Scene-dependent, feedforward eye gaze metrics can differentiate technical skill levels of trainees in laparoscopic surgery.

INTRODUCTION: In laparoscopic surgery, looking in the target areas is an indicator of proficiency. However, gaze behaviors revealing feedforward control (i.e., looking ahead) and their importance have been under-investigated in surgery. This study aims to establish the sensitivity and relative importance of different scene-dependent gaze and motion metrics for estimating trainee proficiency levels in surgical skills. METHODS: Medical students performed the Fundamentals of Laparoscopic Surgery peg transfer task while recording their gaze on the monitor and tool activities inside the trainer box. Using computer vision and fixation algorithms, five scene-dependent gaze metrics and one tool speed metric were computed for 499 practice trials. Cluster analysis on the six metrics was used to group the trials into different clusters/proficiency levels, and ANOVAs were conducted to test differences between proficiency levels. A Random Forest model was trained to study metric importance at predicting proficiency levels. RESULTS: Three clusters were identified, corresponding to three proficiency levels. The correspondence between the clusters and proficiency levels was confirmed by differences between completion times (F2,488 = 38.94, p < .001). Further, ANOVAs revealed significant differences between the three levels for all six metrics. The Random Forest model predicted proficiency level with 99% out-of-bag accuracy and revealed that scene-dependent gaze metrics reflecting feedforward behaviors were more important for prediction than the ones reflecting feedback behaviors. CONCLUSION: Scene-dependent gaze metrics revealed skill levels of trainees more precisely than between experts and novices as suggested in the literature. Further, feedforward gaze metrics appeared to be more important than feedback ones at predicting proficiency.

NAPS Fusion: A framework to overcome experimental data limitations to predict human performance and cognitive task outcomes.

In the area of human performance and cognitive research, machine learning (ML) problems become increasingly complex due to limitations in the experimental design, resulting in the development of poor predictive models. More specifically, experimental study designs produce very few data instances, have large class imbalances and conflicting ground truth labels, and generate wide data sets due to the diverse amount of sensors. From an ML perspective these problems are further exacerbated in anomaly detection cases where class imbalances occur and there are almost always more features than samples. Typically, dimensionality reduction methods (e.g., PCA, autoencoders) are utilized to handle these issues from wide data sets. However, these dimensionality reduction methods do not always map to a lower dimensional space appropriately, and they capture noise or irrelevant information. In addition, when new sensor modalities are incorporated, the entire ML paradigm has to be remodeled because of new dependencies introduced by the new information. Remodeling these ML paradigms is time-consuming and costly due to lack of modularity in the paradigm design, which is not ideal. Furthermore, human performance research experiments, at times, creates ambiguous class labels because the ground truth data cannot be agreed upon by subject-matter experts annotations, making ML paradigm nearly impossible to model. This work pulls insights from Dempster-Shafer theory (DST), stacking of ML models, and bagging to address uncertainty and ignorance for multi-classification ML problems caused by ambiguous ground truth, low samples, subject-to-subject variability, class imbalances, and wide data sets. Based on these insights, we propose a probabilistic model fusion approach, Naive Adaptive Probabilistic Sensor (NAPS), which combines ML paradigms built around bagging algorithms to overcome these experimental data concerns while maintaining a modular design for future sensor (new feature integration) and conflicting ground truth data. We demonstrate significant overall performance improvements using NAPS (an accuracy of 95.29%) in detecting human task errors (a four class problem) caused by impaired cognitive states and a negligible drop in performance with the case of ambiguous ground truth labels (an accuracy of 93.93%), when compared to other methodologies (an accuracy of 64.91%). This work potentially sets the foundation for other human-centric modeling systems that rely on human state prediction modeling.

Building an emotion regulation recommender algorithm for socially anxious individuals using contextual bandits.

