding of everything that led up to that moment for the decision to restrain. Access: Effective interventions There is a body of research that contributes to the overall conversation on how to successfully design and implement interventions, but there is little application (Oliver & Reschly, 2010; Sprick & Borgmeier, 2010). Intervention design requires an iterative process. Five steps for this process are offered including, "1) Initial Intervention Development, 2) Preparation for Implementation, 3) Implementation, Feedback, and Revision, 4) Data Based Refinement, and 5) Further Refinement with Divergent Sample" (Kern, et al., 2011, p. 609). The playbook for how to provide scholars meaningful learning experiences while meeting them where they are and providing the system that they need is arguably not the challenge. There are educators in every school who exemplify the above. 52 Barriers such as time, guiding principles of leadership, professional learning opportunities, and state-mandated requirements contribute to creating systems and structures within the school and classroom environment which then result in staff who choose to or not to engage with the learning (Oliber & Reschly, 2010; Sprick & Borgmeier, 2010). Aldrup et al. (2018) reviewed how teacher exhaustion can lead to an increase in viewing misbehavior as a stress factor with increased sensitivity to misbehavior. When stress increases, individuals lose the ability to think clearly (Fox, et al., 2022, Ch.6 p.4). As teacher stress increases, this causes lowered cognition from oxygen reduction which compounds to see greater and greater struggles in the classroom (de Ruitera et al., 2020). The paradox that is created within the group suggests that teachers see no way to reduce the disruptions and misbehaviors within the classroom, making the natural progression to consider that it is the scholar who needs to be removed (Smith & Berg, 1987). As this stress increases, individuals are unable to reach their prefrontal cortex where creativity for solving the challenges within the classroom lives (Jensen, 1998; Siegel, 2010; Shanker, 2017). This disengagement leads to an inability to engage the systems and learning that could improve activities and structures within the classroom. Without these increases in structures and improved activities, scholars continue to misbehave more and more. The cycle continues and becomes more challenging (Fox, et al., 2022). Siegel (2010) explained the importance of parents, or in this case, a teacher's, attunement to a developing mind (p. 129). It is important for adults to 53 be attuned to a scholar's emotional state so that the responding staff can intervene and support earlier. This emotional attunement and early intervention or support teaches the scholar how to regulate and to be connected with a community who will support because the community is a trustworthy place. Approaching internal work: Self-regulation informs the staff response and how responding staff members intervention will land. Dr. Shanker (2016) works with parents and teachers on the importance of self-regulation. When the energy from the responding adult is tense, stressed, or out of sync with the child, the escalation grows larger. Dr. Shanker (2016) discussed the importance of supporting adult regulation first. He further identified five pillars from which to consider selfregulation: Biological, Social, Prosocial, Cognitive, and Emotional (pp. 7477). Each of these pillars supports the adult in helping the child to co-regulate. By identifying and addressing stressors, the crisis response never happens or lessens. McGonigal (2013) teaches how to make stress productive, the importance of how individuals frame stress, and the bodys response to stress. The simple activities of how an individual thinks of stress, holds their posture, and considers breathing all contribute to how stress is managed for responding staff. This applies to those practitioners who work with scholars who may be prone to aggressive behavior to get their needs met. After a stressful event or series of events with a scholar who escalates to a point of physical aggression, it is important for responding staff to take the time to care for themselves and 54 recenter. Failure to do this can lead to retriggering the crisis event and continuing the escalation longer (Fox, et al., 2022). ProACT (Fox, et al., 2022) teaches the importance of taking care of self through a series of exercises in its training to address this need in the short-term and long-term to reduce the effects of stress. In addition to restoring after an escalated situation, the work of those who support scholars with emotional and behavioral challenges can experience secondary trauma. Secondary trauma is the stress that is gained from hearing the stories and experiencing the stresses that others experience (see discussion of emotional contagions above). Baker, et al., (2020) provided six techniques for managing secondary trauma for human rights investigators. The six techniques included: process with a team who understands the work, limiting exposure to graphic content, drawing boundaries between work and personal life, bringing positivity into the work, receive guidance from veterans of the work, exploring a variety of techniques to find what works (p. 301). How stress is processed, and secondary trauma supports the responding staff to invest and remain in the field. Again, what is missing from the literature is the internal work the teacher must undergo to put all of these elements into practice. It is not enough to say that this is what is needed. Bermejo, et al., (2013) explored how teachers experience the demands of their work through the lens of stress. Their findings suggested that increased engagement could effectively prevent teacher burnout. It stands to reason that if a teacher is actively seeking to improve their craft and engage in additional learning on how to best navigate the challenges affecting 55 their overall happiness in their work, they will see improvements in their performance and thus the performance of their scholars. This overall increase in effectiveness then feels good and decreases the stress that is negatively impacting performance. McGonigal (2013) discussed how the body prepares the individual for stress and how perception of this will positively or negatively impact the responding staffs ability to engage. Again, the paradox suggests that internal work impacts the ability to engage and thus be effective in the work when supporting those with behavioral and emotional challenges. Working through staff push back When teachers push academics, or a classroom conflict occurs, and an escalation begins, a teacher can feel either fear or frustration. Emotional contagions, if not managed, impact response (Pert, 1997; Christakis & Fowler, 2009). That fear or frustration leads to concerns for safety (Musu et. al, 2019). A narrative is written in the feeling that can lead to a worst-case scenario (Fox, et al., 2022). Collectively, community members are concerned for safety. The concern for safety has entered the school settings and became a major topic of discussion and consideration (Musu et. al, 2019). With identified emotional and behavioral challenges gaining more conversation in the media, there is an overall sense in communities that anyone could be the next to end up the victim in a tragic scenario, or to be the hero of the story in which the tragic event was stopped. School systems are increasing the use of School Resource Officers in an effort to help keep campuses safe. In a report from the American Civil Liberties Union (2019), "14 million students are with police but no counselor, 56 nurse, psychologist, or social worker" (p. 4). This model suggests a wait for the victim approach for support to scholars with emotional and behavioral challenges, which indicates that the system is giving exactly what it is designed to give - victims. When working with scholars who have emotional and behavioral challenges, struggles arise that can include concerns around safety of self and others. Concerns include loss of life, injury, and litigation. When using restraint as a response, staff need to be made aware of the risks to this intervention (Fox, et al., 2022). Scholars in the care of mental health institutions and in school settings have lost their lives from asphyxiation in or after a restraint. There have been cases where a scholar lost their life days after the restraint from an untreated cardiac episode or from lowered oxygen in the body from the stress of that incident. Autopsies were later able to identify that the restraint was the cause of the death. Nunno et al., (2021) shared that, "The leading cause of death among the fatalities discovered from 1993 to 2018 was asphyxia" (p. 5). Nunno et al., (2021) went on to share that, Of the fatalities, 56 were boys with a mean age of 14.4 years, and 23 were girls with a mean age of 15.0 years. We were able to verify the race/ethnicity of 51 child fatalities, while information on race/ethnicity was not available for the remaining 28. (p. 5) Children such as, Cornelius Frederick died in a restraint at age 16. Max Benson died in a restraint at age 13. Angellika Arndt died in a restraint at age seven, Michael Renner Lewis III died in a restraint at age 15. Corey Foster died in a 57 restraint at age 16. Jonathan King died from hanging himself while in seclusion at age 13. (Alliance Against Restraint & Seclusion, 2022). In each of these cases, guardians turned their scholar to the agency for help. Responding staff came to work that day intending to provide that help. These individuals did not intend these outcomes, and some are now serving time in prison because of their response. In their pilot study, Bonner et al. (2002) acknowledged that clients were confused by staff responses to their reactions. There was fear of accessing mental health services because of the fear of this response and possible deadly outcomes. Another fear around loss of life comes regarding responding staff injury. Staff also enter the workplace concerned for their own safety. When working in a public school or unlocked facility, staff will report concerns for weapons brought into the school (Musu et al., 2019). With greater exposure to conversations around gun violence, staff enter the workplace afraid for their safety. Many have associated gun safety with mental health. There are mixed results in how much gun safety and mental health correlate. According to NAMI California (2019), "Every time we experience a tragedy involving guns, people with mental illness are drawn into the conversation. The truth is that the vast majority of aggression is not perpetrated by people with mental illness" (para. 1). Statistically speaking, gun violence is most attributed to experiencing bullying behaviors. The second reason for gun violence is medication misuse or side-effects with the third being prescribed medications that are not being taken (Lee, 2013). While most 58 shootings are attributed to interactions between peers, there are still components of mental health that cannot be ignored. Everytown for Gun Safety (2022) advocates the importance of positive and affirming school communities and stated that "Safe schools are built on trusting relationships among students, staff, and administrators" (para. 14). Additionally, mental health counselors and response teams who can assess potential danger and intervene early are also critical to this work. Every Town for Gun Safety (2022) stated that "School-employed mental health professionals serve as a critical resource for students as they navigate the education system and the challenges of emotional and social development." Crisis response is about early identification and meeting needs (Fox, et al., 2022). Included in safety measures are steps that should be considered carefully that may not support the reduction of aggressive acts, including gun violence. Within this list is the use of School Resource Officers (SROs) and arming teachers. Here, it is noted that SROs have not made a difference in the deterrence of gun violence or the prevention of deaths (Livingston et al., 2019). The use of the SRO is more likely to be overapplied to black and brown scholars at a rate of three times more often (Weisburst, 2019). How staff responds in dayto-day interactions and who is called into an escalated situation matters. ProACT (Fox, et al., 2022) teaches that, sometimes team members call for all available help; this may also be known as a show of force, show of support, or a code. While the team may feel safer in a group, without a clear understanding of their roles, the 59 energy and addition of many responding staff members rushing into a room will usually escalate the potential for assault. (Chapter 7, p. 14) This is why careful consideration to how responding staff enter a space is needed. In addition to these safety concerns, staff can have the concern of litigation for use, or lack of use, of restraint and seclusion. In training, ProACT (Fox, et al., 2022) crisis intervention asks trainees to consider the risk of restraint and the risk of not choosing restraint. The answer to the question is the same on both sides, and litigation for restraint is one of those concerns. According to the Office of Civil Rights (2022), there are one hundred four open cases in the United States under the category of "Disability - Restraint and Seclusion". The Alliance Against Seclusion and Restraint will also advocate fiercely that there is no place for these practices in treatment institutions and school settings. In the case of Cornelius Fredrick, 20 employees are being sued with three convicted with criminal charges. (Romine & Sturla, 2021). Three staff were indicted, and the school closed its doors two months after the death of Max Benson. The case is open today (Serrin, 2022). Angellika Arndt's death was ruled a homicide with the center she was treated at closing its doors and staff serving time (Zehnder, 2006). Again, the staff did not intend these outcomes. The restraints resulted in a worst-case scenario, was traumatic for all to experience, and ultimately compounded with years worth of litigation and unintended consequences to life. More and more scholars and families are saying that the use of restraint as a treatment for emotional and behavioral challenges is not effective. 60 Many school systems are looking to School Resource Officers (SROs) as the answer to the reduction in escalation and increased sense of safety. While it may feel more comfortable for some to involve a SRO in an escalation with a scholar, this response will often serve to escalate the situation further. This is a version of outsourcing the work and yields results that do not support the growth of the scholar. The SRO is not trained to treat mental health or behavioral challenges. The SRO is minimally trained to counsel someone who is dysregulated. According to the National Association of School Resource Officers (2022) "it is a best practice to use a triad concept to define the three main roles of school resource officers: educator (i.e., guest lecturer), informal counselor/mentor, and law enforcement officer" (NASRO, 2022). If there is an actual threat, then the plan to keep people safe has already failed. The SRO's role is to pacify a threat to the safety of the school. Staff have taken the liberty of applying this need for protection to include scholars who are having an escalated moment. As ProACT teaches, just because a person feels intimidated, does not mean that the acting out individual is operating in a way with intent to intimidate (Fox, et al., 2022, Chapter 17. p. 12). These clear distinctions teach responding staff to be mindful of feelings about an incident. Responding staff can add energy when they label an acting out individual as intimidating. More often the case is that the frustrated individual contributed to a feeling of intimidation. To combat these natural feelings, the focus turns to understanding how the brain functions and the internal forces at play that keep responding staff from engaging in a safe and productive way with an escalated individual. 61 NEUROSCIENCE A foundation in neuroscience will support staff reactions and responses. Scientists have improved ways to measure what is happening within the internal world in the past fifty years. The brain is composed of three main areas: brainstem, limbic area, and cortex. Dr. Siegel (2010) offers a hand model of the brain to support understanding. Hold up the hand and pull the thumb in, the thumb represents the limbic region including the hippocampus and amygdala. The four fingers represent the cerebral cortex with the ring, the middle fingers represent the middle part of the prefrontal cortex, and the wrist represents the brainstem and spinal cord. Bending the four fingers over the thumb represents the ways the brain is connected and how these areas all communicate with one another (p. 15). When a brain is experiencing dysregulation, certain areas become less aware of one another. This chaos or rigidity in the brain then leads to defensive reactions (Siegel, 2010, p. 132). Siegel (2010) goes on to explain that "we build a fence around our awareness so that we don't feel the anxiety or fear associated with feeling our feelings" (p. 132). These "automatic" strategies and responses happen without awareness or desire for them to be the response (Siegel, 2010, p. 132). By increasing awareness, responding staff can better interrupt its patterns and find an option for how to respond. Mindfulness strategies are the tools for interrupting these automatic responses. In addition to this understanding, Siegel (2010) explained that past experiences continue to influence present experiences (p. 150). These areas that impact implicit memory include: "perception, emotion, bodily sensation, behavior, 62 mental models, and priming" (Siegel, 2010, p. 150). Implicit memory, as Siegel (2010) explained, impacts attitude, beliefs, and brings forth prejudices and biases. This causes responding staff to "feel conviction that our beliefs and reactions are based on our present good judgment" (Siegel, 2010, p. 152) as opposed to the influence of past experiences. It requires great intentionality to be aware of and integrate implicit memories into working memory so that individuals are no longer impacted by them. The more traumatic an experience the more impacted when the implicit memory begins to surface. There is no choice as to when implicit memories will shade perception of what is happening in direct experience. They are just there, in the background of thinking. Jensen (1998), Siegel (2010), Wilson and Conyers (2016), and Desautels and McKnight (2019) all focus on the importance of teaching with the brain in mind. With increased understanding of what the mind is and how the brain functions individuals are pushed to consider practices both personally and professionally and how they show up daily. Thinking about thinking is the definition of metacognition. The term metacognition was first coined by John Flavell (1979). Wilson and Conyers (2016) outline how to teach metacognition to young learners using the analogy "drive their brains" as an access point. Lessons include identifying how the brain works and an understanding of the brain's plasticity. Neuroplasticity was coined by Jerzy Konorski in 1948 (Livingston, 1966), though people today continue to have beliefs based on the adage, "You can't teach an old dog new tricks." There is also a tendency to use traits and genes as a way of saying that staff and scholars have no control over behavior. 63 The mentality of "I am what I am" takes hold as if there is no choice or skill needed in how that trait or genes presents. Environmental conditions impact how and if a gene or trait will present (Jensen, 1998; Siegel, 2010). Understanding how the brain interprets the world has changed dramatically, especially in the last fifty years. Neuroanatomists introduced just how malleable the brain is and continues to be throughout life (Diamond et al., 1967). Where previously, individuals believed the brain to be fixed at a certain age, scientists now understand that there are opportunities for growth and change throughout a lifetime. Dr. Ramachandran (2009) explained that the sophistication of the mirror neuron is what has pushed society to survive and thrive. He hypothesized that it is through the activity of mirror neurons that information is passed along quickly. This happens because of imitating others. It is through this that culture is born. It could also be hypothesized that change in the system and response will occur much faster than expected thanks to mirror neurons. This is why modeling is so important! Teacher Reactions and Self-Regulation The outcome of the work around restructuring a specialized, alternative day school program was most visible in teacher reactions and responses toward scholars behavior. The following are areas explored in that learning. In public school institutions, staff responses cause a disconnect between colleagues, staff and scholars, and families and staff (Costello, et al., 2010). The interplay of mirror neurons and emotional contagions can lead to unintended consequences (Pert, 1997; Siegel, 2010). The ripple effect through emotional contagions, if not 64 put in check, can disrupt the learning environment far beyond a single moment. Siegel (2010) explained the above phenomenon as leading to chaos and rigidity within us. Siegle (2010) explained that "when differentiation is blocked, integration cannot occur. Without the movement toward integration, the entire system moves away from complexity-away from harmony-and into rigidity" (p. 66). Emotional health then is explained as the "function of integration" (Siegel, 2010, p. 69). The role of educators then becomes to learn what the brain is, how it functions, and how to find integration in times of rigidity and chaos. (Siegle, 2010). Once responding staff can find that within, they are better equipped to teach it to scholars. A mental health diagnosis that falls into rigidity is explained as depression, cutoffs, and avoidance. Chaos is explained as agitation, anxiety, and rage (Siegel 2010, p. 138). Desautels (2020) said, "Behavior management is not about students. Behavior management is about the adults" (p. 33). She further stated that "A dysregulated adult cannot regulate a child" (Desautels, 2020, p. 33). Responding staff can support regulating another when they are, "Being aware and adjusting [our] tone of voice, posture, and facial expressions" (Desautels, 2020, p. 34). Desautels (2020) clarified that "[a]ll of this creates an emotional contagion" (Desautels, 2020, p. 38) which can support the continuation of the escalation, or a sense of calm. Regarding responding staff, Desautels (2020) explained that [A]ll adults working with traumatized children and youth must become intentional and reflective in their actions if they are to succeed in helping to heal our most wounded students" (p. 40). Asking a child to clip up or down as a blame and 65 shame system does nothing to provide feedback, connect, teach, or regulate. How responding staff approach unexpected, chaotic behavior will determine if the situation improves or not. How the responding staff approaches a situation will also foster or hinder brain development. Brains go through a process of pruning some connections and growing new ones. This happens the most when younger and then lessens. Within scholars: [t]he normal remodeling of the brain is intensified by stress, and it can unmask or create problems during this vulnerable period. This makes the nine middle prefrontal functions-from fear modulation to empathy and moral awareness- somewhat unpredictable, so that emotional self-regulation can be challenging for any teenager. (Siegel, 2010, p. 87) Collectively, there is some understanding of what the brain goes through when growing up. What responding staff often fail to recognize is that the malleability of the brain indicates that this same pruning and reconnecting process exists as responding staff learn to interpret and respond to stimuli differently. Responding staff also experiences periods of struggling with emotional self-regulation. The experience of disequilibrium as responding staff attempt to respond differently to a scholar's escalation feels uncomfortable. Being patient and showing a level of support and grace will make a huge difference when it comes to building a new habit. de Ruitera, et al. (2020) found that teachers who have previous negative interactions with scholars were more likely to move toward anger for that scholar 66 in future disruptions. When a scholar is more than disruptive and moves toward an aggressive action, one could make a similar assumption that teachers carry this interaction with them into the future. ProACT (Fox, et al., 2022) teaches responding staff to be mindful of this, as predicting behaviors is dangerous. Yet, it goes against human nature and conventional wisdom to believe that a person who is prone to aggressive behaviors as a means for getting needs met will refrain from this on a whim, or from traditional disciplinary actions (detention, loss of recess or lunch with peer group, suspension from school or class). Instead, staff respond with assertive communication, empathy, and choice in how to navigate strong emotions. (Fox, et al., 2022, Chapter 5 pp. 6-7). ProACT (Fox, et al., 2022) provides a framework for assessing and responding to assaultive behaviors. This assessment requires the use of a selfcontrol plan implemented repeatedly for the responder to remain in the space of present, current, and direct experience. ProACT (Fox, et al., 2022) acknowledges that responding staff need to be wary of internal dialogue and reduced cognition that occurs when responding staff move away from baseline and into fight, flight, and freeze mode. By remaining calm and in the moment responding staff can calm a situation down and use only the amount of force needed to help the individual who is escalated to calm. When implementing this with school-aged children, it becomes imperative that responding staff approach escalation in this way. Remaining calm and lending calm (Fox, et al., 2022, Chapter 7. pp. 8-9) supports brain development through the responding staffs emotional contagions 67 (Pert, 1997; Christakis & Fowler, 2009; Siegel, 2010) and helps to create neural pathways that mirror healthy engagement (Ramachandran, 2009). A scholar with maladaptive behaviors can be disruptive to the learning environment. This same scholar, if triggered enough, can demonstrate aggression toward self, peer, or teacher in the learning space. This traumatic encounter paired with a lack of training leaves teachers to feel unequipped to work with the scholar. It leaves peers feeling unsafe in a space where they are asked to grow. The escalated scholar, once calmed, feels ashamed or angered by their response and toward the stimuli that escalated them to begin. This can lead to feelings that their community failed to protect them. The scholar expresses feelings unheard and unaccepted. This same scholar may have been met in response by staff who restrained and possibly secluded them. The added trauma of this response toward the scholar leaves them feeling unsafe at school. Now, the cycle of this type of escalation and response has hit a pattern and staff feel unable to break the cycle (Fox, et al., 2022, Chapter 4 p.4). Moving Forward Hiam Ginott (1972) wrote, I've come to the frightening conclusion that I am the decisive element in the classroom. It's my personal approach that creates the climate. It's my daily mood that makes the weather. As a teacher, I possess a tremendous power to make a child's life miserable or joyous. I can be a tool of torture or an instrument of inspiration. I can humiliate or heal. In all situations, it is my response that decides whether a crisis will be escalated or de- 68 escalated and a child humanized or dehumanize. Responding staff have a responsibility to create a daily mood that supports the joyous and healthy engagement of all scholars. This is especially true for those who struggle with social and emotional challenges. Some themes from this literature review that will help to guide how this study proceeds are as follows: 1. Structure through Multi-Tiered Systems of Support and collaboration with mental health agencies need careful consideration. 2. Dignity and respect for how staff respond are the hallmark of professionalism. 3. Careful consideration of stress levels and trauma experienced will support responding staff. 4. Connection and working alongside the escalating scholar supports long-term goals and overall safety. 5. Understanding neurology supports staff understanding of themselves and the scholars they serve. 6. Restoring and how responding staff engage with that practice will determine effectiveness. 69 METHODOLOGY Overview of the Design This retrospective causal-comparative research study worked to uncover the factors that impacted responding staff in approaching an escalated scholar in a specialized, alternative education day school program over seven years. The years before the shift in philosophy and practice in the program and local area specialized, alternative education day school programs served as the controls for this study. Each subsequent year served as the subsequent comparison group. The researcher presented data demonstrating restraint and seclusion use in local area-specialized day programs. While the demographics and intake process for the comparison programs varied, the data served as a comparison of what occurred in programs comparable to this one during the period of change. A retrospective list of professional learning that responding staff engaged with each year was gathered as information to help uncover the learning priorities. The staff covered topics through the years: Lesson design, Overall structure, Crisis management system, MTSS, Regulation, Staff restoration, Restorative practices, Trauma-informed care, Co-teaching structures, Brain function, Culturally responsive practices, and Strategies for increasing academic discourse. Upon completion of the data analysis, responding staff were surveyed to capture their perspectives of the period as they experienced it and their thoughts on the programs evolution. 70 All staff attended each training session, as make-up training sessions are scheduled at later dates and times. Identifying which staff participated in which training session varies due to staff attrition. New staff are trained by veteran staff in the on-boarding process to support the continuation of the knowledge and skill gained from previous years experience. Additionally, training materials are readily available to support staff in increasing their depth of understanding when they identify the need. This specialized, alternative day school program is located in one of Indiana's large, urban school districts. In 2016, it underwent a change in leadership and practices for how staff respond toward those with emotional and behavioral challenges. After seven years of this work, the school went from 327 incidents of restraint and seclusion in 2015-2016 to 43 incidents in 2022-2023. The current principal served as the assistant principal during that change in 2016. Surrounding schools and similar programs are adding additional seclusion spaces to their public school institutions and have other variables that cannot be controlled for in an experimental design. Instruments Used and Rational Credit Attainment: Gained through frequency count of credits offered to scholars at the start of the semester divided by the number awarded after the semester ended. The number offered and awarded is pulled directly from the schools data management system. Attendance Rate: A percentage generated from the schools data management system. The percentage is reported after every year ends. 71 Responding Staff Injuries: Data is pulled from the district Occupational Safety and Health Administration (OSHA) reports. The frequency count of incident reports that included the use of restraint and, or seclusion submitted for scholars who attended the year preceding the programs restructuring is presented and dates back to the 2011-2012 school year. The frequency of reports submitted in each subsequent year after the programs restructuring is presented through the 2022-2023 school year. Frequency count of incidents from the highest level of incident noted in the incident report is presented beginning in the 2016-2017 school year and up to the 2022-2023 school year. The frequency broken down by the same definition prior to the 2016-2017 school year is not available due to the school no longer having access to these incident reports because of the appropriate record destruction of these reports. Any incident noted that documents Levels 1 and 2 behaviors were screened out for the purposes of this study. This research focused on physical aggression, verbal aggression, elopement, and sexualized behaviors at Levels 3 and 4 as described in Appendix A. Levels of dangerousness for physical aggression are defined from ProACT (Fox, et al., 2022) definitions noted in Appendix A. The frequency of restraint and seclusion from the local area, similar programs as reported to the media served as one control. The incidents 72 reported reflect the time period of the restructuring for this specialized day school program. The ATLAS tool in Appendix G allowed staff to reflect on the data collected for this research and provide initial feedback and reflection. Surveys from responding staff were collected based on when staff were hired before and throughout the restructuring process. Population and Sample The population studied was the staff of a specialized, alternative day school program within an urban, public school setting, who remained through the challenges faced in the restructuring that began in 2016 and those hired since. The staff was represented by years of experience with the abovementioned work. See Table 12. Scholars who attended this program are all served with an Individualized Education Program (IEP) and have behaviors of concern that cannot be supported in the general education school setting, as determined by their case conference committee. The case conference committee included the parent, a representative from the specialized day program and the referring school, service providers such as a mental health therapist, occupational therapist, speech and language therapist, and a public school agency representative. The case conference committee may have included additional staff, such as an Assistant Director of Special Education, a School Psychologist, an advocate, or anyone else the parent may have invited to the meeting. See Intervention Evaluation section for data representing who was served in the specialized day program. 73 Data Collection All incident reports for scholars who attended the program were housed in a Google Drive built and maintained by the school staff. Each scholar had a folder where incident reports were housed and included additional information needed for the appropriate support and care of that scholar. Each scholars folder, from the 2016-2017 school year through the study, are maintained. If a scholar withdrew, transferred, or transitioned back to their homeschool, these files were archived in a Google folder labeled "No Longer Attending." This allowed the school to maintain records if the documents were called into question or needed for review. It also allowed staff to pull a scholar back into the active drive if they returned to the program. A blank copy of an incident report used by school staff was included in Appendix B. Directions for the review and coding of incident reports were included in Appendix C. Directions are always available to staff. The school's administrators monitored and tracked the use of restraint and seclusion daily. This looked like a frequency count of incidents that occurred. This was recorded on a spreadsheet so that all incidents were accounted for. The building-level administrators have maintained in-flight recordings because these incidents were not recorded in the data management system. The district also collected these incidents through a Google form submitted for engaging in the use of restraint or seclusion. One inter-rater reliability test that the school administrator engaged in was to review the list of incidents recorded in the Google form against the list recorded in the spreadsheet and clean up any 74 missed recordings at the close of each school year. The district's behavior specialist also worked to cross-reference the reporting systems on behalf of the program as another inter-rater reliability test to the self-reporting system. The use of reports from the specialized, alternative day school data management system included attendance rate each year, credits earned, and demographic information to demonstrate who was served in the program at intervals through the study. Scholar demographic information was shared from the 2011-2012, 2016-2017, and 2022-2023 school years to demonstrate change over time. The school district changed data management systems in the 20152016 school year. Therefore, a full picture was not always available as not all information was recorded in the data management system prior to the 2016-2017 school year. Credits earned have been recorded each semester since the 2015-2016 school year. This process looked like counting what was offered to each scholar at the start of the semester. Consideration was given to plans that included a reduced day. A scholar on a reduced day may not have the same access to credit-bearing classes as one who attended a full school day, depending on what was determined by that student's case conference committee. Once individualized plans are considered, the number of credits offered was counted. Then, the data secretary ran a report from the specialized alternative day school programs data management system to indicate the number of credits awarded. This was counted for each scholar. The total number of credits earned was divided by the total number of credits offered. This percentage was then recorded 75 as credits earned during the semester and overall for the year. By recording a percentage, the staff could better compare year-to-year, knowing that the number of scholars who engaged fluctuated. Responding staff injuries were reported through incident reports and the program's school nurse. If a responding staff member was injured, regardless of an accident or from supporting an intense behavior from a scholar, they reported this to the nurse as standard protocol. The nurse then provided treatment in the school setting and may have referred the individual to outside medical care. Depending on the severity, the outside medical care ranged from requiring emergency transport to the individual making an appointment to be assessed. This was recorded through an initial report with the nurse or elevated to a referral and reported to the Occupational Safety and Health Administration (OSHA). In some cases, the staff had an injury that became worse over time and needed outside care. If a scholar were injured, the parent was notified with the recommendation for outside medical care unless it were a true medical emergency in which medics were notified before the parent. These reports were then saved with the nurses' records, and a copy was provided to the buildinglevel administrator. At the close of the calendar year, the building-level administrator was provided a copy of the report summary from OSHA. The OSHA report showed the number of injuries reported. The report does not break down whether the injuries occurred from an accident or from supporting an intense behavior. 76 Once data was gathered, it was presented to the staff for reflection through the ATLAS tool and later through an individual survey. Survey data was only recorded using how long the interviewee had worked with the specialized alternative day program as indicated by a range of hire dates described above. All other identifiers were not recorded. Survey data was recorded in a Google Form. No demographic information was recorded from incident reports or nurse reports. Only the intensity of these reports was recorded and analyzed. The specialized day program has employed the participants in this research study for one year to up to twenty years. A frequency distribution chart indicating the number of staff and students per year is provided in the data section as a reference point for the reader. See Table 8. Limitations and Delimitations: Limitations: The principal and evaluator of the specialized alternative day school program conducted the research. At the time of the study, the researcher had been serving in a building-level leadership role since 2016. Efforts to control for the principal and evaluator of the program as the researcher include having a third party proctor and administer the survey with the researcher away to ensure that staff are able to reflect and respond, or not participate, without their evaluator being present. Procedures for this are described in the Ethics and Human Subjects and Data Management Issues section below. The researcher has been a certified ProACT (Fox, et al., 2022) trainer since 2016 and maintained that certification through the study. The researcher 77 taught this curriculum for crisis management several times a year since 2016. The researcher participated in two continued certification training courses in ProACT since 2016, the most recently completed in July of 2022. The committee to support this work's development includes a ProACT national trainer, Dr. Rick Van Acker. Underreporting has been a concern and has potential when working with restraint and seclusion as a response. While incident reports are reviewed to count levels 3 and 4 behaviors, it is unclear how many level 3 behaviors occurred without documentation. This is because level 3 behaviors are documented at the teachers discretion, as Appendix A defines. It is possible that the team was working on an intervention plan and decided not to document the behaviors formally while focusing on the fidelity of implementing the new intervention plan. The willingness of a school district and its governing body to explore topics around restraint and seclusion is a potential limitation for generalizability. This school district agreed at all levels to explore and innovate within the school experience in the interest of safety and better outcomes for the unique scholars served. A statistical analysis that compared staff injury could not be completed as neither the data for the control programs nor the years prior to the 2016-2017 school year could be located due to the destruction of records. A trendline was included to support the overall reduction in staff injury since the start of the intervention. Data from incidents of suspension from the year before the intervention, the start of the intervention, and the close of the recording period 78 were located and, therefore, included in the analysis as another reference point for the reader. Delimitation: The researcher looked at responding staff and data from scholars in a specialized, alternative day school program that supports scholars with emotional and behavioral challenges in an urban, public PK-12 school program in Indiana. Incidents and outcome data available from scholars who accessed the program from the 2011-2016 school year through the 2022-2023 school year were used within the findings and conclusions. Generalizing this work across settings would require time and an adaptive leadership approach (Heifetz, 1994). The population size was limited by the number of scholars and staff in the organization during the respective years of this study. The population size was transient as scholars would enroll, withdraw, and transition within and out of the program throughout the year. The staff body was transient as some decided not to continue work at the school for personal and professional reasons throughout the years of this study. See Table 8 for population size. Validity and Reliability Validity: Historical validity was impacted through the end of the 2019-2020 school year and the 2020-2021 school year when COVID-19 resulted in several rounds of virtual learning for all schools, including this specialized, alternative day school program. During the 2020-2021 school year, the governor of Indiana permitted specialized, alternative day school programs to hold limited in-person 79 instructional time under strict guidelines. This included half of the group attending on alternate days. While the same data was collected, it was skewed because of this. Historical validity was impacted in the 2021-2022 school year when the program was short five of 26 teachers due to attrition. The school was unable to hire candidates for the full school year. The program continued to struggle to hire all replacement staff, especially certified teachers, through the 2022-2023 school year. There was a threat to instrumentation regarding how incident reports were collected. The reports themselves had several iterations with the program evolution. Regarding intensity, what was collected remained the same as defined by ProACT (Fox, et al., 2022) physical aggression definitions from the 2016-2017 school year through the duration of the study. Recording interventions and outcomes evolved and became a stronger practice in Spring 2022. Therefore, what was available was reported for review. The validity of grading is a concern when considering the subjectivity from one teacher to another and their relationship with a scholar. This specialized day program utilized an online learning platform where a licensed content teacher provided feedback and grading to the scholar participating in the program. The teacher most knowledgeable about the scholars needs was the specialized day program teacher who provided access, small group instruction, and accommodations in the learning experience. The content teacher never met the scholar as they had other responsibilities in another building in the school system 80 during regular school hours. Grading and feedback occurred in the evenings while the specialized day program was closed and was submitted electronically. While this structure does not fully control teacher subjectivity, it is important to note that the teachers most knowledgeable about the scholar and invested in the success of the specialized day school program were not the ones awarding credits. Reliability: This research attempted to identify the variables that led to the reduction of the use of restraint and seclusion in a specialized, alternative day school program. Utilizing a combination of data gained from the data management system, the program-built system for reporting incidents, OSHA reports, and surveys of staff who experienced the changes is an attempt to ensure reliability. Data triangulation aimed to discover insights on how to make the changes for future responding staff. The potential for generalizability was communicated through overall demographic information, resource access, and other factors contributing to the overall shift in practice. The restructure was modeled after a private, specialized, alternative day school program, Centennial School, which worked in partnership with Lehigh University (Kern, et al., 2011). The researcher was the evaluator and building-level leader of the group participants in the restructuring and research. The researcher served as the Assistant Principal from the Fall of 2016 through the Spring of 2019 and as the Principal from the Fall of 2019 through the time this research was conducted. The researcher entered this work with the belief that it is possible and imperative 81 that responding staff learn to work with this most vulnerable and complex learning community through care that conveys dignity and respect. The researcher already believed doing this work without restraint and seclusion was possible. It was the potential for a restructuring of the program that drew the researcher into this work. This suggests that any attempt to duplicate this work may require leadership who already believe it to be possible to accomplish. Ethics and Human Subjects and Data Management Issues Staff were presented with written notice of the intent of the research study and that their participation or lack of participation would have no bearing on their evaluation because the researcher would not be present for the survey. Staff were informed that their responses were recorded electronically and that names were not recorded in the submitted responses. The only identifying information was the self-selected range for being hired into the program. Teachers were permitted to opt out of submitting a survey and pass on any question as no response was required, allowing staff to leave answers blank. Data was stored with the researchers personal accounts and outside the work environment. No identifiable information was collected outside of years of experience within the range described in Measurement in the Overview. Participants' names were not recorded during the collection process. The number of possible participants for each range and how many agreed to be interviewed is reported in Table 12. A third party introduced the protocol outlined in Appendix D and was available to answer any questions as staff completed the survey. The researcher left the 82 building at the time of collection so that staff who opted out could move about the building without fear of discovery. 83 INTERVENTION DEVELOPMENT This study intended to uncover the internal work of staff and external outcomes within a specialized, alternative day school program that underwent restructuring with the intent to reduce and eliminate the practice of restraint and seclusion with the most vulnerable and complex learners. It aimed to capture the professional learning opportunities that the staff underwent as a collective, the outcomes captured over these seven years, and a post-analysis self-recorded survey to gain the perspectives of staff whose lived experiences with these changes. The professional learning opportunities were gathered by reviewing past professional learning plans submitted to the programs Director each year. The structural changes came through iterations as the group processed the needs and adjusted the operation of the program. The collection of professional learning and structural changes was reviewed by the staff of the program and updated based on feedback gained. The group responsible for the evolution of the program included the student body, who provided crucial feedback, families who accessed the program, staff who worked in the program, and the building-level administrators who synthesized the feedback to support the design of the structures used. Feedback was gained in formal and informal settings as the building-level leaders worked to diagnose the system (Heifetz, 1994). The iterative process included diagnosing the system and adjusting the system to support the work. This practice continued beyond the research period. 84 The data collected, and the trends described are in the data collection section. Data was presented in two all-staff meetings. Staff were seated with the groups they worked closest to within their day-to-day activities. Staff received a graphic organizer to capture their thoughts from the data reviewed. (See Appendix F for ATLAS Protocol). The researcher facilitated moving from one section of the ATLAS Protocol to the next. The presentation of the data, along with the ATLAS Protocol, took one hour and spread over two all-staff meetings. Teams were asked to copy the ATLAS Protocol without labeling who they were on it. They typed their responses in each section of the ATLAS Protocol. They then printed the completed protocol and turned it in to the main office secretaries before passing it back to the researcher. No personal names, team names, or other identifiable information was provided in the protocol responses. Staff gathered for a follow-up meeting where a third party introduced the protocol for a self-guided survey located in Appendix D. The survey questions are in Appendix E. These were loaded into a Google Form that was recorded in the researcher's personal account and did not record names. Staff identified themselves based on years of experience described in Measurement in the Overview. The self-guided survey was scheduled for one meeting and emailed along with the protocol to those who wished to complete it independently. The window for completing the survey was one week. Considerations for protecting external validity include acknowledging that staff observed the present leadership change at the start of restructuring the program. District administrators made it clear that remaining in this work would 85 require staff to change their practice. Staff underwent multiple iterations during this time as they collectively worked to uncover what that change would look and feel like. This multiple-threat interference made it difficult for the researcher to capture everything that occurred to make these changes. This was the reason that a survey with staff was included. This attempt to gain additional perspectives was a protection against external validity by gaining insights to round out what could be gained through data review alone. The population selection was not random, as the reports filed from all who participated in each year are those who are studied. Participation in this specialized, alternative day school program required the participant to have already been identified with special education services and recommended by their case conference committee to attend this program due to the intensity of their behaviors. Therefore, when looking at a program-wide restructuring, there is no way to randomize who received the interventions and who did not. A clear depiction of the scholar this program serves, the resources (financial and human capital) available, the in-take process utilized, and the overall structure would help other institutions to know how this program compares and differs from their program. 86 INTERVENTION PROPOSAL Table 1 Steps and Timeline for Study Start Step Resources Upon successful proposal defense Gain IRB approval. Finalize letter to the School Board for approval. Submit with IRB the day after successful defense. Upon approval from IRB Submit for Superintendent approval and present the intent of the study to the Learning Community described above. Finalize community consent. Submit after IRB approval is gained. Upon approval from Superintendent Gain consent from the community for their participation in the interview survey for the purpose of the study. Schedule a staff meeting to review the data gathered. Submit a letter explaining what is being studied and what will be used for this study. Week one Collect pre-intervention data that will be pulled from before the 2015-2016 school year. This could include attendance records, high school credit attainment, staff injuries frequency and intensity, scholar injuries frequency, and intensity, and use of restraint and seclusion Collect data from area specialized day school programs as a comparison to this program. Reports run from the data management system, nurses office, and collect incident reports. Begin putting into tables and spreadsheets for ease of analysis. Week one Collect the titles of the books staff engaged with each year of program restructuring. Present district administration guidance and resources provided each year of the program's restructuring. Present professional learning opportunities that staff engaged with through outside agencies at each Review previous year professional development (PD) plans, reach out to the principal who served from 20162018, and Survey staff to see what 87 year of the program's restructuring. PD helped in the Present the changes in staff that occurred journey through in each year of the program's the restructure. restructuring. Present in-school structures such as the master schedule, goals for that year, how resources were structured and utilized within the school day, and what systems were in place to support the work at each year of the restructure. Week two Collect yearly restructure data that will be pulled from the 2016-2017 school year, the 2017-2018 school year, the 20182019 school year, the 2019-2020 school year, the 2020-2021 school year, the 2021-2022 school year, and the 20222023 school year. This will include attendance records, high school credit attainment, OSHA injury reports, and levels of dangerousness reported in incident reports based on frequency and intensity (all described above) for each school year. Review the incident reports from each respective year. Create a spreadsheet with the frequency count from each year. Week three Analyze the data before and after the study is completed using descriptive data. Present and make adjustments based on feedback gained. Week four Create tables, graphs, and charts of Spreadsheets, findings. Schedule a time to present to the pivot tables, and specialized, alternative day school statistical analysis program staff. tables. Week five Present the findings and what can be Presentation of the inferred from them. Lead the group in above data sets. reflection of data using the ATLAS protocol. Let staff know they will be invited to a self-guided survey to gain additional insights. Weeks six seven Provide the survey to staff as described above. Pull any additional insights into what contributed to the restructuring. 