OBJECTIVES: Poor emotion regulation (ER) has been implicated in many mental illnesses, including social anxiety disorder. To work towards a scalable, low-cost intervention for improving ER, we developed a novel contextual recommender algorithm for ER strategies. DESIGN: N = 114 socially anxious participants were prompted via a mobile app up to six times daily for five weeks to report their emotional state, use of 19 different ER strategies (or no strategy), physical location, and social context. Information from passive sensors was also collected. METHODS: Given the large number of ER strategies, we used two different approaches for variable reduction: (1) grouping ER strategies into categories based on a prior meta-analysis, and (2) considering only the ten most frequently used strategies. For each approach, an algorithm that recommends strategies based on one's current context was compared with an algorithm that recommends ER strategies randomly, an algorithm that always recommends cognitive reappraisal, and the person's observed ER strategy use. Contextual bandits were used to predict the effectiveness of the strategies recommended by each policy. RESULTS: When strategies were grouped into categories, the contextual algorithm was not the best performing policy. However, when the top ten strategies were considered individually, the contextual algorithm outperformed all other policies. CONCLUSIONS: Grouping strategies into categories may obscure differences in their contextual effectiveness. Further, using strategies tailored to context is more effective than using cognitive reappraisal indiscriminately across all contexts. Future directions include deploying the contextual recommender algorithm as part of a just-in-time intervention to assess real-world efficacy. PRACTITIONER POINTS: Emotion regulation strategies vary in their effectiveness across different contexts. An algorithm that recommends emotion regulation strategies based on a person's current context may one day be used as an adjunct to treatment to help dysregulated individuals optimize their in-the-moment emotion regulation. Recommending flexible use of emotion regulation strategies across different contexts may be more effective than recommending cognitive reappraisal indiscriminately across all contexts.

Evaluation of a Virtual Training to Enhance Public Health Capacity for COVID-19 Infection Prevention and Control in Nursing Homes.

CONTEXT: Between April 2020 and May 2021, the Centers for Disease Control and Prevention (CDC) awarded more than $40 billion to health departments nationwide for COVID-19 prevention and response activities. One of the identified priorities for this investment was improving infection prevention and control (IPC) in nursing homes. PROGRAM: CDC developed a virtual course to train new and less experienced public health staff in core healthcare IPC principles and in the application of CDC COVID-19 healthcare IPC guidance for nursing homes. IMPLEMENTATION: From October 2020 to August 2021, the CDC led training sessions for 12 cohorts of public health staff using pretraining reading materials, case-based scenarios, didactic presentations, peer-learning opportunities, and subject matter expert-led discussions. Multiple electronic assessments were distributed to learners over time to measure changes in self-reported knowledge and confidence and to collect feedback on the course. Participating public health programs were also assessed to measure overall course impact. EVALUATION: Among 182 enrolled learners, 94% completed the training. Most learners were infection preventionists (42%) or epidemiologists (38%), had less than 1 year of experience in their health department role (75%), and had less than 1 year of subject matter experience (54%). After training, learners reported increased knowledge and confidence in applying the CDC COVID-19 healthcare IPC guidance for nursing homes (≥81%) with the greatest increase in performing COVID-19 IPC consultations and assessments (87%). The majority of participating programs agreed that the course provided an overall benefit (88%) and reduced training burden (72%). DISCUSSION: The CDC's virtual course was effective in increasing public health capacity for COVID-19 healthcare IPC in nursing homes and provides a possible model to increase IPC capacity for other infectious diseases and other healthcare settings. Future virtual healthcare IPC courses could be enhanced by tailoring materials to health department needs, reinforcing training through applied learning experiences, and supporting mechanisms to retain trained staff.

Relationships between trait emotion dysregulation and emotional experiences in daily life: an experience sampling study.

Few studies have examined how trait emotion dysregulation relates to momentary affective experiences and the emotion regulation (ER) strategies people use in daily life. In the current study, 112 college students completed a trait measure of emotion dysregulation and completed experience sampling and end-of-day surveys over a two- to three-week period, asking about their emotional experiences and ER strategy use. Participants completed a total of 3798 experience sampling (in-the-moment) and 995 nightly diary surveys. We examined the top 40% of each participant's reported instances of negative affect (to capture times when emotions more likely need regulation). Results indicated that a higher level of trait emotion dysregulation was associated with the following in-the-moment responses: (a) higher level of negative affect; (b) greater desire to change emotions; (c) more attempts to regulate emotion; (d) higher relative endorsements of avoidant (e.g. thought suppression) versus engagement (e.g. acceptance) ER strategy use; and (e) lower perceived effectiveness of ER. Further, individuals with a higher (vs. lower) level of trait emotion dysregulation were less able to identify emotions over the course of the day. Findings demonstrate how trait emotion dysregulation may predict emotional experiences and ER in daily life.

Leveraging Mobile Sensing to Understand and Develop Intervention Strategies to Improve Medication Adherence.