88 Weeks eightten Bring the contributing factors to the restructuring back to the current staff of the specialized, alternative day school program. Present this information gained and ask for anything missed. Analyze feedback provided for themes and principles. Weeks ten twelve Adjust from feedback provided by staff. Write up the analysis. 89 INTERVENTION EVALUATION Who is Served in this Program? The Figures below represent the races and genders of the student body during the 2011-2012, 2016-2017 (start of intervention), and 2022-2023 (end of study period) created from the total number of students who attended in each respective school year. Figure 1 Note: Bar Graph represents the percentage of each racial group in attendance at the specialized day program during the 2011-2012 school year as compared with the 2016-2017 school year which started the intervention and the 2022-2023 school year end of intervention study period. 90 Special Note: Figure was updated after presentation to staff for the purpose of historical view for the reader. Only 2022-2023 demographic data was presented to staff at the time of the ATLAS tool and survey. Figure 2 Note: Bar Graph represents the percentage by gender in attendance at the specialized day program during the 2011-2012 school year as compared with the 2016-2017 school year which started the intervention and the 2022-2023 school year end of intervention study period. Special Note: Figure was updated after presentation to staff for the purpose of historical view for the reader. Only 2022-2023 demographic data was presented to staff at the time of the ATLAS tool and survey. 91 Figure 3 Note: Percentages of grade levels served by school year include the 2011-2012 school year as compared with the 2016-2017 school year (start of intervention) with the 2022-2023 school year (end of intervention recording period). Special Note: Figure was updated after presentation to staff for the purpose of clarification for the reader. 92 Figure 4 Note: Percentages of primary disability categories served by school year include 2016-2017 school year (start of intervention) with the 2022-2023 school year (end of intervention recording period). Unable to retrieve the 2011-2012 school year primary disability categories served from the schools data management system as the district was previously in a different system. Special Note: Figure was updated after presentation to staff for the purpose of clarification for the reader. 93 Figure 5 Note: The chart represents the average daily attendance reported through the programs data management system after each school year. Attendance was not recorded in the data system prior to the 2016-2017 school year. Presentation of Data on Historical Perspective: Restraint and seclusion were common practices in crisis management and practice before this change, as reported below. Data collected from credit attainment as an indicator of what the teachers could do with the time they gained from no longer responding in this way was also presented. 94 Figure 6 Note: Pre-intervention data from the 2011-2012 school year through the 20152016 school year compared with post-intervention data from the 2016-2017 school year through 2022-2023 school years. Numbers indicate incidents where restraint or seclusion was used as they were reported to the school system, the consultant who supported the change in programming, and the Office of Civil Rights. Since the change in 2016, the numbers have been cross-checked between two tracking systems, one used by the school leaders and one from the report submitted to the Office of Special Services, which then reports to the Office of Civil Rights. Incident reports are submitted for each of these events. Individual reports are shared with the families. Special Note: Figure was updated after presentation to staff for the purpose of clarification for the reader. Line added to delineate pre and post intervention for ease. 95 One similar Indiana school program (control 1) had 1,049 incidents of seclusion and 423 incidents of restraint between the 2017-2018 school year and the 2021-2022 school year for a total student body of around 300 full-time students. Another Indiana similar school program (control 2) reported 919 incidents of seclusion and 413 incidents of restraint between the 2018-2019 school year and the 2021-2022 school year for a total student body of between 160 and 200 students (Gaines, 2023). Table 2 Incidents of Restraint and Seclusion as Compared with Similar Programs 2017-2018 through 2021-2022 Seclusion Restraint Control 1 1049 423 *Control 2 919 413 Program 140 141 Note: Table represents the use of restraint and seclusion for three alternative day public day school programs in Indiana. Total number of incidents for both restraint and seclusion that spanned the 2017-2018 through the 2021-2022 school year are represented. Special Note: Control 2 reported one less school year of restraint and seclusion as compared with Control 1 and the program being studies. 96 Special Note: Table was updated after presentation to staff for the purpose of historical view for the reader. Numbers were presented in paragraph format at the time of the ATLAS tool and survey. Table 3 Multinomial Test for use of Seclusion df P Multinomial 687.866 2 < .001 Descriptives for use of Seclusion Program Observed Expected: Multinomial 1 919 702.667 2 1049 702.667 3 140 702.667 95% Confidence Interval Lower Upper 874.091 964.285 1003.540 1094.474 118.376 164.205 Note. Confidence intervals are based on independent binomial distributions. Descriptives Plot for Use of Seclusion 97 Note: Nonparametric multinomial test of statistical significance run to demonstrate the intervention was statistically significant for reduction of the use of seclusion as a response from staff. Special Note: Statistical analysis was added after the presentation to the staff for the purpose of analysis. Table 4 Multinomial Test for use of Restraint df P Multinomial 157.224 2 < .001 Descriptives for use of Restraint Program Observed Expected: Multinomial 1 413 325.667 2 423 325.667 3 141 325.667 95% Confidence Interval Lower Upper 382.503 443.953 392.382 454.014 120.044 164.074 Note. Confidence intervals are based on independent binomial distributions. Descriptives Plot for use of Restraint 98 Note: Nonparametric multinomial test of statistical significance run to demonstrate the intervention was statistically significant for reduction of the use of restraint as a response from staff. Special Note: Statistical analysis was added after the presentation to the staff for the purpose of analysis. Figure 7 Note: Total incidents of out of school suspension and total number of out of school suspension days issued as compared by the 2015-2016 (preintervention), 2016-2017 (start of intervention, and 2022-2023 (end of recording period) school years. Special Note: Figure was updated after presentation to staff for the purpose of historical view for the reader. Only 2022-2023 demographic data was presented to staff at the time of the ATLAS tool and survey. 99 Table 5 Multinomial Test for Suspension Incidents df P Multinomial 12.957 2 0.002 Descriptives for Suspension Incidents YTD Observed Expected: Multinomial 1 195 163.667 2 166 163.667 3 130 163.667 Descriptives Plot for Suspension Incidents Note: Nonparametric multinomial test of statistical significance run to demonstrate the intervention was statistically significant for reduction of the use of incidents of suspension. Special Note: Statistical analysis was added after the presentation to the staff for the purpose of analysis. 100 Figure 8 Note: Total incidents of staff injuries reported to OSHA submitted as compared by year with trendline. Incidents reported do not distinguish between accident and injury from responding to intense behavior. Special Note: Unable to retrieve OSHA reports prior to 2016 based on appropriate destruction of records. 101 Table 7 Reconnect Debrief Staffing Practice Replace Restoration FGDM Update Plan Totals 6 29 6 10 19 40 11 14 714 37 7 22 12 34 21 19 152 571 35 19 10 32 28 10 12 717 86 52 24 18 38 27 41 286 19-20 353 19 4 17 12 25 12 16 458 69 36 17 30 37 21 23 233 20-21 317 12 4 11 24 19 448 95 38 13 15 51 10 18 240 21-22 374 18 11 22-23 423 15 16-17 256 2 3 8 8 9 3 17-18 597 17 18-19 7 54 E4 0 E3 4 V4 2 V3 5 S4 5 S3 7 P4 2 291 P3 TOTALS Incidents Totaled Including Reported Restorative Activities 7 15 77 25 40 567 116 29 19 12 62 18 37 293 9 32 14 88 16 22 619 129 16 25 22 70 27 32 321 Totals 2891 118 56 95 107 321 101 125 3814 539 183 125 111 296 124 176 1554 Note: Incidents of behavior through incident report review for each school year. Incidents of reported restorative activity when reported on incident reports. Behavior code definitions are located in Appendix A. Incidents prior to 2016-2017 are not available due to proper destruction of records. System for recording incidents with common definition prior to 2016-2017 did not exist. 102 Figure 9 Note: Percentage indicates the number of credits earned of what was offered for that semester. Awarding credits can arguably be subjective as grades come from the teacher who participates in the specialized day program. All curriculum provided to the specialized day program comes from the larger metropolitan school district, which was a part of where grading practices were monitored by the building-level administration and the high school counselor assigned to this program. Another challenge was having dual licensed teachers, both a content specialist and a special education instructor. To support the courses without content specialists, the program would partner with the online learning program in the same metropolitan school district to provide courses. These courses were then graded by a licensed teacher outside of the specialized day program and 103 through the online learning program. The role of the special education teacher then was focused on finding an access point to the learning for that student. Table 8 Staff and Student Body by School Year School Year Staff Open positions Total Students Full Transition Served Back 2016-2017 22 certified 36 classified 0 138 19 2017-2018 22 certified 38 classified 0 139 7 2018-2019 24 certified 27 classified 1 certified 2 classified 143 16 2019-2020 26 certified 25 classified 125 10 4 classified 2020-2021 27 certified 19 classified 0 123 13 2021-2022 27 certified 16 classified 3 certified 1 classified 135 12 2022-2023 27 certified 15 classified 1 certified 1 classified 121 9 Note: The table includes working staff and the number of attending students for that school year. The total includes support staff including the office, nurses, and building sub. Staff total includes open positions noted. Student total includes all students served at any point during that school year. The transition total only included students who made a full transition and does not include students who made a partial transition where they worked part of their day in one setting and the other part in the specialized day program. 104 Data Review Findings from Professional Learning, ATLAS, and Survey Table 9 Professional Learning by School Year 2016-2017 Trauma Informed Care Walkthroughs focused on positive feedback for scholars. CHAMPS expectations clear ProACT crisis intervention, first year 2017-2018 Book studies: Costello, B., Wachtel, J., Wachtel, T. (2010). Restorative circles in schools: Building community and enhancing learning a practical guide for educators. International Institute for Restorative Practices. Fisher, D., Frey, N. (2008). Better learning: Through structured teaching. ASCD. Souers, K., Hall, P. (2016). Fostering resilient brains: Strategies for creating a trauma-sensitive classroom. ASCD. Hoerr, T. R., (2017). The formative five: Fostering grit, empathy, and other success skills every student needs. ASCD. Wilson, D. Conyers, M. (2016). Teaching students to drive their brains: Metacognitive strategies, activities, and lesson ideas. ASCD. Stutzman-Amstutz, L., Mullett, J.H. (2005). The little book of 105 restorative discipline for schools: Teaching responsibility' creating caring climates. The Little Books of Justice & Peacebuilding. 2018-2019 Continued with same books as 17-18. 2019-2020 Partnered with IDOE INSERT TITLES to develop MTSS Book Study & follow up PD: Souers, K., Hall, P. (2016). Fostering resilient brains: Strategies for creating a trauma-sensitive classroom. ASCD. 2020-2021 Desautels, L., McKinight, M. (2019). Eyes are never quiet: Listening beneath the behaviors of our most troubled students. Wyatte-MacKenzie. Goldstein, A. (1988). Prepare curriculum. Research Press. Seclusion Study Counsel PLC Leveling Up and Progress Monitoring PLC Academic Progress and Showcasing PLC Improv led by Jim Ansaldo from the Indiana Institute on Disability and Community through Indiana University. Staff regulation and self-care 2021-2022 Goal writing and Data collection Braaten, S., (1998). Behavioral objective sequence. Sheldon Braaten. Mental Health & Wellness pulled from Dr. Desautel's work. Building classroom libraries and conducting mini-lessons. 106 Kagan, L., Kagan, S., Kagan, M. (2015). Kagan structures. Kagan Publishing. Jackson, R., (2019). Becoming the educator they need. ASCD. Instructional Technology work 2022-2023 Singh, A. A., (2019). The racial healing handbook: practical activities to help you challenge privilege, confront systematic racism & engage in collective healing. New Harbinger Implicit Bias PLC Parent Communication PLC Assessment PLC Once a nine weeks ProACT training mini-refresher Table 10 Structural Changes by School Year 2016-2017 Clubs Added Students in secondary begin moving class to class (departmentalized teaching structure) Level system was introduced to make clear how to progress through the program. Mission & Vision updated. Lesson design 2017-2018 Master Schedule restructure to include built-in regulation times 2018-2019 Change in leadership 107 2019-2020 Begin remodel process. Restructure for COVID-19 Dr. Azziz presentation Compassion Fatigue group study 2020-2021 Conclude remodel. Restructure for COVID-19 2021-2022 Built behavior tracking system Change in leadership 2022-2023 Build effective sensory spaces to support our work with Zones of Regulation. Effective use of visuals. Effective use of behavior tracking system. Table 11 ATLAS Participants Possible Total Total Participants Participants Percentage Day 1 31 27 87% Day 2 31 26 84% 108 ATLAS Findings The ATLAS tool provided the researcher with time to gain feedback on the data presented and insights into what would help to capture the story of the evolution of this program. Initially, the participants were able to provide insights that included missing data points that could help clarify the graph structure that, when addressed, supported the descriptive data presentation for the reader. Clearing up these points helped the researcher to present more meaningful data. Implications for the classroom that were identified by staff in the ATLAS tool included: Common language developed through a social-emotional learning curriculum adopted by the school district. Universal expectations System for the development of intervention plans Meeting needs over gaining compliance. Working with families to develop strategies. Data tracking with fidelity to support the development of effective plans. Develop an understanding of when to push and when not to Modeling healthy relationship skills Defined program including the level system to transition back to homeschool. Communication of progress with families. Understanding the root of the behavior Only using restraint as a last resort and for as long as necessary 109 Strategies that are effective before the use of restraint Staff complete a reconnect after an incident to better understand. Staff looking at behavior objectively and be aware of implicit bias (gender, race, ability) Well-developed skills for crisis communication. ProACT Training Debriefing incidents School-wide goal that is tracked and communicated regularly. Understanding of family and cultural norms Survey Participants Staff were provided with open-ended questions. Seventy-four percent of the possible participants gave the researcher feedback to support the project. Table seven shows where participants were hired to work at this specialized day program. The table does not indicate years of experience in teaching, only when the staff member joined this teaching assignment. 110 Table 12 Participants in Survey by Year Hired to Program Possible Total Total Participants Participants Responding Percentage Prior to Spring 2016 8 7 88% Fall 2016 Spring 4 2 50% 8 8 100% 11 6 55% 31 23 74% 2019 Fall 2019 Spring 2022 Fall 2022 Spring 2024 Total Survey Findings Themes were pulled from reading staff survey responses. A theme was identified when more than one staff member referenced that specific theme in their response. The table represents the number of times a theme was identified from the open-ended questions. A few statements that staff made in response are listed below each question. 111 Table 13 Themes from Survey Theme Identified Total Participants who Endorsed Theme Educating the whole child 2 Restorative Practices 13 Safer for staff and students 6 Dignity and respect focus 7 Trauma reduction 2 Journey was difficult 4 Paradigm shift 7 Accountability for staff and students 2 SEL Curriculum 4 Debriefing 2 Regulation Strategies 9 Equity and implicit bias work 3 Staff Regulation and Self Care 3 ProACT 20 Data Review and Goal Setting 7 Observation of peers 2 Brain Research 4 Collaboration with Team 22 Building Connections 10 Visuals 2 112 Clear Levels Progression 2 Engaging lessons 5 Clear expectations 2 Administration trust of staff 8 Change in Culture 5 Having a positive impact 9 Community voices are heard 3 Patience 1 Environmental Changes 4 Survey Question One: Upon reviewing the school data from before the changes, through the changes, and seeing similar data from similar programs, what impressions do you have about the journey to restructure the school? I think that treating students with dignity and respect, and refraining from using restraint and seclusion unless absolutely necessary likely gives students the opportunity to understand and improve behavior and make transitions back to home schools. I think that it was a very great journey to restructure that is helping scholars to feel more dignified and successful and to help staff feel more in control and safe. It was necessary. Restraint and seclusions can cause irreparable harm to both staff and students physically, emotionally, and mentally. Utilizing ProACT and doing a deep dive on SEL skills has allowed for fewer staff injuries and less 113 P4s. Our school is not perfect. We have students who are suspended more than other students for doing the same thing, but we're a work in progress. It was a process but the approach to finding the structure that it needed didn't end after tough times throughout the process. There was a persistence that was needed to find what would benefit the scholars in the best way possible. The data is reflective of the work we do and are passionate about in this building. The number of restraints and seclusion going down drastically is a sign that working with scholars who face complex needs can be done with dignity and respect for all involved. The journey has made the staff more accountable to the supports and accommodations that they need to follow for the scholar to be successful. Looking back, I feel that some of the time, scholars were restrained for compliance to staff needs rather than their needs. Over time the number of injury for myself has decreased significantly as well as the severity. Survey Question Two: What has been the most impactful professional learning you engaged with to create this change? Most impactful for me personally has been professional learning in Teaching Students to Drive Their Brains, The Zones of Regulation and Restorative Practices all of which have supported students in understanding themselves, their brains, their emotions, alternatives to aggression and conflict resolution for the purpose of achieving growth and restoring opportunity and relationships. 114 Survey Question Three: What have been the most impactful changes to structure? (Consideration structure, instructional practices, crisis response, connection, and restoration)? I think restoration has been a huge piece for scholars and staff. It helps our scholars to kind of "see it through" and see that staff is here to support them even after crisis. I think it can also help staff further understand confusing situations where they may not have understood why a scholar reacted the way they did. It also can help to repair relationships and trust for both scholars and staff. There is also a benefit to restoration that includes cleaning up or fixing broken items because it allows students to see that the things they take part in don't just "go away". Crisis response and connection. Staff seek to make connections and use verbal de-escalation or planned ignoring rather than adding to the problem. They utilize each other's strengths and ask for advice. The most impactful changes to the structure have been how teams are now given more support within the classroom rather than letting students navigate when they need support. It was first set out that students would use an SSC or SELF center [outside of class support] to help regulate their emotional needs. It has been more beneficial to have supports in the classroom to help scholars navigate their learning with support within the classroom so they're able to practice the skills they need in order to transition. 115 The most impactful has been being able to trust staff to constantly make change as the plans/needs change while honoring the staffs' unique approach to use research and best practice in their respective spaces. The use of co-teaching and always having someone to bounce ideas off of has been very helpful. It is also helpful to have another set of eyes and have that person be able to check your ego as needed. Survey Question Four: What has been the most impactful change you have experienced in what you observed of your colleagues? (Consideration structure, instructional practices, crisis response, connection, and restoration)? Focusing on having everyone use the same SEL curriculum to ensure that all students are receiving the baseline skills and building upon them so when they get to the next grade level, the teachers can continue building upon these skills. The two biggest impacts have been the response to a crisis and restoration. The staff does a great job in understanding where each scholar is at and what they're needing in that moment to stay safe. There's a lot of work that goes into navigating students to participate in restoration which can be in the moment or at another time when conversations happen. Watching the staff that has stayed and the staff that has left, has shown that this work has to be a passion at this point since what this building does is outside the everyday norms in most educational settings has guided my own decisions to continuously show up and be the strongest advocate I can be. The 116 changes of allowing families to also see that staff are not jumping to the every day status quo of suspensions, blame, and what one parent said, "shame" is clear in the way we see progress and growth with scholars and their families. Survey Question Five: What makes this school a place where you want to continue to work? I want to continue working here because I feel like you can actually see the impact you have on students. After working through other organizations in different roles and hearing the things they say and the way they feel about behavior and watching the way they react to crisis, working here makes me feel good and like we're on the right track to help students in a dignified way. I also feel very supported and seen and validated by my co workers. I want to continue to work here because of the growth we get to see in our scholars. Our scholars at times are passed off with very little to no interventions along the way which can be a disservice to them and their education. The feeling that you get when you see your students succeed or make improvements to better themselves, not only academically but socially and mentally. I LOVE the philosophy!!! This school and the core of acknowledging and walking the talk is unlike any other place I have taught. The trauma I have faced from previous experiences as a special education student and as a special education teacher made me want to leave because my pedagogical beliefs and understanding of mental 117 health were not honored in other buildings. My lack of drive to punish in restraint, push back on why my scholars were being removed from classes for things that could be best addressed in the classroom, and work on meeting my scholars where they needed to be met in terms of basic needs was seen as hindrance. Knowing that the work we do every day in this building is not replicated in other places keeps me working here because while exhausting, I can sleep knowing that at least while in this building our scholars and I are receiving and growing in ways mental health and brain research show are appropriate and needed to thrive. The people here want to help these scholars and truly believe in helping them. I want to continue working here so that one day I will be comfortable enough in my skill set to step in to other [district] buildings and help the staff there grow and change to help support scholars in a way that meets scholar needs not staff needs. It is rewarding, If I can help one child or more to be successful. Empathy and love for the students and staff we serve at [this school]. We all need someone who will accept us for who we are, value us despite our experiences or circumstances and love us enough to speak truth and spur us to grow for our own benefit. Survey Question Six: What advice do you have for a school considering similar work so that they are equipped to begin? Do your research, go to many other schools and programs and see it in action MANY times. Meet the staff and students and speak with them to ask 118 questions. Get someone who has experience to help you set up and open your new program/school and be a guide/mentor. Be realistic that significant and real change takes time in all things. Dig deep in yourself before you begin this work to see how you truly feel about being able to "give up control" and your communication and compassion skills. Prepare for it to get worse before it gets better. Build a tight-knit community and lean on them. Make sure you are ready to carry someone else's load on some days because some days you will need someone to carry your load. If conventional ways don't work, try unconventional ways. Make sure your partnering schools should practice ProAct. There have been several instances where students transitioned out of this environment back to their home school only to be placed in padded, secluded rooms and restrained to the floor. That is trauma that is not needed. From my time at [this school], I have seen the physical environment and staff responses have a significant impact on student behavior, so those places would be a good starting point. Consider how the environment is affecting student's sensory needs. These can often be small changes (volume level, light, seating, etc). Consider how staff responses to students as they are the beginning stages of escalation (be a down arrow). It is okay to question yourself. It is okay to question the process. It is okay to have to come back to the drawing board again and again. Be flexible and allow yourself the dignity and grace you extend to your scholars. 119 Be willing to have open conversations. Make sure the staff understands the why, and understands how they will be safe and protected. They have to understand that admin has their backs. Hang on! It is a very personal journey. Work together, every voice matters, every person counts and it takes a village. Where one falls short, another builds up. Educating and loving students hand in hand is the beginning of creating a culture where everyone feels valued safe and capable of growth and success no matter their beginning! Group Concerns Seven responses included themes that reflected the group's concerns throughout the journey. These included not feeling as though all staff voices were heard through the restructuring process, distrust of the new administrative team, data not telling the full story, fear of repercussions for doing the wrong thing, lowered morale, and feeling unsafe. Adaptive leadership requires the acknowledgment of the challenges faced when moving a program away from a system of control and toward one of acceptance and meeting needs. These themes represent the challenges that the community will face throughout the journey toward change and beyond. Holding the fears of the group is a crucial step in an adaptive challenge (Heifetz,1994). Combining the ATLAS and the Survey Results Within the survey results, twelve themes were endorsed five or more times. These are included on Table 9. 120 Table 14 Top Themes Identified from Survey Results Theme Number of times a participant identified a theme Collaboration with Team 22 ProACT 20 Restorative Practices 13 Building Connections 10 Having a positive impact 9 Regulation Strategies 9 Admin trust staff 8 Data Review and goal 7 Paradigm Shift 7 Dignity and Respect 7 Focus Safer for students and 6 staff Change in culture 5 The analysis and implications will explore the items endorsed by eight or more staff members. These included Collaboration with Team, ProACT, Restorative Practices, Building Connections, Having a positive impact, Regulation Strategies, and Administration trust in staff. 121 Discussion of the Analysis of the Findings This study aimed to answer the question, how did a specialized, alternative day school program in a public school system in Indiana utilize professional learning and feedback gained from diagnostic work to restructure, thereby reducing the use of restraint, seclusion, and injuries that required outside medical attention while increasing academic attainment over the course of seven years? The null hypothesis stated that over the course of seven years, professional learning, feedback gained from diagnostic work, and a restructure to reduce the use of restraint, seclusion, and injuries that required outside medical attention or academic attainment have had no impact. The null hypothesis was rejected based on the nonparametric multinomial analysis conducted on restraint and seclusion compared to two specialized day school programs in Indiana. For the specialized day program, several known factors came to light that launched the change. Two events were significant enough to draw the attention of lawyers and upper administration. These events led to discussions that change was needed, though no one knew what that meant. One event occurred in 2011, the other in 2015. During the 2014 school year, a consultant, Dr. Van Acker, was called to observe and work with the programs team and their leadership. The consultant made many recommendations, and ultimately, it was up to the program evaluators and the school system's leadership to decide what steps to take. 122 The start of the 2016-2017 school year brought the beginning of those changes. New leadership for the building and a new crisis intervention program were introduced. Admittedly, the leadership and the new programming were not ready at the start of the school year, which created a difficult transition period for all. It was not until the start of the second semester that the team had any sense of direction. Families expressed frustration about why the team would not restrain their children and stop them from misbehaving. There was pushback from staff who expressed a significant feeling of loss through not being heard, disbelief that the changes would keep them safe, and feeling that things were being done to them rather than with them. The journey toward change began to take shape in the 2017-2018 school year, as evidenced by the learning the team engaged with beginning that year. In the collective conversation and individual reflections, staff endorsed seven themes eight or more times: Collaboration with the Team, ProACT Crisis Intervention, Restorative Practices, Building Connections, Having a Positive Impact, Regulation Strategies, and Administrative Trust in Staff. Each will be discussed further next, including any training provided to support growth in each theme identified. Collaboration with the team had several iterations for structure and format. It became evident early on that there was a need to dedicate time to problemsolving, creating the plan, developing visuals for both staff and scholars to support the implementation of plans, and providing practice with the plan for both staff and scholars. With so many needs in one schoolhouse, finding time for 123 teams to collaborate was very difficult. Creating a system to support this is ongoing work as the team balances the school's needs with those of the individuals it aims to serve and the staff who choose to work there. Within the comments around collaboration, staff identified subthemes that included the co-teaching structure, increased accountability for staff to follow the plans, debriefing incidents that included scholar perspective, utilizing team members' strengths, and opportunities to share ideas and try new things with support. The specialized day school program has dedicated time as part of the Multi-Tiered Systems of Support structure so the team could process challenges. Dedicated time for collaboration includes time for case conferences, debriefing incidents, staffing to seek solutions, staffing to train the team on a new plan, and a monthly check-in to review the targeted focus for each scholar. ProACT (Fox, et al., 2022) training focuses on how to look at the environment, manage internal stressors, and gain the scholar's perspective to reduce and manage stress so that restraint and seclusion are not needed. The training provides a framework for the team to work from rather than an approach or a technique. The real work comes after the training when the team uses the framework to help process incidents and reduce challenges that lead to escalation. Within ProACT, subthemes identified by the staff included crisis communication, the use of a self-control plan, teaching around the impact of stress on cognition, identifying alternatives to aggression, and the debriefings that helped to provide a different perspective on acting out behavior with the intent to teach new skills for both staff and scholars. 124 The staff endorsed restorative practices as a theme (Peace Learning Center, Stutzman-Amstutz & Mullett, 2005; Costello, Wachtel, & Wachtel, 2010). This training looks to support scholars in processing the challenges faced in the learning environment. The structure allows staff and scholars to request a reconnect, proactive circle, responsive circle, or family group decision-making meeting to support the challenges that the staff or scholars are processing. Having a foundation in understanding states of emotion and how the brain functions supports restorative practices by allowing staff and scholars time to process and express their feelings in a productive way (Pert, 1997; Christakis & Fowler, 2009; Siegel, 2010). The foundation for advocacy supports the community in growing together. Building connections was endorsed by staff as having positively impacted this journey. Ideas discussed with the staff around building connections include creating space for positive relationships to grow. There was intentionality in the environment's design to ensure the representation of all scholars in our learning spaces. That intentionality looks like childrens books that include BIPOC and LGBTQIA+, access to meeting needs through tools in the classroom and visual structures, and intentional hiring practices that aim to diversify the staff. Staff are encouraged to slow down to build relationships by all members of the learning community. Staff are always permitted to stop what they are doing to hold a circle, have a one-on-one conversation with a scholar, allow movement, provide a snack, and do anything else needed to help strengthen that relationship. 125 Christakis and Fowler (2009) and Desautels and McKnight (2019) discussed the importance of emotional connection. Allowing teachers the time and space for this work allows scholars to share their inner world. Through this sharing, the staff created supports and plans that allowed the scholar access to the learning environment. Without this work, barriers to learning continue, and teachers become exacerbated, attempting to manage the class, the curriculum, and individual needs. The school had a system designed to allow instruction to continue for the scholars who are ready for learning and opportunities for building connections at any time for the scholars who are not yet ready for learning. The team intentionally slowed down to become more effective with the individual. Staff endorsed having a positive impact on why they remain in this work. Staff responses shared that in previous teaching experiences, they did not have support in exploring the work of mental health or ideas that supported scholars with unique needs. Staff further identified that they enjoyed working in a smaller environment where the growth of scholars is celebrated regularly. Staff commented that this is a population of students who are passed on and that the outcomes would have been bleak without programming like this. The sense of purpose that drove the staff helped them continue to show up despite the hardships. Training that supported staff retention included understanding the lines of code passed down from generation to generation, both for staff and scholars. These subconscious lines of code direct our movements without the staff endorsing it as the best response. The invitation then became how to explore 126 and move away from lines of code that create systems of control and push toward systems that create an environment of acceptance. When an escalated scholar is approached this way, solutions are discovered alongside the scholar, and more effective intervention plans are designed. The training is focused on the individual's journey over a technique or strategy. This deep, internal work was based on understanding how the brain functions for self and others and the impact of emotion on responses as reflected in the works of Jensen (1998), Siegel (2010), Wilson and Conyers (2016), and Desautels and McKnight (2019), and ProACT crisis intervention training (Fox, et al., 2022). This was a tall ask of the staff who work in the specialized day school program, as most of us were raised within systems of control. Staff expressed gratitude for being able to explore these ideas on a deeper level. The inner work supports both personal and professional growth. Regulation strategies included focusing on staff to consider the regulation strategies that work for each individual. By providing spaces for staff to regulate, the team could speak more eloquently about what might support a scholar in distress. The foundation for learning how to regulate is built from understanding where emotion comes from and how to regulate states of emotion through understanding how the brain functions, as observed in the works of Jensen (1998), Siegel (2010), Wilson and Conyers (2016), and Desautels and McKnight (2019). The team also reflected on and was honest about the workload and continued to seek solutions to balance the work's emotional needs with the job's 127 demands. It is encouraged that staff tag in/ tag out when feelings of dysregulation begin to emerge. Tagging in and out, endorsed by staff in the survey, is the staff's way of identifying that they are not helpful in a situation and affords them the opportunity to remove themselves from it (Desautels, 2020). By modeling practices around regulation, staff can better teach it at a deeper level and ensure that they are good for those they serve. Trust in administration was endorsed by the staff across all questions. In the survey results, staff identified feeling supported to innovate and having opportunities to learn from mistakes in a protected environment. Administrators were described as active, present, and committed to the work. Administration was further described as considerate, trusting, and respectful toward all staff. A paradox in this theme is that trust from the administration was also identified as an area of fear. It is easy to explain this away by recalling the change in leadership throughout these years, as noted in the changes in structure, Table 10. The paradox would suggest that there is trust and fear of administration and that this duality exists in many spaces as the group explores ideas around trust (Smith & Berg, 1987, p. 119). Overall, being a reflective practitioner appears to be woven within many of the themes outlined by the team. This showed up through statements about learning from the struggling scholar, unpacking implicit bias, working through restoration, co-regulating, and how personal this journey becomes. Being presented with the above information was not enough if staff did not choose to 128 engage in it on a personal level. As one staff responded, Its only going to work if you want it to work. 129 IMPLICATIONS AND REFLECTION The research set out to understand how a specialized, alternative day school program in a public school system in Indiana utilized professional learning and feedback gained from diagnostic work to restructure, thereby reducing the use of restraint, seclusion, and injuries that required outside medical attention while increasing academic attainment over seven years. The specialized day program worked with the researcher to identify the journey toward these changes, providing a list of the professional learning and structural changes engaged through this process. Through the ATLAS tool, small group reflections, and the survey for individual responses, the researcher identified themes the group endorsed supporting this program's evolution. The top seven themes identified were Collaboration with the Team, ProACT Crisis Intervention, Restorative Practices, Building Connections, Having a Positive Impact, Regulation Strategies, and Administrative Trust in Staff. Today, some staff reported that they continue to have doubts and fears about these changes. These came up in some of the responses, especially around ideas of safety. Change is hard. It requires the organization to be reflective and experience loss as it moves toward something new (Heifetz, 1994). Woven within the themes identified, staff endorsed that individuals must be reflective and understand how the brain functions to effectively implement several identified themes. Along the way, the interventions focused staff attention within to see the results experienced today. Staff who work in the specialized, alternative day 130 school program were hired with the understanding, and later the training, to support ideas around safety through how they show up in their work. ProACT (Fox, et al., 2022) training, identified as having the largest impact on the staff, provides time for staff to identify triggers and be aware of how stress impacts cognition. Reflecting within allows for deeper connections with scholars. As staff developed deeper connections, opportunities arose that helped staff create structures and support alongside the scholar. This arguably helped teams to create more successful plans. This came out through themes of regulation strategies, restorative practices, and collaboration. Long, et al., (2001) and Green (2014; 2016) have argued that work done with scholars as partners is more effective. This research has argued that to accomplish this, one must focus inward to be ready to receive the perspectives of scholars who have struggled with social, emotional, or behavioral development. The journey inward helps staff unpack triggers, acknowledge bias, and create tools for managing the impact of staff reactions from these spaces toward scholars. The work then lies in creating a team open to the opportunity to move away from thinking of control and toward acceptance, grace, dignity, and respect. Future research will uncover additional themes to bring into the conversation. It is anticipated that as the program continues to evolve, additional insights will emerge so that one day, the use of restraint and seclusion is eliminated, and outcomes for struggling learners are improved. As Eric Jensen (1998) described, the focus shifts to the elimination of aversive stimuli as the 131 primary focus. Future participants in this work will take the conversation further as reflection and growth from experiences continue. Future research will further explore the reduction of suspension as an aversive reaction toward scholars with emotional and behavioral challenges. This program focused on reducing restraint and seclusion as its starting point for the overall reduction in aversive practices. While further exploration and growth are needed, there is also a clear need for the reduction of incidents that rise to the level of suspendable behavior and the use of suspension as a response from the school agency. Further application of practices that support this reduction of all aversive practices continues to be a need in all school settings. 132 EXECUTIVE SUMMARY Dr. LaTonya Turner This retrospective causal-comparative study aimed to uncover some contributing factors to an evolving specialized, alternative day school program that went from a high incident of restraint and seclusion to a low one. During these changes, the program outcomes included a safer environment for staff and students and increased academic achievement. Through data analysis and surveys of staff, the research explored some of the external and internal work that shifted among staff. The goal is to uncover more of what led to the changes in philosophy and, ultimately, outcomes for the scholars of this program. The interventions put in place to make these changes occurred over the past 7 years. Already identified external factors like a change in leadership, professional learning that the community engaged with, and updates to overall structures for the program are presently available. The staff who experienced these changes provided insights for future teams contemplating similar work. Internal factors and insights to support other programs in approaching this work were identified through themes offered by the staff. The information gained through data analysis and surveys from staff was synthesized into themes. These overarching themes will then serve as the beginning of a roadmap to what is possible. It is believed that the answers already exist for each program looking to make these changes. The missing work is having a model to see that it is possible and understanding what responding staff experience as these changes in practice turn into new habits. Through 133 exploring these inner thoughts and experiences from responding staff, there are lessons learned for others to consider. It is also believed that by exploring past experiences, we can further forge a way forward with the specialized, alternative day school program. Exploring the recent past is an intervention for the group that has already come so far! 134 REFERENCES Ackerman, C.E. (2018, July 25). What is neuroplasticity? A psychologist explains. Positive Psychology. https://positivepsychology.com/ neuroplasticity/#:~:text=The%20term%20%E2%80%9Cneuroplasticity %E2%80%9D%20was%20first,widely%20used%20until%20the%201960s Adams, D., MacDonald, L., & Keen, D. (2019). Teacher responses to anxiety-related behaviors in students on the autism spectrum. Research in Developmental Disabilities, 86, 11-19. https://doi.org/10.1016/j.ridd.2018.12.009 Aldrup, K., Klusmanna, U., Ldtkea, O., Gllnerc, R., Trautwein, U. (2018). Student misbehavior and teacher well-being: Testing the mediating role of the teacher-student relationship. Learning and Instruction. 58, 126-136. https://doi.org/10.1016/j.learninstruc.2018.05.006 Alliance Against Seclusion and Restraint. (2022). Ending the school-to-prison pipeline. Alliance Against Seclusion and Restraint. https://endseclusion.org American Civil Liberties Union. (2019). Cops and no counselors: How the lack of school mental health staff is harming students. American Civil Liberties Union. https://www.aclu.org/sites/default/files/field_document/030419acluschooldisciplinereport.pdf Appleton, A. (2022, November). Indiana teachers want focus on more funding, not social issues in the 2023 legislative session. Chalkbeat. Retrieved 135 April 8, 2023, from https://in.chalkbeat.org/2022/11/22/23474295/indianateachers-unions-legislative-session-priorities-funding-staff-shortages-2023 Baker, E., Stover, E., Haar, R., Lampros, A., Koenig, A. (2020, June). Mental health and human rights. Health and Human Rights, 22(1), https://www.jstor.org/stable/10.2307/26923493 Barnett, D. (2012). A grounded theory for identifying students with emotional disturbance: Promising practices for assessment, intervention, and service delivery. Contemporary School Psychology, 16. Bassuk, E. L., & Gerson, S. (1978). Deinstitutionalization and Mental Health Services. Scientific American, 238(2), 4653. http://www.jstor.org/stable/24955635 Bermejo, L., Hernndez-Francoa, V., Prieto-Urs, M. (2013). Teacher Well-being: Personal and job resources and demands. Procedia - Social and Behavioral Sciences. 84. 132125. doi:10.1016/j.sbspro.2013.06.750. Bonner, G., Lowe, T., Rawcliffe, D., Wellman, N. (2002). Trauma for all: A pilot study of the subjective experience of physical restraint for mental health inpatients and staff in the UK. Journal of Psychiatric and Mental Health Nursing. 9. 465-73. https://doi.org/10.1046/j.1365-2850.2002.00504.x Christakis, N. A., Fowler J.H. (2009). Connected: The surprising power of our social networks and how they shape our lives- how your friends' friends' friends affect everything you feel, think, and do. Back Bay Books. 136 Costello, B., Wachtel, J., Wachtel, T. (2010). Restorative circles in schools: Building community and enhancing learning a practical guide for educators. International Institute for Restorative Practices. Couvillon, M. A., Kane, E. J., Peterson, R. L., Ryan, J. B., & Scheuermann, B. (2019). Policy and program considerations for choosing crisis intervention programs. Journal of Disability Policy Studies, 30(1), 3545. https://doi.org/10.1177/1044207318793158 Civil Rights Collection CRDC (2020, October). Civil rights data collection: The use of restraint and seclusion on children with disabilities in K-12 schools. https://www2.ed.gov/about/offices/list/ocr/docs/restraint-and-seclusion.pdf Civil Rights Data Collection CRDC. (2021, September). Restraint & Seclusion (RSTR) module. Civil Rights Data Collection. https://crdc.communities.ed.gov/#communities/pdc/documents/19919 Davis, A., Oompa. (2016). Simon says. National Poetry Slam Finals, House Slam. https://www.youtube.com/watch?v=zatHOwWBPEI de Ruitera, J.A., Poorthuisb, A.M.G., Aldrupc, K., & Koomena, H.M.Y. (2020, October). Teachers' emotional experiences in response to daily events with individual students varying in perceived past disruptive behavior. Journal of School Psychology, 82, 85-102. DeCarlo Santiago, C., Raviv, T., & Jaycox, L. H. (2018). Creating healing school communities: School-based interventions for students exposed to trauma. American Psychological Association. http://www.jstor.org/stable/j.ctv18phgmp 137 Department of Administrations Governor's Council on Developmental Disabilities. Retrieved October 16, 2022). Parallels in time: A history of developmental disabilities: Part one: The ancient era to the 1950s. https://mn.gov/mnddc/parallels/four/4b/1.html Desautels, L. (2012). How may I serve you? USA. Desautels, L., McKnight, M. (2016). Unwritten: The story of a living system: A pathway to enlivening and transforming education. Wyatte-MacKenzie. Desautels, L. (2020). Connections over compliance: Rewiring our perceptions Desautels, L., McKinight, M. (2019). Eyes are never quiet: Listening beneath the behaviors of our most troubled students. Wyatte-MacKenzie. of discipline. Wyatt-MacKenzie. Diamond, M. C., Lindner, B., & Raymond, A. (1967). Extensive cortical depth measurements and neuron size increases in the cortex of environmentally enriched rats. Journal of Comparative Neurology, 131(3), 357-364. Dima, G., Mesean Schmitz, L., & imon, M.-C. (2021). Job Stress and burnout among social workers in the VUCA world of COVID-19 pandemic. Sustainability, 13(13), 7109. MDPI AG. http://dx.doi.org/10.3390/ su13137109 Duncan, A. (2012, May). Restraint and seclusion: Resource document. US Department of Education. www.ed.gov/policy/restraintseclusion Everytown for Gun Safety. Retrieved July 3, 2022). How can we prevent gun violence in American schools? Every Town for Gun Safety https://everytownresearch.org/report/how-can-we-prevent-gun-violence-in- 138 schools/ Fisher, D., Frey, N. (2008). Better learning: Through structured teaching. ASCD. Flavell, J. H. (1979). Metacognition and cognitive monitoring: a new area of cognitive-developmental inquiry. American Psychologist, 34(10), 90611. https://doi.org/10.1037/0003-066X.34.10.906 Fogt, J., (February 21, 2020). personal communication. Fox, L.E., Johnson, L.M., & Nihart, M.A. (2022). Pro-ACT instructor manual. Pro-ACT, Inc. Fox, L.E., Johnson, L.M., & Nihart, M.A. (2022). Pro-ACT instructor manual: Restraint certification. Pro-ACT, Inc. Gaines, L. V. (June 7, 2023). Indiana parents are kept in the dark as schools isolate and restrain kids thousands of times. NPR, WFYI Indianapolis, PBS. Ginott, H. (1972). Teacher and child: A book for parents and teachers. Avon Books. Goldstein, A. (1988). Prepare curriculum. Research Press. Green, R. (2016). Lost & found: Helping behaviorally challenged students. Jossey-Bass. Green, R. (2014). Lost at school: Why our kids with behavioral challenges are falling through the cracks and how we can help them. Scribner. Hammond, Z. (2015). Culturally responsive teaching and the brain: Promoting authentic engagement and rigor among culturally and linguistically diverse students. Corwin. 139 Health Insurance Portability and Accountability Act (HIPPA). (1996, August 21). Public Law. 104-191. Heifetz, R. A. (1994). Leadership without easy answers. Belknap Press. Henderson, K., Klein, S., Gonzalez, P., & Bradley, R. (2005). Teachers of children with emotional disturbance: A national look at preparation, teaching conditions, and practices. Behavioral Disorders, 31(1), 617. http://www.jstor.org/stable/23890397 Holsapple, N. (2021, May 3). Memorandum: Special education child count. https://www.in.gov/doe/files/memo-sboe-unduplicated-child-count-may-20 21.pdf Indiana Department of Education. (2022, May 12). All assignment codes sections. https://www.in.gov/doe/files/All-AC-Sections-5.12.2022.pdf Indiana Department of Education Office of Special Education. (2019, May). Indiana state board of education: Special education rules, Title 511, Article 7, Rules 32-49. https://www.in.gov/doe/files/art-7-english-may-2019-update-index.pdf Individuals with Disabilities Education Act (IDEA), (1997). United States Congress 1412 [a] [5]. (Reauthorization of the Individuals with Disabilities Education Act of 1990). Individuals with Disabilities Education Improvement Act of 2004, P.L. 108-446, 20 U.S.C. 1400 et seq. (Reauthorization of the Individuals with Disabilities Education Act of 1990). Jensen, E. (1998). Teaching with the brain in mind. ASCD. 140 Kagan, L., Kagan, S., Kagan, M. (2015). Kagan structures. Kagan Publishing. Kauffman, J. M., & Badar, J. (2013). How we might make special education for students with emotional or behavioral disorders less stigmatizing. Behavioral Disorders, 39(1), 1627. http://www.jstor.org/stable/23890740 Kazdin, A. E. (2008). Evidence-based treatment and practice: New opportunities to bridge clinical research and practice, enhance the knowledge base, and improve patient care. American psychologist. 63(3), 146-159. Kern, L., Evans, S. W., & Lewis, T. J. (2011). Description of an iterative process for intervention development. Education & Treatment of Children (West Virginia University Press), 34(4), 593617. DOI: 10.1037/0003066X.63.3.146 https://doi.org/10.1353/etc.2011.0037 Kern, L., George, M., Weist, M. D. (2016). Supporting students with emotional and behavioral problems: Prevention & intervention strategies. Paul H Brookes Publishing. Kuypers, L. M. (2011). The zones of regulation: A curriculum designed to foster self-regulation and emotional control. Think Social Publishing. Larrison, S., Mansfield, T. (2018, October 4). Retrieved January 29, 2021. Multitiered systems of support (MTSS). Indiana Department of Education. Lee, J. H. (2013). School shootings in the US public schools: Analysis through the eyes of an educator. Review of Higher Education & Self-Learning, 6(22). Lendzion, S., (2022). personal communication. Livingston, M.D., Rossheim, M.E., Stidham Hall, K. (2019, June). A descriptive 141 analysis of school and school shooter characteristics and the severity of school shootings in the United States, 19992018, Journal of Adolescent Health 64(6). 79799, https://doi.org/10.1016/j.jadohealth.2018.12.006 Livingston, R. B. (1966). Brain mechanisms in conditioning and learning (Ser. Neurosciences research program bulletin, v. 4, no. 3). Long, N., Wood, M.M., Fecser, F.A. (2001). Life space crisis intervention: Talking with students in conflict. Pro-Ed. Lurye, S. (2020, July 26). How 'the most violent' special education school ended restraint and seclusion. USA Today. https://www.usatoday.com/story/news/education/2020/07/25/disabilitykids-private-special-education-school-centennial-no-restraint/5513768002/ Marzano, R. J. (2007). The art and science of teaching: A comprehensive framework for effective instruction. ASCD. McDowell, M. (2018), Specific learning disability. J. Paediatr Child Health, 54, 1077-1083. https://doi.org/10.1111/jpc.14168 McGonigal, K (2013). How to make stress your friend. TedGlobal. https://www.ted.com/talks/kelly_mcgonigal_how_to_make_stress_your_fri end?language=en Meldon,P. (2017, October 31). Disability history: Early and shifting attitudes of treatment. National Park Service. http://www.nps.gov/articles/disability historyearlytreatment.htm Mohr, W. K., Mahon, M. M., & Noone, M. J. (1998). A restraint on restraints: The need to reconsider the use of restrictive interventions. Archives of 142 Psychiatric Nursing, 12(2), 95106. https://doi.org/10.1016/S0883-9417(98)80059-9 Moore, S. (2016, April 4). Transforming inclusive education. Social Science and Human Research Council: Council of Research Science and Humanities. https://www.youtube.com/watch?v=RYtUlU8MjlY Musu, L., et al. (2019). Indicators of School Crime and Safety: 2018. Justice Research and Statistics Association. https://ncvc.dspacedirect.org/handle/20.500.11990/1242 National Alliance on Mental Illness. (2019, August 5). The truth about mental health and gun violence. Retrieved August 15, 2022. NAMI. https://namica.org/advocacy/criminal-justice-advocacy/the-truth-aboutmental-health-and-gun-violence/ National Association of School Resource Officers. (Retrieved September 17, 2022). National association of school resource Officers, FAQ: https://www.nasro.org/faq/ National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention. (2023, March 8). Improving access to childrens mental health care. Center for Disease Control. https://www.cdc.gov/childrensmentalhealth/access.html Nunno, M.A., McCabe, L.A., Izzo, C.V., Smith, E.G., Sellers, D.E., Holden, M.J., (2021, August.11). A 26 year study of restraint fatalities among children and adolescents in the United States: A failure of organizational structures and processes. Children and Youth Care Forum. 51, 661680. 143 https://doi.org/10.1007/s10566-021-09646-w Office of Civil Rights.(2022) Pending cases currently under investigation at elementary - secondary and post-secondary schools as of September 2, 2022 7:30 am search. Retrieved September 25, 2022. https://www2.ed.gov/about/offices/list/ocr/docs/investigations/openinvestigations/dis2.html Olfson, M., Blanco, C., Wang, S., Wang, S., Laje, G., Correll, C. (2014, January). National trends in the mental health care of children, adolescence, and adults by office based physicians. JAMA Psychiatry, 71(1), 81-90. doi:10.1001/jamapsychiatry.2013.3074 Oliver, R. M., & Reschly, D. J. (2010). Special education teacher preparation in classroom management: Implications for students with emotional and behavioral disorders. Behavioral Disorders, 35(3), 188199. https://doi.org/10.1177/019874291003500301 Peace Learning Center (2021). Peace Learning Center. https://peacelearningcenter.org/ Pert, C. (1997). Molecules of emotions. Charles Scrbner's Sons. Ramachandran, V. (2009). The neurons that shaped civilization. TEDIndia. https://www.ted.com/talks/vilayanur_ramachandran_the_neurons_that_sh aped_civilization?language=en Rambo, S. E., & St. Mary's College of California. Kalmanovitz School of Education. (2020). Spare the rod, educate the child: moving away from zero-tolerance policies in schools (dissertation). Saint Mary's College of 144 California. Redmond, H. (1998). The state of mental health in the USA. International Journal of Mental Health, 27(1), 6279. http://www.jstor.org/stable/41344855 Resnick, A. (2021, October 6). What does it mean to be neurodivergent? Very Well Mind. https://www.verywellmind.com/what-is-neurodivergenceand-what-does-it-mean-to-be-neurodivergent-5196627 Richard, T., & Sanders School and ProACT definitions of dangerous behavior. (2022). IR Guidance 22-23. Rockhill, C., Violette, H., Stoep, V., Grover, S., Myers, K. (2013, November. 6). Caregivers distress: Youth with attention-deficit/hyperactivity disorder and comorbid disorders assessed via telemental health. Journal of Child and Adolescent Psychopharmacology, 23, 379-385. doi: 10.1089/cap.2013.0019 Rojas-Velasquez, D.A. (2017, January) The evolution of restraint in American psychiatry. Yale University Thesis Digital Library. https://elischolar.library.yale.edu/cgi/viewcontent.cgi?article=2167&context =ymtdl Romine, T., Sturla, A. (2021, October 7). Three staff members charged in death of 16-year-old who went into cardiac arrest after being restrained at facility. CNN. https://www.cnn.com/2020/06/25/us/teen-restraint-deathstaff-charged-michigan-trnd/index.html Roseth, C.J., et. al. (2008, November.). Teacher intervention and U.S. 145 preschoolers' natural conflict resolution after aggressive competition. Behavior: Natural Conflict Resolution in Humans, 145(11), 1601--26 https://www.jstor.org/stable/40295891 School Reform Initiative. Retrieved 2023, February 11). ATLAS -Looking at data protocol. Network for College Success, 218-21. https://toandthrough.uchicago.edu/sites/default/files/uploads/documents/N CS_FOT_Toolkit_ISBT_SetB_ATLAS%20Data%20Protocol.pdf Schuyler, K. (2023, May 10). Brownsburg school officials hold meeting with special education parents following staffers arrest. Retrieved 2023, May 16. Fox 59. https://fox59.com/indiana-news/brownsburg-school-officialshold-meeting-with-special-education-parents-following-staffers-arrest/ Serrin, G. (July 22, 2022). Case advances against El Dorado Hills school staff in death of child with autism. KCRA3. https://www.kcra.com/article/case-advances-against-el-dorado-hillsschool-staff-in-death-of-child-withautism/40692074#:~:text=Max%20Benson%20died%20in%20November,a nd%20died%20two%20days%20later Shah, A. A., & Beinecke, R. H. (2009). Global mental health needs, services, barriers, and challenges. International Journal of Mental Health, 38(1), 1429. http://www.jstor.org/stable/41345271 Siegel, D. (2010). Mindsight. Bantman Books. Signer, J. (2017). NeuroDiversity: The birth of an idea. Singer. Lexington, Ky. Smith, K.K., Berg, D. N. (1987). Paradoxes of group life: Understanding 146 conflict, paralysis, and movement in group dynamics. Jossey-Bass. Sprick, R., & Borgmeier, C. (2010). Behavior prevention and management in three tiers in secondary schools. The Charlesworth Group, Wakefield. Sprick, J., Sprick, R., Edwards, J., Coughlin, C. (2021). CHAMPS: A proactive & positive approach to classroom management. Ancora. Stutzman-Amstutz, L., Mullett, J.H. (2005). The little book of restorative discipline for schools: Teaching responsibility' creating caring climates. The Little Books of Justice & Peacebuilding. US Department of Education. (2024, February 16). A history of the individuals with disabilities education act. Ed IDEA, https://sites.ed.gov/idea/IDEAHistory#2000s-10s Wei, Y., Szumilas, M., & Kutcher, S. (2010). Effectiveness on mental health of psychological debriefing for crisis intervention in schools. Educational Psychology Review, 22(3), 339347. http://www.jstor.org/stable/23364147 Weisburst, E.K. (2019, February 7). Patrolling public schools: The impact of funding for school police on student discipline and long-term education outcomes, Journal of Policy Analysis and Management. https://doi.org/10.1002/pam.22116. Wilson, D. Conyers, M. (2016). Teaching students to drive their brains: Metacognitive strategies, activities, and lesson ideas. ASCD. Yergeau, M. (2018). Authoring autism. On rhetoric and neurological queerness (pp. 135174). Duke University Press. http://www.jstor.org/stable/j.ctv1168bxt.7 147 Zehnder, I. (2006, August 2). Death of 7-year old Angellika Arndt results in closure of Northwests Rice Lake clinic. WebWire. https://www.webwire.com/ViewPressRel.asp?aId=17798 148 APPENDICES Appendix A: INCIDENT REPORT (IR) SUBMISSION GUIDANCE Rationale: At times it is necessary to document an incident or series of incidents to help capture all perspectives. It then becomes necessary for staff to be able to document all of the ways that we worked to intervene and support that need after the event transpired. The Incident Report (IR) is Sanders' system for documenting these escalations. Locations: A blank copy of the IR is located in each team folder. How to use: 1. Staff are to make a copy of the report and change the report name to the date of the incident, scholars initials, IR. Example: 2/16/21 TR IR 2. Staff then begin writing up the incident. Only write what you did. 3. Stick to factual statements. Each staff member will add their part to the incident. If more than one staff member was involved in a singular part to the incident, it is appropriate for staff to collaborate and ensure that all parts were gained. 4. Staff do need to tag all persons involved in the incident. This helps to alert others that the IR has already been started. 5. Staff may add lines within the document at any time to clarify and add details to the incident. 6. Always complete parent communication 7. Always complete requested actions. 149 8. Always tag classroom teachers when an IR is submitted for an incident outside of their classroom. 9. Always have the Nurse complete the assessment upon completion when use of restraint or seclusion 10. Always tag admin on any IR submitted. Unsure what happened? Any staff member may ask clarifying questions of any IR submitted. This practice is encouraged. When to submit an IR The following descriptors offer a level for the behaviors that occurred. This reflects the top part of the escalation cycle. IRs are always submitted for level 4 behaviors. IRs are submitted at teacher discretion for any level 3 behaviors. Level 1 & 2 behaviors would not likely have an IR submitted. Always submit an IR when restraint and/or seclusion are utilized in response to behavior. Always request a ProACT debriefing when restraint and/or seclusion are utilized in response to behavior. Always have the Nurse see the scholar who has been restrained or secluded for an assessment. 3rd Nine Weeks definitions and coding beginning 01/2023 Physical Aggression (definitions apply to self-harm as well) 1- Challenging - Toward Object (P1) 2- Minor Threat - Toward person (P2) 3- Minor Attempt - Toward person (P3) 4- Major Threat or Attempt - Toward person (P4) 150 Sexualized Behavior 1 - Sexualized comment not directed at anyone. (S1) 2 - Sexual Minor Threat - Sexualized comment directed at someone specific. (S2) 3 - Sexual Minor Attempt - Sexual comment and gesture directed at someone specific. (S3) 4 - Sexual Major Threat or Attempt - Sexual unwanted contact made. Can include a sexual comment that is specific and directed toward one individual (S4) Verbal Aggression 1- Class Disruption (on going, after redirection) (V1) 2- Minor verbal threat (V2) 3- Minor verbal attempt. Derogatory language directed at a group. (V3) 4- Specific or serious and plausible threats or attempts. Includes derogatory language (racist, sexist, homophobic, etc) directed at an individual. (V4) Elopement 1- Out of Area within the Classroom (E1) 2- Elopement Minor Threat - Out of Classroom (E2) 3- Elopement Minor Attempt - Left Building (E3) 4- Elopement Major Threat or Attempt - Left Property (E4) What is a running IR? 151 Running IR is when there are several smaller escalations throughout the day where the team needs to capture how the day is unfolding in order to properly plan to meet the scholars needs. The teacher would start this at the beginning of the day and request that staff add to it as the day unfolds. It is important to note the time of each small incident throughout the day to assist with planning. When conducting a running IR, be sure to note times of success throughout the day. These insights are just as crucial to understanding the complexities around a scholar as where things are not going well. Multiple scholars When multiple scholars are involved in a singular incident, it is appropriate to write one IR. Then upon the conclusion of the team completing their work within that IR, the TOR will copy it and place a copy in each scholars folder. IRs are submitted. Now what? Take the time periodically to review IRs completed to help look for trends. Review IRs to review any skills teaching that needs additional attention Use IRs to support communication to physicians and outside agencies at parent request o Black out names physically on the doc prior to printing, including initials 152 Appendix B: INCIDENT REPORT Individuals Involved: Descriptors of Incident Start Time: End Time: Date: Location(s): Report Written: Student Name: Witness(es): Antecedents: Student Presented with: Student Requested: Student Denied: Accommodation Provided: Support Provided: What Happened: Student Observable Behavior: Staff Response: Student Observable Behavior: Staff Response: Student Observable Behavior: Staff Response: Behavior highlighting: Minor Threat, Minor Injury, Major Threat or Injury Student Voice: Staff who debriefed with the scholar: Scholar expressed: Injury / Property Damage: Insights gained from staff perspective: 102 Nurse: Parent Contacted: Date: Time: Spoke with: Action Taken or Requested for Team: ProACT debriefing SERT assessment Circle Reconnect / Relate Restoration within the classroom (Specify below) Restoration outside of the classroom (Specify below) Adjustment to primary plan as indicated in MTSS & level up review Family group decision meeting Reconvene to adjust BIP / Crisis plan Actions Requested for Administration: Working the plan. Admin, please be in the know. Administrator Immediate Attention Requested Notes regarding the requested actions: Restoration Activity Assigned: *Required after suspension, restraint, or seclusion. Retag admin. upon completion. Who participated? Outcomes: Agreements: Reflections: Ah-has: Fidelity Practice w/ Staff: Addition to Intervention Plan: Other: Physical Techniques Utilized Evasion Technique Restraint 103 Shield up Back Step Pivot Communication Moved outside striking range Minimize Release Evade Crouch Sidestep Crisis Seclusion Start Time: Student behavior during seclusion: Stop Time: Staff Involved in Seclusion: Include Start time: Student Behavior During the Restraint: Stop Time: Staff Involved: 104 Appendix C: DIRECTIONS FOR INCIDENT REPORT REVIEW 1. Frequency and intensity of reports submitted based on the definition described above will be noted at levels three and four only. 2. Only one count will be recorded for each incident. 3. The highest level of dangerousness will be collected as a frequency count from incident reports that were submitted. 4. The incident will be recorded in a row labeled for that school year. 5. Damage to property is not included in physical attacks unless it involves striking a person. No matter how bad the property damage is, it is not recorded for the purposes of this review. 6. Failed attempts at contact made to people are level two and, therefore, are not counted for the purposes of this review. 7. If multiple level four behavior types are within one incident report, physical attack where contact is made will be recorded for the purposes of this review. For example, making physical contact with self or another in a dangerous way is more dangerous than making a threat of harm that raises to a level four behavior when there is not access to the weapon being threatened in that direct experience. 8. An attempt to record the staff response after the incident in order to support the scholar. The categories are: Reconnect, Debrief, Staffing, Practice Replacement Behaviors, Family Group Decision Meeting (FGDM), and Update Plan. 105 9. If nothing was recorded, largely because this did not begin to be tracked until the 2021-2022 school year, no recording for response will be made. Only what was reported will be recorded. 106 Appendix D: INTERVIEW PROTOCOL Interviewee: _______________________________ Interviewer: _______________________________ Written Consent Collected: YES / NO Introductory Protocol Thank you for agreeing to help discover all that has happened to shift the work at our school program since 2016. This study aims to uncover what supported staff in moving from a system of control to one of support and acceptance. The interview survey is designed to take up to one hour. The survey does not collect identifiable information. The only personal question asked is for you to select a range of years of experience working in this program. You may pass on any questions; no question is required for submission. The interview survey will be recorded in the researcher's personal email. You can leave if you wish to be excused from the interview survey. The researcher has left the building to ensure that you can participate in the best way. Passing or being excused from the interview survey will be done with dignity and respect. No consequence will come for leaving the session. The intent is to allow staff to help capture and tell the tale of our journey for other similar programs to explore. We will begin after staff are given a chance to exit if they choose. Introduction Thank you for agreeing to participate in this interview survey. We look forward to learning from the collective who has experienced and is continuing to experience 107 these changes in our programming. Please complete the survey now. I am available for any questions you have. 108 APPENDIX E: INFORMED CONSENT Informed Consent Title of Research: A Historical View of How One Specialized, Alternative Day School Program Moved Away from the Practices of Restraint and Seclusion Principle Investigator: Taryn Richard Affiliations: MSD of Wayne Township Contact Information: taryn.j.richard@gmail.com Institutional Contact: Institutional Review Board Marian University Christina Pepin irb@marian.edu 1. Introduction and Purpose of the Study This retrospective causal-comparative research study aims to uncover the factors that impacted responding staff in approaching an escalated scholar in a specialized, alternative education day school program over seven years. The study aims to capture the story to help other programs make similar changes. 2. Description of the Research The years before the shift in philosophy and practice in the identified program and local area specialized alternative education day school programs will serve as the controls for this study. Each subsequent year will serve as the comparison group. The researcher will look at data through the years and present this to the staff of the specialized day program for their perspective. The staff will be able to provide insights and reflections that are above just knowing what the group engaged with over these last seven years in professional learning and in-house review of student outcomes. 3. Subject Participation Staff working in the specialized day program will review the data collected and provide the researcher with insights into the program's evolution. No identifiable information is collected from students or staff for this research study. 4. Potential Risks and Discomforts There are no known risks to providing the researcher with insights into the evolution of this program. The researcher has taken measures to protect staff from being identified through this processprotective measures described in item 6. 5. Potential Benefits This research aims to provide similar programs with a path to find similar results. Participating in the study allows the staff to share this fantastic work and the outcomes this collective has experienced. 6. Confidentiality Participants will complete a group reflection using the ATLAS tool. This will be shared in a Google Doc. One member of each small group will copy the document and add their 109 group's reflections. This will be printed and handed to the program's secretary to prevent the researcher from knowing which reflective doc comes from which group. The individual survey is submitted anonymously and with a third-party proctor in a meeting where the researcher will not be present. Submitted individual survey responses are kept with the researcher's personal account, not the work account affiliated with the program's organization. 7. Voluntary Participation and Authorization Participation in this research is voluntary. There is no penalty for not participating. The data for the program's evolution is collected. This request is to gain insights into the lived experiences of those who work within the program in its current iteration. 8. Compensation There is no compensation for participation in this research. 9. Withdrawal from the Study Participants may withdraw from the research at any time with no penalty. They may stop participating, listen to the conversation, or leave the space anytime. I voluntarily agree to participate in this research program. o o Yes No I understand I will be given a copy of this signed Consent Form. Name of Participate (print)___________________ Date: ______________ Signature: _________________________ Proctor for Survey (print) Jaime Wright Date: ______________ Signature: _________________________ Investigator (print) Taryn Richard Date: ______________ Signature: _________________________ 110 Appendix F: INTERVIEW QUESTIONS 1. Upon reviewing the school data from before the changes, through the changes, and seeing similar data from similar programs, what impressions do you have about the journey to restructure the school? 2. What has been the most impactful professional learning you engaged with to create this change? 3. What have been the most impactful changes to structure? (Consideration structure, instructional practices, crisis response, connection, and restoration)? 4. What has been the most impactful change you have experienced in what you observed of your colleagues? (Consideration structure, instructional practices, crisis response, connection, and restoration)? 5. What makes this school a place where you want to continue to work? 6. What advice do you have for a school considering similar work so that they are equipped to begin? 111 Appendix G: ATLAS TOOL WAYNEW3 DataDive Protocol Norms We DO: own it, explain it, interrogate it, validate it We DONT: judge it, dismiss it Data Source: Predict (5 minutes) What do you anticipate seeing in this data set? Describing the Data (10 minutes) What do you see (strengths, deficits, wonderings)? Avoid judgments about quality or interpretations Identify where the observation is being made (e.g., On page one in the second column, third row) Interpreting the Data (10 minutes) What does the data suggest? 3-5 Whys about each suggestion. What are the assumptions we make about that interpretation? o Try to make sense of what the data says and why. Find as many different interpretations as possible and evaluate them against the kind and quality of evidence. o Try to infer what is being worked on and the reason. o Think broadly and creatively. Assume that the data, no matter how confusing, makes sense to some people; your job is to see what they may see. As you listen to each others interpretations, ask questions that help you better understand each others perspectives. Implications for Classroom Practice (10 minutes) What are the implications of this work for teaching and assessment? Consider the following questions: o What steps could be taken next? o What strategies might be most effective? 112 What else would you like to see happen? What does this conversation make you think about in terms of your own practice? About teaching and learning in general? o What are the implications for equity? Phrase implication statements: ____ (stakeholder group) need(s) to _____ (action) o o Reflecting on the ATLAS-Looking at Data (10 minutes) What did you learn from listening to your colleagues that was interesting or surprising? What new perspectives did your colleagues provide? How can you make use of your colleagues perspectives? Did questions of equity arise? How can you pursue these questions further? Are there things you would like to try as a result of looking at this data? Debrief the Process (5 minutes) How well did the process work? What about the process helped you to see and learn interesting or surprising things? What could be improved?  ...

... DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 Running head: SRNA GASTRIC US EDUCATION 1 Marian University Leighton School of nursing Doctor of Nursing Practice Final Project Report for Students Graduating in May 2024 Education for Student Registered Nurse Anesthetists on Preoperative Ultrasound Guided Assessment of Gastric Content Stephen Sai Schandorf Marian University Leighton School of Nursing Chair: Dr. Derrianne Monteiro _____________________________ (Signature) Project Team members: Dr. Vadim Korogoda ______________________________ (Signature) Date of Submission: May 17, 2024 DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 2 Table of Content Abstract3 Introduction..4 Background..5 Problem Statement...7 Organizational Gap analysis 7 Review of Literature8 Theoretical framework/Conceptual model13 Aims/Goals and Objectives...14 Project Design Methods.16 Project site and Population....17 Statistical Tests..17 Ethical Consideration....17 Data Analysis and Results.18 Discussion .20 Conclusion.21 References.23 Appendix A25 Appendix B....26 Appendix C27 Appendix D28 Appendix E38 Appendix F41 DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 3 Abstract Background: Since its introduction by Mendelson in 1946, preoperative fasting has been utilized to produce an empty stomach and decrease the risk of aspiration in the surgical patient. Patient adherence to NPO recommendations, certain comorbidities, and/or medications that can decrease gastric motility increases the risk of aspiration. Additionally, anesthesia induction drugs blunt airway reflexes making patients susceptible to reflux and possible aspiration, resulting in adverse outcomes. Identification of patients at increased risk and prevention of aspiration is therefore imperative for the CRNA to achieve successful perioperative outcomes. Ultrasonography, a safe noninvasive tool frequently used by anesthesiologist can be utilized in identification of patients at increased risk of aspiration. It is however currently underutilized. Purpose: This DNP project aims at teaching SRNAs how to perform an ultrasound gastric assessment as well as develop a check sheet to guide performance of the gastric ultrasound scan (GUS) in order to increase use and patient safety. Method: A 30 minute voice over instructional PowerPoint together with a pretest/posttest survey was deployed to all registered Marian University SRNAs with instructions to complete the pretest prior to reviewing the PowerPoint tutorial and the post test afterwards. Results: Participants knowledge based scores significantly increased from the pretest (M = 50.5, SD = 14.4) to the post test (M = 93.8, SD = 9.3; t = -11.1, p < .001, d = -2.86). Additionally SRNA confidence in performing a GUS significantly increased from the pretest (M = 6.6, SD = 18.5) to the post test survey (M = 57.2, SD = 19.6; t = -7.99, p < .001, d = -2.06). Conclusion: Student participation in the DNP project significantly increased their knowledge on the gastric ultrasound assessment procedure and their confidence for performing the procedure. Keywords: Gastric ultrasound, gastric content, gastric volume DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 4 Education for Student Registered Nurse Anesthetists on Preoperative Ultrasound Guided Assessment of Gastric Content This project is submitted to the faculty of Marian University Leighton School of Nursing as partial fulfillment of degree requirements for the Doctor of Nursing Practice, CRNA track. Anesthesia induction drugs blunt the airway reflexes and diminish the tone of the lower esophageal sphincter (LES) making patients increasingly susceptible to reflux of abdominal content and possible aspiration into the lungs. Pulmonary aspiration of gastric content is a potentially fatal complication of anesthesia during surgical procedures (Reed & Haas, 2020). Pulmonary aspiration is defined as the entry of liquid or solid material into the trachea and lungs, anesthesia-related aspiration occurs when patients without sufficient laryngeal protective reflexes passively or actively regurgitate gastric contents (P.302). Almost half of all patients who aspirate during surgery develop a related lung-injury, such as hypoxia and aspirational pneumonia (Nason, 2015). 10% to 30% of anesthesia related deaths are attributed to aspiration (Reed & Haas, 2020). Per the definition of pulmonary aspiration, increased gastric content resulting in intra procedure emesis can result in pulmonary aspiration, hence should be prevented. Reduction of pulmonary aspiration in surgical patients is a key component of anesthesia practice and the primary goal of preoperative fasting. Fasting guidelines are aimed at producing an empty stomach to reduce the risk of emesis and aspiration. However, adherence to these guidelines is self-reported, which poses some level of uncertainty. Furthermore, specific patient medical and physiological conditions such as diabetes, GERD, hiatal hernia, gastrointestinal obstruction, obesity, sympathetic activation, pain, anxiety and some specific medication therapy can delay the gastric transit time or increase gastric secretions thereby increasing the risk of emesis and/or aspiration under anesthesia even with sufficient fasting. For DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 5 these reasons, the assessment of risk of intra-operative emesis while under anesthesia is somewhat challenging, necessitating a more objective method of assessment. Gastric ultrasonography is a noninvasive and reliable method that can be utilized to assess the qualitative and quantitative nature of gastric content. This method will provide objective information that will inform the anesthesia practitioner on how best to prevent aspiration in patients found to be at high risk. Currently this proven and reliable method of qualitative and quantitative gastric assessment is underutilized in practice. Practicing anesthesia providers agree ultrasound is a great tool for assessing gastric content, but view it as an additional step in their routine and hence do not utilize it. A new approach is therefore needed to improve the utility of this valuable assessment tool. By teaching student registered nurse anesthetists (SRNAs) the knowledge and skills to perform ultrasound guided gastric assessments, newly graduated CRNAs will go out into various practice settings equipped and ready to use their skills, and advocate for its use. The aim of this DNP project is to develop a procedure checklist and teach SRNAs the necessary knowledge and skills to perform an ultrasound guided gastric content assessment. Background According to Reed & Hass (2020), the incidence of pulmonary aspiration varies in the literature from 0.1% to 19% in the adult population. Although the incidence is low, pulmonary aspiration is a serious complication of anesthesia, accounting for 10% to 30% of anesthesia related deaths (Reed & Haas, 2020). Other aspiration related complications including hypoxia and pneumonitis results in prolonged hospital stays, increased healthcare cost, and a decreased quality of life (Reed & Haas, 2020). Historically, the practice of preoperative fasting originated from Mendelsons 1946 study of 44,016 patients showing a 0.15% incidence of pulmonary DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 6 aspiration (LaSala et al., 2020). The current ASA practice guidelines (2017) for preoperative fasting recommends fasting periods ranging from 2 to 8 hours depending on types of food consumed, (Reed & Haas, 2020). In an ideal situation the ASA fasting guidelines provides sufficient time for the stomach to be empty, so as to prevent emesis and/or aspiration during surgical procedures under anesthesia. However, medical and physiological conditions as well as certain medications that delay the gastric emptying time or increase gastric secretions may render patients in a state of increased gastric content with an increased risk for aspiration. According to Nason (2015), the severity of lung parenchyma damage is dependent on the degree of acidity, the volume of the aspirate, and the presence or absence of particulate matter in the aspirated fluid. As little as 50 ml of very low PH regurgitated gastric contents or aspirated material containing particulate matter can be considered a severe aspiration risk. Feighery et al. (2023) also concluded that retained food, the use of monitored anesthesia care (MAC) and general anesthesia (GA) were associated with significantly increased risk of aspiration in patients undergoing esophagogastroduodenoscopies (EGD). Anesthesia related aspiration can be fatal, as such; strategies for preventing occurrence are imperatives for the anesthesia provider. A range of preventive measures including proton pump inhibitors, antihistamine, antacid and gastric pro-kinetic medications as well as rapid sequence induction (RSI) can be employed by anesthetists to prevent pulmonary aspiration and decrease sequelae. In order to inform the anesthesia provider what strategies best suits a particular patient, an objective assessment method is needed. Point of care ultrasonography is a technique familiar to anesthesia providers in the area of regional anesthesia. Ultrasound has been shown as a safe, non-invasive and reliable technique for assessment of gastric content (Evain et al., 2022). Routine use of point of care gastric ultrasound in the preprocedure assessment will DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 7 provide valuable information about the volume and/or quality of gastric content and enable the anesthesia provider to assess the risk of aspiration and better inform the anesthesia provider on the most appropriate aspiration preventive measure. Gastric ultrasound adds objectivity to the subjective, self-reported NPO status and introduces another layer to maintaining patient safety during the perioperative assessment of aspiration risk. Problem statement Anesthesia related aspiration can be fatal; as such, strategies for preventing occurrence are imperatives for the anesthesia provider. Even patients who adhere to the preprocedure fasting guidelines may have medical conditions, take medications and/or be in a physiological state that decreases gastric motility or increases gastric secretions. Without a quantitative and qualitative method of determining gastric content, a true assessment of the risk for emesis/aspiration is therefore challenging for the anesthesia provider during the pre-procedure assessment. Gastric ultrasound offers a safe, non-invasive and reliable technique for assessment of gastric content. A barrier to the use of US guided gastric assessment is the knowledge and skill set needed to perform a proper assessment. This leads to my PICOT... does providing SRNAs with a procedure checklist and teaching needed skills improve student knowledge and confidence to perform an US guided gastric content assessment? Gap Analysis/Needs Assessment Anesthesia providers including SRNAs are keenly aware of the potential danger of an aspiration event during anesthesia. Pre-procedure fasting is the current standard method used to allow time for the stomach to empty before a surgical procedure. However patient comorbidities, physiology and medications can slow gastric motility resulting in residual food, or increase DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 8 secretions, both of which increase the risk of aspiration even in seemingly healthy people. Despite patients reportedly adhering to overnight pre-procedural fasting guidelines, gastric residual food was identified in more than 3% of all patients attending for EGD (Feighery et al., 2022). CRNA education curriculum includes the ASA NPO guidelines, which patients must adhere to decrease the risk for aspiration. However, compliance is self-reported. Based on information provided by the patient of their adherence to NPO, medical diagnoses, and medications taken, anesthesia providers must predict a patients risk for an aspirational event. An objective method for assessing risk of aspiration eliminates these challenges to the anesthesia provider, improving the determination of risk, as well as patient safety overall. According to Tankul et al. (2022), various studies have shown gastric sonography to be highly satisfactory as a reliable source of valuable information of the quality and quantity of gastric content when used by experienced providers, and is also relatively easy to learn. Currently a full tutorial dedicated to use of ultrasound to assess gastric content is not included in the curriculum at the project site. Providing students with the knowledge and skills needed to perform the gastric ultrasound assessment will enhance the curriculum at the project site, improve students confidence, and enhance patient safety in practice during the perioperative period. Point of care ultrasound is a standard of practice with anesthesia providers in the area of regional anesthesia and is a tool that is frequently used with a high level of proficiency. Hence the use of point of care ultrasound applied to assessment of gastric content in the context of pre-procedure evaluation of aspirational risk can easily be taught to SRNAs. DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 9 Literature Review A literature review was conducted in December 2023 using the PUBMED and CINAHL databases, for studies involving the use of ultrasonography for assessment of gastric content. The search words gastric ultrasound, assessment of gastric volume and measurement of gastric volume as well as the BOOLEAN phrase ultrasound AND gastric volume was also used. Inclusion and Exclusion Criteria The various combinations of searchers yielded more than 302 results from 2017 to 2023. The article titles and abstracts were screened for inclusion of studies on preoperative ultrasound assessment of gastric content in relation to NPO fasting guidelines. Articles included in this literature review were primary research carried out in the pediatric and adult population that evaluated preprocedure gastric content using ultrasonography. Duplicate search results were removed and studies with indication other than preoperative gastric volume assessment were excluded. Studies related to pregnancy, neonates and infants < 2 years old were also excluded. A total of 19 articles were included in the literature review (Appendix C is a PRISMA chart of search results). Outcomes measured The studies in the literature measured the preoperative gastric antral cross sectional area, this was used in calculating the gastric volume. Some studies also reported a qualitative assessment of gastric content that graded the stomach as empty, or having clear fluid, thick fluids or solids. Secondary outcomes such as emesis, gastric PH, patient anxiety and pain were also reported DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 10 Summary of the Literature A total of 57 articles were screened for inclusion, of which 19 were included in the literature review (see Appendix C). A breakdown of the included articles is as follows; three (3) articles representing 15.8% of the articles included in the review assessed the accuracy of using ultrasound to determine gastric volume as compared to gastric suctioning. One (1) article (5.3%) studied using ultrasound to assess the volume of ingested fluid, and another 5.3% (1 article) assessed ultrasound versus NPO patients. Three (3) articles (15.8%) studied ultrasound assessment of volume of an ingested fluid over a time period. Three (3) articles (15.8%) studied ultrasound of NPO status versus ingested fluid volume, 3 articles (15.8%) studied ultrasound assessment of NPO patients with conditions that delay gastric emptying versus NPO patients without delayed gastric conditions, 2 articles (10.5%) used ultrasound to assess gastric volume after different periods of fasting (gastric volume from time of last intake), 2 articles (10.5%) studied ultrasound gastric assessment of NPO patients after a period of chewing gum and 1 article studied gastric volume and PH of gastric content. Support for Use of Ultrasound for Gastric Assessment All the studies reviewed utilized ultrasound as a comparative measure for assessing gastric content and/or volume. In fact all the articles reviewed supported the use of ultrasound and concluded that ultrasound is either equally accurate, or a superior tool for assessment of gastric content or volume as compared to NPO or gastric suctioning. This was indicated by Kruisselbrink et al. (2017). They studied the accuracy of ultrasound at determining gastric volume by calculating the gastric volume by ultrasound and comparing it to the volume DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 11 suctioned out. They determined that there was a strong correlation between ultrasound measured gastric volume and the volume of gastric content aspirated via a gastric tube. Van de Putte et al. (2017) also concluded that a larger antral CSA is consistent with higher qualitative grades and therefore an increased risk for aspiration. Their study further indicated that the use of ultrasound was capable of consistently discriminating between different gastric volumes at various time intervals following ingestion of fluids. Tankul et al. (2022) also identified that the diagnostic accuracy of qualitative gastric ultrasound assessment was as high as 96% when performed by trained anesthesiologists. For patents with comorbidities that affect gastric motility, Ultrasound continues to be a tool that can be used to assess or discriminate differences in gastric volume. Sabry et al. (2019) determined that patients with diabetes showed higher median antral CSA and aspirated gastric volume versus control (nondiabetics) and concluded that there was a good correlation between ultrasound calculated gastric volume and volume aspirated via a gastric tube. According to Bouvet et al. (2020), gastric suctioning did not provide a more accurate estimate of residual gastric volume as compared with ultrasound calculated volume. Joshi & Dhamija, (2021) used Gastric ultrasound to quantify gastric volume comparing patients who had fasted overnight to patients who ingested 200ml of clear apple juice 2 hours prior to their assessment. Gastric PH assessed in both groups were not significantly different. Gastric volume in the overnight fasting group was 29.7 8.0 ml. In the group that ingested 200 ml of clear fluid 2 hours prior to their assessment, the gastric volume was 19.2 4.9 ml. The statistically significant reduction in gastric volume after ingesting 200ml of fluid, strongly DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 12 supported fasting guidelines which help in reducing the preoperative discomfort of long fasting time and dehydration of patients without significantly impacting gastric PH. Patient safety In the last decade the guidelines for preoperative fasting has seen some changes with a push to liberalize preoperative fasting to enhance patient recovery, with ERUS protocols recommend, ingestion of carbohydrate containing drinks two hours prior to surgery. Shin et al. (2022) utilized gastric ultrasound to evaluate the safety of drinking carbohydrate containing fluids two hours prior to surgery in older adults. Their study determined that gastric volume was not significantly different between the fasting group (NPO) and the carbohydrate ingestion group (30.2 mL vs 28.4 ml). Mean difference was 1.9 mL (95% confidence interval, 17.9 to 14.2) and concluded that drinking carbohydrate containing fluid two hours prior to surgery is safe. Sanders et al. (2023), conducted a prospective observational study in healthy pediatric patients using gastric ultrasound to quantify the time taken to achieve a gastric volume < 1.5 mLkg1 (the upper limit of normal gastric volume in a fasted patient) after ingesting clear fluid. In this study, participants consumed 250 mL of a clear fluid followed by gastric US at four time intervals: 30, 60, 90, and 120 minutes to calculate gastric volume using the validated equation. They concluded that the total gastric fluid volume was < 1.5 mLkg1 after 60 min, suggesting that the fasting guidelines for the healthy pediatric population was safe and furthermore can be liberalized. Overall the review of the literature strongly supports ultrasound assessment as an accurate method of measuring gastric content in both the adult and pediatric population as well as healthy patients and patients with comorbidities the decrease gastric motility. The literature DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 13 shows that GUS is able to discriminate between changes in gastric volume over time as well as between patients who have fasted and those who have ingested fluids. The literature also shows that gastric ultrasound has been used to evaluate the safety of recent changes in preprocedure fasting guidelines and furthermore provides both a quantitative and qualitative noninvasive method of assessing gastric content and volume. As supported by the evidence in literature, the accuracy of gastric sonography eliminates any guess work in identifying patients with increased risk of aspiration that anesthesia providers may encounter by having a validated quantitative method of assessment, thus improving patient safety. The evidence behind the use of gastric ultrasound in anesthesia practice strongly suggests that acquiring the knowledge and the skills to perform the gastric ultrasound procedure would be highly beneficial to SRNAs. Conceptual framework The conceptual framework that will be used to serve as a guide in the development of this project will be the Knowledge-to-Action (KTA) model. This model was developed by the University of Ottawa as a way to merge the creation of knowledge and its application (White, 2016). The KTA uses a funnel to visually represent the movement of knowledge into higher stages until it is ready to be fully adopted (White, 2016). The KTA model is a planned action theory that is used to plan activities and facilitate change (White, 2016). This project will use the KTA model to compile and condense the existing knowledge on the use of ultrasonography, gastric anatomy and evidence based procedures in the education of SRNAs. The KTA model consists of seven phases that facilitate translation of knowledge to actionable practice (White, 2016). The first phase involves the identification of a problem that needs to be addressed and relevant research. The problem that was identified for this project is DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 14 that there is a knowledge deficit amongst SRNAs for US guided gastric content assessment. The second phase of the KTA model involves adapting existing knowledge for use in practice. The use of point of care ultrasound which is often used in regional anesthesia as well as in other medical specialties will be adapted for use in assessing gastric content and risk of aspiration. In step three of the model, developers address barriers to knowledge use. In this case, SRNAs have to learn to apply ultrasound for assessment of gastric content. In step four the assessment method and tools are tailored to simplify its application by SRNAs by providing an easy to follow checklist for performing the procedure as well as an instructional PowerPoint/video. Levels five through seven of the KTA model monitor use, evaluate outcomes of knowledge use, and sustained use of knowledge. Evaluation of use and outcomes will be addressed through a survey of the effect of the education on SRNA knowledge. By providing SRNAs with the knowledge and skills to perform an UG gastric assessment, student will incorporate this skill into their practice in the clinical setting as well as pass on their knowledge to others. The benefit to using the KTA model is that the seven phases are interconnected (White, 2016). Because of this, the knowledge creation and action cycles can continue to develop to meet the goals of the researchers and their target population. Goals and Objectives The purpose of this DNP project is to improve the quality of care of surgical patients during preprocedure assessment by educating SRNAs on how to perform an Ultrasound guided gastric content assessment. Project Aims: To develop a procedure checklist and teach SRNAs the skills needed to successfully perform an US guided gastric assessment. DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 15 The objectives of the DNP project include the following: 1. Develop a checklist for US guided gastric content assessment. 2. Develop a PowerPoint/video teaching material on how to perform a US gastric assessment 3. Develop pre and post teaching survey 4. Deploy teaching material and survey to SRNAs 5. Analyze survey results SWOT Analysis A SWOT analysis was conducted to identify strength, weakness, opportunity and threats to this DNP project. The analysis is as follows: Strengths: Ultrasound has been used in anesthesia and other medical specialties to provide reliable objective patient data. It is a tool that anesthesia providers including SRNAs are familiar with and use often. It is a skill that is also easy to learn and master by the novice practitioner Tankul et al. (2022). Teaching SRNAs the procedure for gastric assessment will improve identification of patients at risk of aspiration during pre-procedure assessment, enhancing patient safety under anesthesia care, decrease hospital length of stay and cost related to aspirational pneumonitis. Weaknesses: Learning a new skill can always be challenging especially for novice practitioners such as SRNAs. There is also less opportunity for students to practice and maintain the skill as it is not an institutional requirement which could make the skill redundant. Additionally, while it is DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 16 hoped that this new skill will be utilized in practice and potentially lead to a practice change where preprocedure US gastric assessment becomes a routine, it is not guaranteed. Opportunities: This project provides an opportunity to improve current practice and potentially encourage students to pursue additional research studies for use of gastric ultrasound assessment. Threats: Since this project is not conducted in conjunction with an institutional curriculum, there is less incentive to learn the skill, thus, students may pushback, student participation may be low or the project may be rejected altogether. Project Design and Methods The project is designed as an independent study education for student registered nurse anesthetists on US guided gastric content assessment. It involves the use of different instructional modalities to meet different learning needs of student including but not limited to PowerPoint presentation, video, audio, pictorial images, schematics and a procedural checklist. A pretest and posttest survey with multiple choice knowledge check and likert scale questions was used to assess participant knowledge and confidence for performing the US guided gastric content assessment before and after the education and to determine if there is a significant difference between students pretest and posttest score. Methods The project was deployed by the Marian University DNP nurse anesthesia department administrator to all registered SRNAs as an independent study 30 minute voice over instructional PowerPoint and pretest/posttest surveys. Participants were instructed to take the pretest prior to reviewing the PowerPoint and the posttest afterwards. The pretest and posttest questionnaire DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 17 surveyed 3 areas, 14 questions assessed students knowledge on anatomy and procedures involved in performing a GUS, 1 question assessed students confidence level for performing a GUS and 1 question assessed whether students have ever performed a GUS, this question was only asked in the pretest survey and not repeated in posttest for relevance. Students were also asked to provide the last 4 digits of their student ID number only for the purpose of linking pre and post test surveys (Appendix E). The survey results data collected was analyzed for statistical differences in the pretest/posttest scores to determine if there has been a change in students knowledge and confidence for performing the US guided gastric content assessment. Project Population and Site The project was conducted at Marian University, a tertiary Midwestern institution of higher education with over 100 SRNAs at different levels of their training. The project was deployed to all registered SRNAs in nurse anesthesia department of the institution by the departmental administrator to maintain anonymity. Participation by SRNAs was voluntary. Statistical Tests The study utilize a paired sample T-test to analyze the Pre and Post educational survey within the same cohort to determine a difference in the participants knowledge score as well as their confidence score for performing the US guided gastric content assessment. Ethical Considerations and Data Collection The DNP project is designed as an education for SRNAs including a pretest and posttest survey. The survey was conducted through Qualtrics, the Marian University recommended survey engine. No identifiable or demographic information was collected for this project. For the DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 18 purposes of linking participant pretest and posttest surveys, the last 4 digits of the participants student ID number were requested. These are not expected to pose any significant harm to participants requiring ethical consideration. To maintain credibility of survey results, an assessment of the appropriateness of collected data will be conducted. An audit trail will also be used to ensure dependability and confirmability of the survey results (Meadows-Oliver, 2019). Data points that contradict the majority will be analyzed to help eliminate any bias to make sure that the survey findings reflect the data collected and statistical analysis and not the researchers viewpoint (Meadows-Oliver, 2019). Results and Data Analysis After deployment of the DNP PowerPoint presentation and surveys, a period of 4 weeks was used to collect data during which reminders were sent to SRNAs for completion. A total of 17 responses were obtained. 2 of the respondents did not provide the last 4 digits of their students ID and was excluded from the final results. As is customary for SRNA exam scoring, the 14 knowledge based questions on the survey were scored as all or nothing, with no partial credit for multiple selection questions. For each respondent, their score on the knowledge based questions was reported as a percentage. The question on students confidence for performing GUS was coded and scored as follows: Not at all confident = 0; Somewhat not confident = 33; Somewhat confident = 66; Very confident = 100. The question on whether students have ever performed a GUS was a yes/no type question which was reported as a percentage of participants. Q: Have you ever performed an ultrasound assessment of gastric content All participants (100%) reported that they have never performed a GUS assessment DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 19 Q: How confident are you that you can perform an ultrasound assessment of gastric content In the pretest survey 86.7% of respondents (13) indicated they were not at all confident in performing a GUS assessment. 1 of respondents (6.7%), indicated they were somewhat not confident and another 6.7% indicated they were somewhat confident. In the posttest 1 participant (6.7%) indicated they were not at all confident (no change from pretest). 2 participants (13.3%) indicated they were somewhat not confident, both a change from not at all confident in the pretest survey. 12 participants (80%) indicated they were somewhat confident, of which 1(6.7%) had not changed from the pretest response and 1(6.7%) had changed from somewhat not confident. 10 of these responses (66.7%) were a change from not at all confident in the pretest survey. The paired T test showed that the participants perceived level of confidence in performing a GUS had significantly increased from the pretest (M = 6.6, SD = 18.5) to the post test survey (M = 57.2, SD = 19.6; t = -7.99, p < .001, d = -2.06). Q: Students knowledge based score The students pretest knowledge based scores ranged from a low of 21.43% to a high of 71.43% with a mean class score of 50.5%. Post test scores ranged for a low of 71.4% to a high of 100% and a mean class score of 93.8%. All 15 (100%) respondents scored below 83% (B grade) in the pretest. In the post test survey 2 participants (13.3%) scored below 83% while 13 participants (86.7%) scored above 83%. A paired T test showed that the participants knowledge based score had significantly increased from the pretest (M = 50.5, SD = 14.4) to the post test score (M = 93.8, SD = 9.3; t = 11.1, p < .001, d = -2.86). DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 20 Discussion Participants this DNP project were instructed to complete the pretest survey prior to reviewing the GUS PowerPoint presentation to provide a baseline of students knowledge before being exposed to the tutorial. The pretest results were then compared to the post test survey results to determine if there has been a change in students knowledge and confidence to perform the GUS assessment. In the pretest survey, participants knowledge based scores ranged from a low of 21.43% to a high of 71.43% with a mean class score of 50.5%. This is a low score profile considering that the passing grade for SRNA exams is 83% (B) or above. Participants were also asked in the pretest survey if they had ever performed a GUS assessment. All participants (100%) responded no to this question. This indicated that the procedure involved in GUS assessment was a fairly new concept and a reflection of the low pretest knowledge based scores. In the post test survey, the participants scores for the knowledge based questions significantly increased (t = -11.1, p < .001). In the post test, 2 participants (13.3%) scored below 83% while 13 participants (86.7%) representing the majority of participants scored above 83%. Student confidence for performing the GUS assessment also significantly increased from the pretest (M = 6.6, SD = 18.5) to the post test (M = 57.2, SD = 19.6; t = -7.99, p < .001, d = 2.06). In the pretest survey 86.7% of respondents (13) indicated they were not at all confident in performing a GUS assessment, whereas in the post test, 12 participants (80%) indicated they were somewhat confident in performing the GUS assessment after reviewing the PowerPoint. 1 (6.7%) participant who indicated they were not at all confident in the pretest had no change in their confidents in the post test after reviewing the tutorial. 1 (6.7%) participant who responded DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 21 in the pretest that they were somewhat not confident changed to somewhat confident in the post test survey and 1 (6.7%) participant who responded that they were somewhat confident in the pretest had no change in the post test. As indicated by these results, the information provided to SRNAs in the PowerPoint presentation significantly increased both their knowledge on the GUS procedure as well as their confidence to perform the procedure. While these results are an indication that this DNP project was successful at achieving its aims, the sample size of 15 respondents may be no the smaller size to provide a true indication and will have to be tested on a larger sample. This may require a different strategy to increase SRNA participation in the future. Conclusion During the SRNAs training program students learn the intricacies of providing anesthesia care to patients. It is a rigorous period of intense learning when students acquire knowledge on many concepts and hands on skills that are indispensable to anesthesia providers as well as shapes the students future practice. Patient safety is the paramount responsibility of the anesthesia provider and students must learn all and any skills that enhance their ability to maintain the patients safety. The ultrasound assessment of gastric content and volume is no exception. It provides a qualitative and quantitative means to assess gastric content and volume and improves the providers ability to identify patients at increased risk of aspiration which then allows the anesthesia provider to tailor their anesthetic to prevent aspiration, delay or postpone the case. As indicated by the results of this DNP project providing SRNAs with a tutorial on the procedure increases their knowledge and confidence for performing the GUS assessment. DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 22 Incorporating such a tutorial in the SRNAs study will better prepare students to perform the GUS assessment and thus increase utility in practice. DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 23 References Evain, J. N., Allain, T., Dilworth, K., Bertrand, B., Rabattu, P., Mortamet, G., Desgranges, F. P., Bouvet, L., & Payen, J. F. (2022). Ultrasound assessment of gastric contents in children before general anaesthesia for acute appendicitis. Anaesthesia, 77(6), 668673. https://doi.org/10.1111/anae.15707 Feighery, A. M., Oblizajek, N. R., Vogt, M. N. P., Bi, D., League III, J., Buttar, N. S., & Prichard, D. O. (2023). Retained Food During Esophagogastroduodenoscopy Is a Risk Factor for Gastric-to-Pulmonary Aspiration. Digestive Diseases & Sciences, 68(1), 164 172. https://doi.org/10.1007/s10620-022-07536-2 LaSala, V. R., Morgan, M. E., Bradburn, E. H., Vernon, T. M., Maish III, G. O. (2020). The effects of fasting status on the relative risk of pulmonary aspiration in acute care surgery patients. American Surgeon, 86(7), 837840. https://doi.org/10.1177/0003134820940257 Meadows-Oliver, M. (2019). Critically appraising qualitative evidence for clinical decision making. In B. M. Melnyk., & Fineout-Overholt, E. (Eds.), Evidenced-based practice in nursing and healthcare: A guide to best practice (4th ed., pp.189-218). Wolters Kluwer. Nason, K. S. (2015). Acute intraoperative pulmonary aspiration. Thoracic Surgery Clinics, 25(3), 301307. https://doi.org/10.1016/j.thorsurg.2015.04.01 Reed, A. M., & Haas, R. E. (2020). Type 2 diabetes mellitus: Relationships between preoperative physiologic stress, gastric content volume and quality, and risk of pulmonary aspiration. AANA Journal, 88(6), 465471 Tankul, R., Halilamien, P., Tangwiwat, S., Dejarkom, S., & Pangthipampai, P. (2022). Qualitative and quantitative gastric ultrasound assessment in highly skilled regional DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 24 anesthesiologists. BMC Anesthesiology, 22(1), 19. https://doi.org/10.1186/s12871-02101550-z White, K. M. (2016). The science of translation and major frameworks. In K. M. White, S. Dudley-Brown, & M. F. Terhaar (Eds.), Translation of evidence into nursing and health care (2nd ed., pp. 2555). Springer Publishing Company. DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 25 Appendix A DNP PROJECT GANTT CHART DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 26 Appendix B SWOT Chart Strengths Use of ultrasound is familiar to anesthesia providers including students. It is an easy skill to learn and master even for novice practitioners Potential to improve patient safety, hospital length of stay and cost if utilized. Weaknesses Challenge for students learning a new skill Possibility for skill to become redundant without it being an institutional requirement. Opportunities Improvement of current practices More research studies to strengthen need for use of gastric ultrasound assessment. Threats Pushback by students to learn a new skill that is not part of the institutional curriculum. DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 27 Appendix C Screening Identification PRISMA flow chart of literature search results Records identified from*: Databases CINAHL (n = 57) PUBMED (n = 18) Records screened (n = 57) Records excluded (n = 11) Reports sought for retrieval (n = 46) Reports not retrieved (n = 2) Reports assessed for eligibility (n = 44) Included Records removed before screening: Duplicate records removed (n = 18) Records removed for other reasons (n = 0) Reports excluded: Retracted (n = 1) Maternal/pregnancy (n = 9) Neonatal/premature (n = 3) Nonrelated/postop (n = 11) Non English language (n = 1) Studies included in review (n = 19) From: Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. doi: 10.1136/bmj.n71 DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 Running head: SRNA GASTRIC US EDUCATION 28 Appendix D Synthesis Matrix Citation Research Design & Level of Evidence Kruisselbrink, R., Arzola, C., Jackson, T., Okrainec, A., Chan, V., & Perlas, A. (2017). Ultrasound assessment of gastric volume in severely obese individuals: a validation study. BJA: The British Journal of Anaesthesia, 118(1), 7782. https://doi.org/10.1093/bja/ aew400 Randomized blinded experimental study Level 2 The oret ical / Con cept ual Fra me wor k N/A Purpose / Aim Popul ation / Sampl e size n=x Major Variables Instrument s / Data collection Results Evaluate performance of model in predicting gastric volume in severely obese subjects (BMI > 35) N = 38 BMI, Antral CSA of pre and post gastric volume after predetermi ned fluid ingestion (0 400 ml) Ultrasound ; Antral CSA, qualitative grading; NG suction volume Strong correlation between predicted sonographic gastric volume and suctioned volume (concordance correlation coefficient of 0.82 and Pearsons correlation coefficient of 0.86) in severely obese people DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION Tankul, R., Halilamien, P., Tangwiwat, S., Dejarkom, S., & Pangthipampai, P. (2022). Qualitative and quantitative gastric ultrasound assessment in highly skilled regional anesthesiologists. BMC Anesthesiology, 22(1), 19. https://doi.org/10.1186/s128 71-021-01550-z Sabry, R., Hasanin, A., Refaat, S., Abdel Raouf, S., Abdallah, A. S., & Helmy, N. (2019). Evaluation of gastric residual volume in fasting diabetic patients using gastric ultrasound. Acta Anaesthesiologica Scandinavica, 63(5), 615619. https://doi.org/10.1111/aas. 13315 Van de Putte, P., Vernieuwe, L., Jerjir, A., Verschueren, L., Tacken, M., & Perlas, A. (2017). When fasted is not empty: a retrospective cohort study of gastric content in fasted surgical patients. BJA: The British Journal of Anaesthesia, 118(3), 363 371. https://doi.org/10.1093/bja/ aew435 29 Prospective cohort study Level 4 N/A prospective observational study Level 4 retrospective cohort study Level 2 N/A Asses interrater agreement between anesthesiologist performing US gastric content measurement N = 47 Empty stomach, 100ml ,200ml, 300ml clear fluid and solid food Ultrasonog raphy, antral CSA Overall success rate of all gastric content categories was 96%. Tendency for deviation of results between raters increased with increasing gastric volume Evaluate residual gastric volume in fasting diabetics N = 50 Antral CSA, calculated gastric volume, aspirated gastric volume Ultrasonog raphy, aspiration of gastric volume via NG tube Diabetic group showed higher median antra CSA and aspirated gastric volume versus control. Good correlation between calculated gastric volume and aspirated content Evaluate the incidence of full stomach in a population of fasted patients presenting for elective surgery, using bedside gastric ultrasound. N= 538 Gastric volume, Antral CSA, Ultrasonog raphy, antral CSA, full or empty stomach, antral grade 6.2% of elective surgical patients present with a full stomach. Increasing antral grade was correlated with larger antral cross-sectional area and higher gastric volume (P<0.001). DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 30 Sander, T., Urmson, K., Langford, L., OBrien, J., Bajwa, J. S., Walker, M. E., & Leswick, D. (2023). Determining residual gastric volume in healthy children using ultrasound. Canadian Journal of Anaesthesia / Journal Canadien dAnesthsie, 70(8), 1323 1329. https://doi.org/10.1007/s126 30-023-02526-y prospective observational study Level 4 Shin, H. J., Koo, B. W., Lim, D., & Na, H.-S. (2022). Ultrasound assessment of gastric volume in older adults after drinking carbohydratecontaining fluids: a prospective, nonrandomized, and noninferiority comparative study. Canadian Journal of Anaesthesia / Journal Canadien dAnesthsie, 69(9), 1160 1166. https://doi.org/10.1007/s126 30-022-02262-9 Nonrandomiz ed and noninferiority comparative study Level 4 N/A N/A Quantify the time to achieve a gastric volume < 1.5 mLkg1 after clear fluid ingestion in healthy children N =33 Gastric volume, time elapsed Ultrasound guided antral CSA at 30, 60, 90 and 120 mins Evaluate the safety of drinking carbohydratecontaining fluids two hours prior to surgery in older adults using ultrasonography. N = 60 Gastric content and volume Ultrasound guided gastric antral CSA, Gastric volume Mean gastric volume per weight (mLkg1) at baseline was 0.51 mLkg1 (95% CI, 0.46 to 0.57). The mean gastric volume was 1.55 mLkg1 (95% CI, 1.36 to 1.75) at 30 min, 1.17 mLkg1 (95% CI, 1.01 to 1.33) at 60 min, 0.76 mLkg1 (95% CI, 0.67 to 0.85) at 90 min, and 0.58 mLkg1 (95% CI, 0.52 to 0.65) at 120 min. Total gastric volume was < 1.5 mLkg1 after 60 min Mean (standard deviation) gastric volume was not significantly different between the fasting group and the carbohydrate ingestion group (30.2 mL vs 28.4 ml). Mean difference was 1.9 mL (95% confidence interval , 17.9 to 14.2), and the upper limit of the 95% CI was lower than the prespecified non-inferiority limit ( = 50 mL) DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 31 Jae Yong Jeong, Jin Hee Ahn, Jae-Geum Shim, Sung Hyun Lee, Kyoung-Ho Ryu, Sung-Ho Lee, Eun-Ah Cho, Jeong, J. Y., Ahn, J. H., Shim, J.-G., Lee, S. H., Ryu, K.-H., Lee, S.-H., & Cho, E.-A. (2021). Gastric emptying of preoperative carbohydrate in elderly assessed using gastric ultrasonography: A randomized controlled study. Medicine, 100(37), 17. https://doi.org/10.1097/MD. 0000000000027242 Randomized controlled study Level 2 N/A Assess the safety of drinking carbohydratecontaining fluids two hours prior to surgery in older adults by comparing the residual GV between patients who fasted and patients who ingested carbohydratecontaining fluids two hours preoperatively. N = 58 Gastric content grades 0, 1, 2, antral CSA and aspirated gastric volume Ultrasound guided Antral CSA Abdul Kadir, M. Z., Cheah, S.K., Mohamad Yusof, A., Mohd Zaki, F., & Teo, R. (2022). Ultrasound-Determined Residual Gastric Volume after Clear-Fluid Ingestion in the Paediatric Population: Still a Debatable Issue. Children, 9(5), 639N.PAG. https://doi.org/10.3390/child ren9050639 Non randomized comparative study Level 3 N/A Evaluate the RGV after 1 and 2 h of clear fluid fasting. and parents satisfaction concerning clear fluid fasting time at 1 and 2 h. N = 99 Antral CSA. Residual gastric volume (RGV), Time (1hr & 2hr). parent satisfaction Ultrasound guided Antral CSA after 1 and 2 h of clear fluid. $ point satisfaction Likert scale. Incidence of grade 2 stomach was 13.8% in NPO group and 17.2% in carbohydrate group (P = .790). Antral CSA in the supine position was larger in carbohydrate group than in NPO group (4.42 [3.72 5.18] cm2 vs 5.31 [4.35 6.92] cm2, P = .018). Antral CSA in the RLD position was not different in NPO and carbohydrate groups (P = .120). There was no difference in gastric volume (2 [07.5] vs 3 [013.4], P = .331) in NPO group versus carbohydrate group. RGV was significantly higher at T1 compared to T2 (p < 0.001). No significant difference was seen between T0 and T2 (p = 0.30). Parental satisfaction was similar at T1 and T2 (p = 0.158). DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 32 Valero Castaer, H., Vendrell Jord, M., Sala Blanch, X., & Valero, R. (2021). Preoperative bedside ultrasound assessment of gastric volume and evaluation of predisposing factors for delayed gastric emptying: a casecontrol observational study. Journal of Clinical Monitoring & Computing, 35(3), 483489. https://doi.org/10.1007/s108 77-020-00489-9 Casecontrol observational study Level 4 N/A Assess differences in gastric fluid volume between fasted patients with or without predisposing factors for delayed gastric emptying. N = 53 Bouvet, L., Zieleskiewicz, L., Loubradou, E., Alain, A., Morel, J., Argaud, L., Chassard, D., Leone, M., & Allaouchiche, B. (2020). Reliability of gastric suctioning compared with ultrasound assessment of residual gastric volume: a prospective multicentre cohort study. Anaesthesia, 75(3), 323330. https://doi.org/10.1111/anae .14915 Cohort study Level 4 N/A Compare the reliability of aspiration via a nasogastric tube with ultrasound for assessment of residual gastric volume. N = 61 Gastric residual volume between patients with delayed gastric emptying (DGEF) versus patients without delayed emptying. Gastric residual volume, aspirated gastric volume Ultrasound Antral CSA , gastric fluid volume No differences were found between patients with or without delayed gastric emptying factors. Gastric fluid volume was 35.21 32.69 mL in the DGEF versus 53.50 30.72 mL in the non-DGEF group (p = 0.08). Average volume per unit of weight was 0.61 0.46 mL/kg. Ultrasound guided antral CSA, Gastric residual volume, aspirated gastric volume Gastric suctioning did not provide an accurate estimate of residual gastric volume compared with ultrasound, with a mean bias of 66.6 ml and a 95% agreement band ranging from 218 ml to 351 ml. DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION Miller, A. F., Levy, J. A., Krauss, B. S., Gravel, C. A., Vieira, R. L., Neuman, M. I., Monuteaux, M. C., & Rempell, R. G. (2021). Does Point-of-Care Gastric Ultrasound Correlate With Reported Fasting Time? Pediatric Emergency Care, 37(12), e1265e1269. https://doi.org/10.1097/PEC. 0000000000001997 Demirel, A., zgnay, . E., Eminolu, ., Balkaya, A. N., Onur, T., Klarslan, N., & Gaml, M. (2023). Ultrasonographic Evaluation of Gastric Content and Volume in Pediatric Patients Undergoing Elective Surgery: A Prospective Observational Study. Children, 10(9), 1432. https://doi.org/10.3390/child ren10091432 33 Non randomized cross sectional study, Level 4 N/A Prospective observational study Level 4 N/A Assess gastric volumes in pediatric ED patients, with the goal of determining the feasibility of this technique and the relationship between gastric volume and reported last oral intake. Evaluate the incidence of a high risk stomach characterized by ultrasound identification of solid matter and/or an estimated gastric fluid volume exceeding 1.25 mL/kg in elective procedures. N= 103 Antral CSA, gastric residual volume, Time since last intake Ultrasound guided antral CSA, Gastric residual volume, Time from last intake A weak inverse correlation between fasting time (either liquid or solid) and estimated gastric volume ( = 0.33) was observed, with no significant difference based on type of intake (solids, = 0.28; liquids, = 0.22). N=97 Gastric volume, Gastric content Ultrasound guided Antral CSA, gastric volume, Gastric content, BMI and age median fasting duration was 4 h for liquids and 9 h for thick liquids and solids. Solid content was absent in all the children. median antral CSA in the RLD was 2.36 cm2, with a median gastric volume of 0.46 mL/kg. A moderate and positive correlation was observed between the antral CSA and BMI for Grade 0 patients. A strong and positive correlation was evident between the antral CSA and age, DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION Valencia, J. A., Cubillos, J., Romero, D., Amaya, W., Moreno, J., Ferrer, L., Pabn, S., & Perlas, A. (2019). Chewing gum for 1 h does not change gastric volume in healthy fasting subjects. A prospective observational study. Journal of Clinical Anesthesia, 56, 100105. https://doi.org/10.1016/j.jcli nane.2019.01.021 Dupont, G., J. Gavory, P. Lambert, N. Tsekouras, N. Barbe, E. Presles, L. Bouvet, and S. Molliex. 2017. Ultrasonographic Gastric Volume before Unplanned Surgery. Anaesthesia 72 (9): 111216. doi:10.1111/anae.13963. 34 Observational prospective analytical study Level 4 N/A Investigate whether gum-chewing has significant impact on the gastric volume of healthy adults. N = 55 Prospective cohort study Level 4 N/A Ultrasound N= measurement of 300 gastric antral crosssectional area and estimate gastric volume in patients before unplanned surgery after at least a six-hour fast. Gastric volume, Gastric content Ultrasound guided Antral CSA, gastric volume, Gastric content Gastric antral CSA, Ultrasound guided Antral CSA, The proportion of subjects who presented a completely empty stomach (Grade 0 antrum) was similar at baseline and after 1 h of gum-chewing [81% vs. 84%, p = 0.19, CI 95% (12%, 16%)]. Among those subjects who had visible fluid at baseline, the volume remained unchanged The median (IQR [range]) area was 333 (241-472 [281803]) mm2, a mean (SD) estimated volume of 45.8 (34.0) ml. CSA exceeded 410 mm2 in 92/263 (35%) measurements. Body mass index and morphine administration were associated with larger gastric areas on multivariable linear regression analysis, with beta coefficient (95%CI) 0.02 (0.01-0.04), p = 0.01, 0.23 (0.01-0.46), p = 0.04, respectively. DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 35 Leviter, J., Steele, D. W., Prospective Constantine, E., Linakis, J. G., cohort study Amanullah, S., & Macy, M. L. Level 4 (2019). Full Stomach Despite the Wait: Pointof care Gastric Ultrasound at the Time of Procedural Sedation in the Pediatric Emergency Department. Academic Emergency Medicine, 26(7), 752760. https://doi.org/10.1111/ace m.13651 N/A to use gastric point ofcare ultrasound (POCUS) to assess gastric contents and volume, summarize the prevalence of "full stomach," and explore the relationship between fasting time and gastric contents at the time of procedural sedation N= 116 Gastric antral CSA, Gatric Volume and content, fasting time Ultrasound guided Gastric antral CSA, Gastric volume and qualitative gastric content, Fasting time Bouvet, L., Loubradou, E., Desgranges, F.-P., & Chassard, D. (2017). Effect of gum chewing on gastric volume and emptying: a prospective randomized crossover study. BJA: The British Journal of Anaesthesia, 119(5), 928 933. https://doi.org/10.1093/bja/ aex270 N/A To assess whether gum chewing affects gastric emptying of 250 ml water and residual gastric fluid volume measured 2 h after ingestion of water N = 20 Gastric antral CSA, gastric volume Ultrasound guided, timed CSA and gastric volume after chew gum or not. randomized observer-blind crossover trial Level 3 Median fasting time was 5.8 hours. 69% of evaluated scans (95% confidence interval [CI] = 60%77%), were categorized as having a full stomach (solid content or volume >1.2ml/kg). Each hour of fasting was associated with lower odds (odds ratio = 0.79, 95% CI = 0.650) of a full stomach. However, the knowledge of fasting time alone provides little ability to discriminate between risk groups No significant difference between chewing gum and control. Mean (sd) was 23 min in the Control and 21 min in the Chewing gum session (P=0.52). Total gastric emptying time of water was 42 min in the Control session and 39 min in the Chewing gum session (P=0.25). DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 36 Joshi, Y., & Dhamija, S. (2021). Randomized Control Clinical Trial of Overnight Fasting to Clear Fluid Feeding 2 Hours Prior Anaesthesia and Surgery. Indian Journal of Surgery, 83(1), 248254. https://doi.org/10.1007/s122 62-020-02369-7 Randomized control parallel group study. Level 3 N/A Arif, N. M., Nazihah Sayed Masri, S. N., Nur Yazmin Yaacob, Yeoh Chih Nie, Mahdi, S. N., & Izaham, A. (2021). Gastric Antrum Ultrasonography Measurement in Healthy Adults at 1 and 2-hours Fasting Time After Ingesting Glucose-loaded Clear Fluids...Malaysian Society of Anaesthesiologists & College of Anaesthesiologists, AMM, Annual Scientific Congress August 6-8, 2021. Non randomized cohort study Level 4 N/A Compare gastric volume using ultrasonography and pH of gastric aspirate by pH strip in patients after overnight fasting and after ingestion of 200 ml clear apple juice, 2 h prior to non-abdominal surgery as primary and secondary objective, respectively. compare gastric volume estimation in healthy fasting adults at different time interval after consuming lychee flavored beverage N = 60 Gastric volume and gastric PH Ultrasound guided CSA, Gastric volume and gastric PH Mean gastric volume was 29.7 8.0 ml in overnight fasting (grp A) and 19.2 4.9 ml in the 2h fluid group (grp B) which was statistically significant (p < 0.00001). Mean gastric PH was statistically insignificant p < 0.1268 (group A was 1.4 0.5 and group B was 1.6 0.5). N= 255 Gastric volume Ultrasound guided CSA, Gastric volume at baseline (after 8H fastin and 1 and 2 hours post ingestion of 250ml fluid (grp 1 & 2) Median of residual gastric volume per body weight after fasting for Group 1 was 1.3 (1.0 - 1.8) which was significantly higher than median of residual gastric volume in Group 2, with 1.1 (0.8 - 1.4) (p=0.001) DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION Okabe, T., Terashima, H., & Sakamoto, A. (2017). What is the manner of gastric emptying after ingestion of liquids with differences in the volume under uniform glucose-based energy content? Clinical Nutrition, 36(5), 12831287. https://doi.org/10.1016/j.cln u.2016.08.014 37 Non Randomized study Level 4 N/A Examine the effects of different volumes of liquids (200ml, 400ml, 600ml) with a uniform energy (200kcal) content on gastric emptying. N=8 Gastric CSA and volume Ultrasound guided CSA, Gastric volume, Time after fluid ingestion Mean gastric volume decreased exponentially to nearly 0 ml 70 min after ingestion of 200 ml, 90 min after 400 ml and 100 min after 600 ml . DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 Running head: SRNA GASTRIC US EDUCATION Appendix E Pretest Questionnaire (Correct responses are highlighted) 1. Please Provide the last 4 digits of your Marian University ID# 2. Pulmonary aspiration accounts for what percentage of anesthesia related death a. 1 5% b. 5 10% c. 10 30% d. 20 40% e. 30 50% 3. The stomach has how many layers a. 2 b. 3 c. 4 d. 5 e. 6 4. CRNA responsibilities for preventing aspiration include (Choose 2) a. Identification of aspiration risk b. Ensuring patient NPO compliance c. Reduction of aspiration risk d. Canceling the case 5. Advantages of gastric ultrasound include a. Safe b. Noninvasive c. Accurate d. All of the above 6. Anatomical parts of the stomach include (Chose 3) a. Infundibulum b. Pyloric antrum c. Pyloric fundus d. Body e. Fundus 7. True/False. The pyloric antrum is the most proximal part of the stomach a. True b. False 38 DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 8. The Antrum of the stomach is inferior to which organ on the ultrasound scan a. Aorta b. Pancreas c. Liver d. Sternum e. Colon 9. Ultrasound scanning should be done with the patient in what position (Choose 2) a. Supine b. Prone c. Right lateral decubitus d. Left lateral decubitus 10. Qualitative gastric assessment grade 0 corresponds with (Choose 2) a. Empty Antrum in supine position b. Empty antrum in RLD position c. Clear liquid in supine position d. Clear liquid in RLD position e. Thick fluid/solid in antrum 11. Qualitative gastric assessment grade 1 corresponds with (Choose 2) a. Empty Antrum in supine position b. Empty antrum in RLD position c. Clear liquid in supine position d. Clear liquid in RLD position e. Thick fluid/solid in antrum 12. Qualitative gastric assessment grade 2 corresponds with (chose ) a. Empty Antrum in supine position b. Empty antrum in RLD position c. Clear liquid in the antrum d. Thick fluid/solid in antrum 13. True/false. The quantitative gastric assessment is validated for only non-pregnant adult a. True b. False 14. True/False. A full stomach is a high risk for pulmonary aspiration a. True b. False 15. The upper limit of normal gastric volume in the fasted individual is a. 0.5 ml/kg b. 1.0 ml/kg c. 1.5 ml/kg d. 2.0 ml/kg e. 2.5 ml/kg 39 DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 40 16. How confident are you that you can perform an ultrasound assessment of gastric content a. Not at all confident b. Somewhat not confident c. Somewhat confident d. Very confident 17. Yes/No. Have you ever performed an ultrasound assessment of gastric content a. Yes b. No DocuSign Envelope ID: 82B3D178-9C76-4948-AA0B-C385E25DC5D2 SRNA GASTRIC US EDUCATION 41 Appendix F Result Tables Paired Samples T-Test pre vs post test knowledge score 95% Confidence Interval T value -11.1 df 14 p - value < .001 Mean difference -43.3 pretest knowldge score post test knowledge score T-value -7.99 df 14 p -value < .001 SE difference 3.91 Lower -51.7 Upper -34.9 95% Confidence Interval Effect Size -2.86 N Mean Median SD SE 15 15 50.5 93.8 50 100 14.4 9.3 3.72 2.4 Paired Samples T-Test pre vs post test students confidence score 95% Confidence Interval Lower Upper Mean difference SE difference Effect Size -50.6 6.33 -64.2 -37 -2.06 pretest confidence score post test confidence score N 15 15 Mean 6.6 57.2 Median 0 66 SD 18.5 19.6 SE 4.78 5.06 Lower -4.02 Upper -1.69 95% Confidence Interval Lower Upper -2.96 -1.14  ...

... The item referenced in this repository content can be found by following the link on the descriptive page.  ...

... The item referenced in this repository content can be found by following the link on the descriptive page.  ...

... Original Publication Building Interprofessional Competencies Through a Collaborative Prescribing Activity With Osteopathic, Pharmacy, and Physician Assistant Students Veronica Vernon, PharmD, Brian W. Skinner, PharmD*, , Patricia S. Devine, PharmD, Lori Fauquher, MS, PA-C, Emily Young, MD *Corresponding author: bskinner@marian.edu Co-primary authors Abstract Introduction: Medication errors can lead to signicant adverse events. Nearly 50% of medication errors occur during the prescription-writing stage of the medication use process, and effective interprofessional collaboration and communication are key to reducing error in this process. Methods: We developed a three-part, 60-minute, interprofessional education activity providing medical, physician assistant, and pharmacy students the opportunity to practice collegial interprofessional communication surrounding prescribing practices. Learners met virtually initially as a large group and divided into small groups facilitated by a health professional. Part 1 involved reviewing two prescriptions prepared by learners; part 2 was a discussion about the education, roles, and responsibilities of each profession; and part 3 focused on identifying prescription errors in examples provided by faculty. Students completed a post-pre survey measuring their perception of learning the Interprofessional Collaborative Competency Attainment Survey (ICCAS) areas. Results: Of 317 participants (151 doctor of osteopathy, 68 master of physician assistant studies, and 98 doctor of pharmacy students), 286 completed the post-pre survey, for a 90% response rate. Students reported statistically signicant (p < .001) increases in all 20 questions spanning the six ICCAS areas. Discussion: The virtual format allowed multiple institutions to participate from various locations. It broadened the learners experience by fostering interaction among those with varied perspectives and allowed collaboration between locations and programs that otherwise could not have participated. The activity introduced students to virtual collaboration and key telehealth skills, enhancing their condence and familiarity with virtual interactions in a professional setting. Keywords Prescription Writing, Case-Based Learning, Communication Skills, Interdisciplinary Medicine, Online/Distance Learning, Practice Management, Interprofessional Education Educational Objectives By the end of this activity, learners will be able to: 1. Respectfully exchange critiques of written prescriptions to identify errors. 2. Collaborate on prescription-writing best practices to improve communication, ensure clear and accurate transmission of medication information between health care professionals, and minimize the risk of errors. Citation: Vernon V, Skinner BW, Devine PS, Fauquher L, Young E. Building interprofessional competencies through a collaborative prescribing activity with osteopathic, pharmacy, and physician assistant students. MedEdPORTAL. 2024;20:11403. https://doi.org/10.15766/mep_2374-8265.11403 3. Analyze the unique challenges and responsibilities associated with prescription writing in the context of interprofessional collaborative teams. 4. Recognize the value of working across disciplines for patient-centered care. 5. Discuss the different training backgrounds, roles, and challenges experienced by pharmacy, physician assistant, and osteopathic professionals. Introduction Medication errors can lead to signicant adverse events, including death. The annual global cost of treating them is an estimated $42 billion.1 Nearly 50% of medication errors occur during the prescription-writing stage of the medication use process, such as selecting the wrong medication, route, dose, Copyright 2024 Vernon et al. This is an open-access publication distributed under the terms of the Creative Commons Attribution-NonCommercial license. 1/8 or frequency.2,3 The third World Health Organization (WHO) Global Patient Safety Challenge is medication without harm, which focuses on reducing medication-related harm by 50% over 5 years. The strategic framework for the Global Patient Safety Challenge was created to reduce medication errors, including those at the prescribing stage. Two domains of this framework related to health care professionals are education and training and collaboration and teamwork.1 Interprofessional education (IPE), as dened by the WHO, occurs when multiple professions learn about, from, and with each other in order to improve collaboration and health outcomes.4 A variety of rich IPE experiences are described in the literature. We reviewed MedEdPORTAL for publications that t our criteria for an IPE-related prescription-writing activity. A nonIPE prescription-writing activity was available,5 along with IPE activities encompassing other aspects of patient care,6,7 but none t our specic needs. We found two additional articles outside MedEdPORTAL that involved pharmacy and medical students in prescription-writing exercises. The rst article involved a workshop during which nal-year pharmacy students taught second-year medical students how to write prescriptions.8 The second article described collaboration between rstyear medical students and third-year pharmacy students on pharmacogenomics patient cases, and a portion of this activity provided an opportunity for pharmacy students to give feedback on written prescriptions.9 As a result, we sought to create an IPE activity for osteopathic medicine (DO), physician assistant (PA), and pharmacy students. All three disciplines have standards related to IPE.10-14 Pharmacy schools must include interprofessional learning opportunities for students and are assessed on this as part of their accreditation.13 Additionally, a common thread in the curricula of the three programs is prescribing and interpreting prescriptions. Enter and discuss orders and prescriptions is an entrustable professional activity (EPA) for entering residency for medical students.15 Therapeutic management and planning is a required graduation competency for PA students according to the PA Education Association,16 and fulll a medication order is an EPA for pharmacy students upon graduation.17 Thus, an IPE activity centered on writing prescriptions would be benecial to each profession. Our activity provides students with an experience that spans the four Interprofessional Education Collaborative (IPEC) competency domains: Interprofessional Teamwork and TeamBased Practice, Interprofessional Communication Practices, Roles and Responsibilities for Collaborative Practice, and Values/Ethics for Interprofessional Practice.18 Our activity allows all three disciplines to teach and learn interprofessional competencies through a written prescription activity. We developed a three-part, 60-minute, interdisciplinary activity that provides DO students and PA students the opportunity to receive from peers enrolled in pharmacy school real-time feedback regarding their ability to write a prescription that is accurate and clear and meets state legal requirements. Simulated practice of prescribing has been shown to increase condence and collaboration when other health care professions are represented in the activity.8,19 As the PA students are in the nal semester of their training, this also allows them to share what they have learned from their clinical rotations with medical and pharmacy students still in the didactic portion of their own curricula. Additionally, this activity provides an opportunity for different health profession students to discuss their respective education and training, their role in the health care team, and challenges unique to their future profession. We piloted the program in the spring of 2022 with the three programs and modied it in spring 2023 for use with 317 students (151 DO, 68 PA, and 98 pharmacy students). Prior to implementation, each institution taught information related to prescription writing relevant to their specic discipline, the timelines of which varied based on curricular design (Appendix A). Methods This activity consisted of three parts: Part 1 involved writing two prescriptions; part 2 was a group discussion on education, roles, and responsibilities of DOs, PAs, and pharmacists; and part 3 focused on error recognition in written prescriptions. The students were divided into three large groups, scheduled sequentially within the same afternoon, to accommodate the number of facilitators available. There was a 15-minute break to allow for turnover between each session. The large groups were then divided into 17-18 smaller cohorts, with trainees from each profession in each. In parts 1 and 2, each cohort was led by a pharmacist, PA, or physician who facilitated discussion and ensured coverage of the learning objectives. Faculty were recruited to facilitate based on their interest in IPE. To expose students to facilitators from all three professions, we returned to a large-group activity for part 3, after the smallgroup breakout sessions. This was more feasible than providing facilitators from each profession for all cohorts. We hosted the activity virtually using a videoconferencing system due to space constraints and geographic distance between the two universities involved. Students from each profession needed to have an entry-level understanding of the legal requirements of prescription writing, Copyright 2024 Vernon et al. This is an open-access publication distributed under the terms of the Creative Commons Attribution-NonCommercial license. 2/8 knowledge of commonly used abbreviations, and the ability to calculate patient-specic dosages when using a drug reference. DO students had received a self-paced prescription-writing module in a previous semester. PA students had coursework embedded in clinical medicine courses during their didactic year as well as training in prescription writing in their health care communications course. They also practiced application of prescription writing during clinical rotations throughout the second year of the curriculum. Pharmacy students learned legal requirements, formatting, and common issues regarding prescription writing in the rst professional year of their curriculum and applied this content throughout the curriculum in labs, courses, and rotations. Additionally, all learners had prior IPE collaboration events within their respective curricula. The pharmacy students had completed a longitudinal four-semester IPE course series that contained IPE content and experiences. PA and DO students had participated in multiple IPE experiences spanning the didactic and clinical years at their respective programs. We utilized third-year pharmacy and second-year DO students in their last semester of didactic education before entering clinical rotations, along with PA students currently completing clinical rotations in their last semester before graduation. Instructions to Learners Learners received an overview of the activity that included learning objectives along with the required preclass assignment (Appendix B) and their assigned groups (Appendix C). We reminded students about the event and provided a link to the video-conferencing service for their respective sessions. We instructed them to come in professional dress, to have their camera on and microphone muted but available, and to change their display name to include their name and future credentials (i.e., PharmD, DO, or PA). Students had a copy of their completed preclass assignment on their devices and were prepared to share their screen during the small-group sessions. Administrative Support One faculty member and staff from each program provided a roster of students and facilitators from each program; a staff member sorted the students into groups of ve to seven, ensuring each group had all three types of learners. The hosting institution was responsible for creating a videoconferencing link for each large-group session to be shared with all students and facilitators. One faculty member during each session created the breakout rooms, assigned participants to their respective rooms, took attendance, monitored the session in case a participant unintentionally was disconnected and needed assistance to rejoin, and called the participants back to the main room at the appropriate time (Table 1). Lead Faculty and Facilitators Each program assigned a lead faculty member to ensure all content was posted to their online learning management system in advance of the event and to answer student questions. As students entered the videoconference, instructions were displayed, reminding students of the expectations regarding camera and microphone usage, as well as expectations regarding their display name (Appendix D). The lead faculty provided a brief overview of the days schedule and each group members expectations (Appendix D). Facilitators received a copy of the facilitator guide (Appendix E) and reviewed all materials before the session. During part 1, facilitators asked student volunteers from either the PA or DO program to share their prescriptions with the group and solicit feedback. Pharmacy students provided feedback rst on each shared prescription, followed by the other students. If an error or discrepancy was noted on the prescription, students were encouraged to help identify the proper changes as well as to explain how to calculate a patient-specic dose using reference material. Multiple students shared their prescriptions as time allowed. Facilitators were responsible for ensuring each student had an opportunity to receive direct feedback and for maintaining the time between activities. During part 2 of the exercise, facilitators encouraged students from each profession to briey describe their education and training, their role in the health care team, and the challenges they expected to face in their professions. The questions were meant to be a starting point for discussion, not an exclusive list. The focus of part 2 was on students sharing their experiences and expectations as they entered their profession. Once all participants had returned to the main room, the lead faculty began part 3 by providing a series of prescriptions with several errors and asking students to identify what concerns Table 1. Facilitation Timeline Activity Welcome and introduction Part 1: prescription writing Part 2: roles of the profession Part 3: prescription errors Debrief Survey distribution Suggested Time Location 5 minutes 15 minutes 15 minutes Main room Breakout rooms Breakout rooms 15 minutes 9 minutes 1 minute Main room Main room Main room Copyright 2024 Vernon et al. This is an open-access publication distributed under the terms of the Creative Commons Attribution-NonCommercial license. 3/8 or issues they saw (Appendix D, cases 3-5). The lead faculty encouraged students from each profession to participate in the discussion. Time was reserved at the end of the activity for a debrief about what each group of students had learned about the other professions with which they had interacted. We shared the link to the post-pre survey at the end of the activity via the chat function and email after the event and encouraged students to complete the survey before exiting the videoconferencing meeting (Appendix F). Survey Development A post-pre survey assesses students perception of change after an educational activity or intervention. A perception of change survey can assess knowledge and skills, personal attributes, or their impact on future behavior. Although similar to a pre-post survey, a post-pre survey design has several advantages both in administration and in elimination of knowledge bias. Unlike a pre-post survey that has to be administered both before and after an educational event, a post-pre survey is administered once at the conclusion of the event. With a pre-post survey design, students may unintentionally rate their level of knowledge higher due to a lack of awareness of opportunity for growth in an area. A post-pre assessment allows students to utilize their current level of knowledge as a benchmark for their initial level of knowledge prior to the intervention.20 Our post-pre survey was developed based on the validated Interprofessional Collaborative Competency Attainment Survey (ICCAS) and had been utilized for other IPE activities.21,22 Across 20 items using a 7-point IPEC Competency Domains Likert scale (1 = strongly disagree, 7 = strongly agree), the ICCAS measures six competencies similar to the competencies developed by the IPEC (Figure). Additionally, four open-ended questions were included to provide formative feedback for improvement (Appendix F). The mean and standard deviation were calculated and reported for each Likert-scale question. Inferential statistics were determined with IBMs SPSS (version 29). The Wilcoxon signed rank test for nonparametric data was used to evaluate the paired responses to the post-pre survey Likert-scale responses, with an alpha set at .05. This project was determined to be exempt by institutional review boards at both institutions. Additionally, learners had an opportunity to provide open-ended feedback regarding the event to drive future improvements. Results Together, faculty from three professions at two universities developed an IPE prescription-writing virtual event that was piloted in January 2022 and implemented as a required component in each professions curriculum in February 2023. In total, 317 students participated in the 2023 activity: 151 DO second-year students, 68 PA masters second-years, and 98 doctor of pharmacy third-year students. Of the 317 participants, 286 completed the post-pre survey measuring their perception of learning the ICCAS areas, for an overall response rate of 90%. Students reported statistically signicant increases (p < .001) in all 20 items across six ICCAS areas (team functioning, ICCAS Competency Areas Team Funconing Interprofessional Teamwork and Team-Based Pracce Collaboraon Interprofessional Communicaon Pracces Communicaon Roles and Responsibilies Roles and Responsibilies for Collaborave Pracce Collaborave Paent/Family-Centered Approach Values/Ethics for Interprofessional Pracce Conict Management/Resoluon Figure. Comparison of IPEC competency domains and ICCAS competency areas. Abbreviations: IPEC, Interprofessional Education Collaborative; ICCAS, Interprofessional Collaborative Competency Attainment Survey. Copyright 2024 Vernon et al. This is an open-access publication distributed under the terms of the Creative Commons Attribution-NonCommercial license. 4/8 communication, collaboration, roles and responsibilities, collaborative patient/family-centered approach, and conict management; Table 2). While not designed as a formal qualitative analysis, responses to the open-ended questions were reviewed. Responses to What do you see as the value of interprofessional collaboration? included comments about improving patient outcomes and patient safety, enhancing ones own comfort level when working with other professions, and learning about others perspectives in prescribing medications. Common responses to the question What did you learn about, with and/or from the other professions in your group? were how to work as a team, others perspectives, and different roles in health care. When asked how this event could be improved, many commented positively on the event structure, content, and/or facilitators, including Favorite IPE event so far! Suggestions for improvement were (1) centering the activity around a patient case that included writing prescriptions for the patient, (2) providing more time for discussion, and (3) offering more IPE events similar to this one. Seventeen faculty from all three professions participated as smallgroup facilitators. Faculty were not formally surveyed; however, during the postevent informal quality assurance discussion, most facilitators felt their groups had been very engaged, asked insightful questions of each other, and appreciated each others contributions. Discussion With limitations on in-person activities predominating in the educational environment during the COVID-19 pandemic, planning interactive IPE events became very challenging. The implementation of this activity in the setting of these restrictions allowed students to interact with the benet of virtual learning and continue to work toward their program goals in IPE. Having transitioned to the virtual platform initially out of necessity, we noted that this mode of learning could offer additional advantages to an event of this kind. Constraints such as limited health profession programs in the same geographic area, space limitations, and insufficient numbers of faculty facilitators often create barriers to in-person events of this size. Our activity allows for either an in-person event or an easy transition to a virtual event that can be used across many disciplines and across state lines, opening up more opportunities for learners in various health care specialties to participate. The virtual format also requires fewer faculty facilitators to effectively manage a large Table 2. Students Self-Assessment Responses to the Prompt Please Rate Your Ability for Each of the Following Statements Prior to and After Todays Event (N = 286) Statementa Communication Promote effective communication among members Actively listen to interprofessional professional team members ideas and concerns Express my ideas and concerns without being judgmental Provide constructive feedback to interprofessional team members Express my ideas and concerns in a clear, concise manner Collaboration Seek out interprofessional team members to address issues Work effectively with interprofessional team members to enhance care Learn with, from, and about interprofessional team members to enhance care Roles and responsibilities Identify and describe my abilities and contributions to the interprofessional team Be accountable for my contributions to the interprofessional team Understand the abilities and contributions of interprofessional team members Recognize how others skills and knowledge complement and overlap with my own Collaborative patient/family-centered approach Use an interprofessional team approach with the patient to assess the health situation Use an interprofessional team approach with the patient to provide whole person care Include the patient/family in decision-making Conict management/resolution Actively listen to the perspectives of interprofessional team members Take into account the ideas of interprofessional team members Address team conict in a respectful manner Team functioning Develop an effective care plan with interprofessional team members Negotiate responsibilities within overlapping scopes of practice Preperception M (SD) Postperception M (SD) p (0.9) (0.8) (0.9) (1.0) (0.9) <.001 <.001 <.001 <.001 <.001 5.5 (1.2) 5.8 (1.1) 5.7 (1.2) 6.1 (1.1) 6.3 (0.9) 6.3 (0.9) <.001 <.001 <.001 5.6 5.8 5.6 5.7 6.1 6.2 6.3 6.3 (1.2) (1.0) (1.0) (0.9) <.001 <.001 <.001 <.001 5.5 (1.3) 5.6 (1.3) 5.4 (1.6) 6.1 (1.2) 6.1 (1.3) 5.8 (1.6) <.001 <.001 <.001 5.9 (1.2) 6.0 (1.1) 5.9 (1.4) 6.3 (1.0) 6.4 (0.9) 6.2 (1.4) <.001 <.001 <.001 5.7 (1.3) 5.6 (1.4) 6.2 (1.2) 6.0 (1.2) <.001 <.001 5.3 6.0 5.9 5.5 5.7 (1.2) (1.1) (1.2) (1.3) (1.2) (1.3) (1.2) (1.2) (1.2) 6.1 6.4 6.3 6.1 6.1 a From the Interprofessional Collaborative Competency Attainment Survey; rated on a 7-point Likert scale (1 = strongly disagree, 2 = moderately disagree, 3 = slightly disagree, 4 = neutral, 5 = slightly agree, 6 = moderately agree, 7 = strongly agree). Copyright 2024 Vernon et al. This is an open-access publication distributed under the terms of the Creative Commons Attribution-NonCommercial license. 5/8 group of students in breakout rooms than would be necessary for a live event. Our activity had results similar to other IPE events described in the literature. An IPE activity involving PAs and pharmacy students found an increase in the perception of team member value, efficiency, and accommodation.23 Durham, Lie, and Lohenry noted an increase in the understanding of the respective roles of team members with medical, pharmacy, and PA students in their IPE activity.6 The students in our activity demonstrated growth in all areas as measured by the ICCAS. Many aspects of the activity met the intended objectives. The virtual format allowed multiple institutions to participate, broadening the learners experience by fostering interaction with others with varied perspectives from other locations and programs. This format supported many student participants when space limitations prevented physically convening a group the same size in one location. The activity also introduced students to virtual collaboration and skills essential to the ever-growing eld of telehealth, enhancing their condence and familiarity with virtual interactions in a professional setting. Moreover, the threepart format provided the exibility to rearrange the activity or use only part of it, if desired. In addition, participation from multiple learners with prescribing ability generated more diverse discussions and perspectives on prescription writing. Assigning prework to learners before the interactive event proved to be effective in assuring they were well prepared to contribute to more in-depth and wellrounded discussions during the event. This preparation step also produced more time for discussion during the event. The postpre survey indicated improvement across all IPE competency areas and an increased appreciation for the various participating professions. We noted several limitations that could be improved on for future cohorts and further enrich the learning potential of the activity. While virtual learning platforms work well to allow multiple learners from various locations to participate, survey results indicated that some learners felt they would gain more from the in-person interaction of live events. Virtual forums tend to limit bidirectional communication, especially from the breakout groups to the main facilitator. They also necessitate more stringent time constraints that may end breakout room discussions abruptly. Many groups ran out of discussion time in their breakout sessions, a limitation that might improve in a live setting where the event could be longer with more time dedicated to small-group discussions. The virtual design also requires additional facilitator preparation, such as planning breakout session groups, assigning students to those groups during or before the event, and managing the technology of the virtual platform. While a smaller number of facilitators can manage breakout rooms virtually, they are required to move from group to group, jumping in and out of breakout rooms, which can interrupt learners discussions. Finally, while this activity challenges learners to assess and practice their prescriptionwriting knowledge and skills, it does not simulate electronic prescription writing, arguably a more applicable method used in clinical practice. The post-pre survey assessment method is reliable for showing student perceptions of their knowledge level. The response rate was very high, with 90% of students completing the survey following the activity. Requiring students to complete the survey before leaving the event contributes to higher response rates and stronger data. Although the survey provided reliable outcome data on learners perspectives, it did not adequately reect their mastery of content related to prescription writing. Future surveys should include a question assessing how the activity affected students abilities and comfort with prescription writing. Other opportunities for future research include development and execution of a qualitative analysis of learner perspectives on how the activity helps them learn about prescription writing or interprofessional competencies and/or capturing the viewpoint of facilitators. After repeating this learning activity a second year and receiving positive outcomes, all three of our programs plan to continue offering the session moving forward. This learning activity provides an opportunity to collaborate with multiple health professions across different institutions. It can be used by any pharmacy program in coordination with PA students and osteopathic medical students. Additionally, it can be adapted for other health professions with prescriptive authority, including allopathic medical students and nurse practitioner students. Students can be enrolled at the same or different institutions without restriction on geographical locale when using a videoconferencing system. Survey results demonstrated a statistically signicant increase in all 20 questions spanning six interprofessional collaboration competency areas, and students reported generally positive experiences regarding multiple aspects of the activity, including its implementation, the value of collaboration, and the roles of each health profession. If implementing the activity, institutions should ensure adequate educational technology support is available to facilitate its virtual delivery. Copyright 2024 Vernon et al. This is an open-access publication distributed under the terms of the Creative Commons Attribution-NonCommercial license. 6/8 Appendices A. Timeline for Implementation.pptx B. Prework Instructions.docx C. Example Groups.xlsx D. Introductory Slides.pptx E. Facilitator Guide.docx F. Post-Pre Survey.docx All appendices are peer reviewed as integral parts of the Original Publication. Veronica Vernon, PharmD: Assistant Professor, Department of Pharmacy Practice, Butler University College of Pharmacy and Health Sciences; ORCID: https://orcid.org/0000-0002-5638-3851 Brian W. Skinner, PharmD: Associate Professor of Internal Medicine, Department of Clinical Sciences, Marian University College of Osteopathic Medicine; ORCID: https://orcid.org/0000-0003-3814-7887 Patricia S. Devine, PharmD: Professor, Department of Pharmacy Practice, and Director of Interprofessional Education, Butler University College of Pharmacy and Health Sciences Lori Fauquher, MS, PA-C: Assistant Professor, Physician Assistant Program, Butler University College of Pharmacy and Health Sciences Emily Young, MD: Associate Professor of Pediatrics, Department of Clinical Sciences, Marian University College of Osteopathic Medicine Disclosures None to report. Funding/Support None to report. Ethical Approval The Butler University Institutional Review Board and the Marian University Institutional Review Board deemed further review of this project not necessary. References 1. Medication without harm. World Health Organization. Accessed April 2, 2024. https://www.who.int/initiatives/medicationwithout-harm 2. Tariq RA, Vashisht R, Sinha A, Scherbak Y. Medication dispensing errors and prevention. StatPearls. Updated February 12, 2024. Accessed April 2, 2024. https://www.statpearls.com/point-of-care/24883 3. Medication Errors: Technical Series on Safer Primary Care. World Health Organization; 2016. Accessed April 2, 2024. https://iris.who.int/bitstream/handle/10665/252274/ 9789241511643-eng.pdf?sequence=1 4. Framework for Action on Interprofessional Education & Collaborative Practice. World Health Organization; 2010. Accessed April 2, 2024. https://iris.who.int/bitstream/handle/ 10665/70185/WHO_HRH_HPN_10.3_eng.pdf?sequence=1 5. Ekong M, Frazier J, Oholendt K. How to write prescriptions. MedEdPORTAL. 2014;10:9982. https://doi.org/10.15766/mep_2374-8265.9982 6. Durham M, Lie D, Lohenry K. Interprofessional care: an introductory session on the roles of health professionals. MedEdPORTAL. 2014;10:9813. https://doi.org/10.15766/mep_2374-8265.9813 7. Gill AC, Cowart JB, Hateld CL, et al. Patient safety interprofessional training for medical, nursing, and pharmacy students. MedEdPORTAL. 2017;13:10595. https://doi.org/10.15766/mep_2374-8265.10595 8. Allen SM, Kachlic MD, Parent-Stevens L. Pharmacy students teaching prescription writing and nonprescription product selection to medical students. Am J Pharm Educ. 2020;84(3):6972. https://doi.org/10.5688/ajpe6972 9. Calinski DM, Hoefer C, Kisor D. An interprofessional education experience to promote the role of the pharmacist in precision medicine. Curr Pharm Teach Learn. 2021;13(10):1370-1375. https://doi.org/10.1016/j.cptl.2021.07.017 10. ACCM Standards of Accreditation for Schools of Medicine. Accreditation Commission on Colleges of Medicine; 2022. Updated June 2023. Accessed April 2, 2024. https://accredmed. org/wp-content/uploads/2023/05/STANDARDS-OFACCREDITATION-FOR-SCHOOLS-OF-MEDICINE-May-2023.pdf 11. American Osteopathic Association Commission on Osteopathic College Accreditation. Accreditation of Colleges of Osteopathic Medicine: COM New & Developing Accreditation Standards. American Osteopathic Association; 2023. Accessed April 2, 2024. https://osteopathic.org/index.php?aam-media=/wp-content/ uploads/2023-COM-New-and-Developing-AccreditationStandards.pdf 12. Accreditation Standards for Physician Assistant Education. 5th ed. Accreditation Review Commission on Education for the Physician Assistant; 2019. Updated September 2023. Accessed April 2, 2024. https://www.arc-pa.org/wp-content/uploads/2023/ 10/Standards-5th-Ed-September-2023.pdf 13. Accreditation Standards and Key Elements for the Professional Program in Pharmacy Leading to the Doctor of Pharmacy Degree (Standards 2016). Accreditation Council for Pharmacy Education; 2015. Accessed April 2, 2024. https://www.acpe-accredit.org/pdf/Standards2016FINAL.pdf 14. Osteopathic Considerations for Core Entrusable Professional Activities (EPAs) for Entering Residency. American Association of Colleges of Osteopathic Medicine; 2016. Accessed April 2, Copyright 2024 Vernon et al. This is an open-access publication distributed under the terms of the Creative Commons Attribution-NonCommercial license. 7/8 2024. https://www.aacom.org/docs/default-source/med-eddocuments/core-epas.pdf?sfvrsn=b6145397_4 interprofessional simulation-based learning activity. Adv Simul (Lond). 2017;2:14. https://doi.org/10.1186/s41077-017-0047-0 15. The Core Entrusable Professional Activities (EPAs) for Entering Residency. Association of American Medical Colleges. Accessed April 2, 2024. https://www.aamc.org/about-us/mission-areas/ medical-education/cbme/core-epas 20. Post-pre survey resources. Simon Fraser University Institute for the Study of Teaching and Learning in the Disciplines. Accessed April 2, 2024. https://www.sfu.ca/istld/faculty/resources/postpre.html 16. Core Competencies for New Physician Assistant Graduates. PA Education Association; 2019. Accessed April 2, 2024. https://paeaonline.org/wp-content/uploads/2023/06/corecompetencies-for-new-pa-grads-097119.pdf 21. Archibald D, Trumpower D, MacDonald CJ. Validation of the Interprofessional Collaborative Competency Attainment Survey (ICCAS). J Interprof Care. 2014;28(6):553-558. https://doi.org/10.3109/13561820.2014.917407 17. Medina MS, Farland MZ, Conry JM, et al. Finalizing the work related to the Curriculum Outcomes and Example Objectives and Entrustable Professional Activities (COEPA) document: the report of the 20222023 Academic Affairs Standing Committee. Am J Pharm Educ. 2023;87(8):100560. https://doi.org/10.1016/j.ajpe.2023.100560 22. Schmitz CC, Radosevich DM, Jardine P, MacDonald CJ, Trumpower D, Archibald D. The Interprofessional Collaborative Competency Attainment Survey (ICCAS): a replication validation study. J Interprof Care. 2017;31(1):28-34. https://doi.org/10.1080/13561820.2016.1233096 18. Interprofessional Education Collaborative. Core Competencies for Interprofessional Collaborative Practice: 2016 Update. Interprofessional Education Collaborative; 2016. Accessed April 2, 2024. https://www.ipecollaborative.org/assets/2016-Update.pdf 19. Cooke C, Gormley GJ, Haughey S, Barry J. Tracing the prescription journey: a qualitative evaluation of an 23. Won KJ, Tsu LV, Saldivar S, Beuttler R, Walsh A. The effect of interprofessional simulations on pharmacy and physician assistant students learning of advanced cardiac life support concepts. Curr Pharm Teach Learn. 2023;15(5):521-527. https://doi.org/10.1016/j.cptl.2023.05.002 Received: November 1, 2023 Accepted: February 16, 2024 Published: May 17, 2024 Copyright 2024 Vernon et al. This is an open-access publication distributed under the terms of the Creative Commons Attribution-NonCommercial license. 8/8  ...

... The item referenced in this repository content can be found by following the link on the descriptive page.  ...