Interventions to improve medication adherence have had limited success and can require significant human resources to implement. Research focused on improving medication adherence has undergone a paradigm shift, of late, with a shift towards developing personalized, theory-driven interventions. The current research integrates foundational and translational science to implement a mechanisms-focused, context-aware approach. Increasing adoption of mobile and wearable sensing systems presents new opportunities for understanding how medication-taking behaviors unfold in natural settings, especially in populations who have difficulty adhering to medications. When combined with survey and ecological momentary assessment data, these mobile and wearable sensing systems can directly capture the context of medication adherence in situ, including personal, behavioral, and environmental factors. The purpose of this paper is to present a new transdisciplinary research framework in medication adherence, highlight critical advances in this rapidly-evolving research field, and outline potential future directions for both research and clinical applications.

Do I really feel better? Effectiveness of emotion regulation strategies depends on the measure and social anxiety.

BACKGROUND: Effective emotion regulation (ER) is important to long-term healthy functioning, but little is known about what constitutes effective ER in the moment or how social anxiety symptoms and different strategies influence short-term effectiveness outcomes. METHODS: Intensive ecological momentary data from N = 124 college students illustrate how different ways of operationalizing ER effectiveness leads to different conclusions about the short-term effectiveness of different strategies in daily life. RESULTS: When effectiveness is operationalized as the degree to which participants judged that their ER attempts made them feel better, social anxiety severity was negatively associated with effectiveness, and avoidance-oriented strategies were judged to be less effective than engagement-oriented strategies. In contrast, when effectiveness is operationalized as the degree of change in self-reported affect following ER attempts, social anxiety severity was not related to effectiveness, and avoidance-oriented strategies were more effective than engagement-oriented strategies. Social anxiety and ER strategy type did not interact in either model, regardless of how effectiveness was measured. CONCLUSIONS: The study highlights discrepancies when examining two common but distinct ways of measuring the same overarching effectiveness construct, and raises intriguing questions about how forms of psychopathology that are intimately tied to emotion dysregulation, like social anxiety, moderate different ways of measuring the effectiveness of ER attempts.

Relative Mortality Analysis Of The "Golden Hour": A Comprehensive Acuity Stratification Approach To Address Disagreement In Current Literature.

OBJECTIVE: This study sought to address the disagreement in literature regarding the "golden hour" in trauma by using the Relative Mortality Analysis to overcome previous studies' limitations in accounting for acuity when evaluating the impact of prehospital time on mortality. METHODS: The previous studies that failed to support the "golden hour" suffered from limitations in their efforts to account for the confounding effects of patient acuity on the relationship between prehospital time and mortality in their trauma populations. The Relative Mortality Analysis was designed to directly address these limitations using a novel acuity stratification approach, based on patients' probability of survival (PoS), a comprehensive triage metric calculated using Trauma and Injury Severity Score methodology. For this analysis, the population selection and analysis methods of these previous studies were compared to the Relative Mortality Analysis on how they capture the relationship between prehospital time and mortality in the University of Virginia (UVA) Trauma Center population. RESULTS: The methods of the previous studies that failed to support the "golden hour" also failed to do so when applied to the UVA Trauma Center population. However, when applied to the same population, the Relative Mortality Analysis identified a subgroup, 9.9% (with a PoS 23%-91%), of the 5,063 patient population with significantly lower mortality when transported to the hospital within 1 hour, supporting the "golden hour." CONCLUSION: These results suggest that previous studies failed to support the "golden hour" not due to a lack of patients significantly impacted by prehospital time within their trauma populations, but instead due to limitations in their efforts to account for patient acuity. As a result, these studies inappropriately rejected the "golden hour," leading to the current disagreement in literature regarding the relationship between prehospital time and trauma patient mortality. The Relative Mortality Analysis was shown to overcome the limitations of these studies and demonstrated that the "golden hour" was significant for patients who were not low acuity (PoS >91%) or severely high acuity (PoS <23%).

Using Mobile Sensing to Test Clinical Models of Depression, Social Anxiety, State Affect, and Social Isolation Among College Students.

BACKGROUND: Research in psychology demonstrates a strong link between state affect (moment-to-moment experiences of positive or negative emotionality) and trait affect (eg, relatively enduring depression and social anxiety symptoms), and a tendency to withdraw (eg, spending time at home). However, existing work is based almost exclusively on static, self-reported descriptions of emotions and behavior that limit generalizability. Despite adoption of increasingly sophisticated research designs and technology (eg, mobile sensing using a global positioning system [GPS]), little research has integrated these seemingly disparate forms of data to improve understanding of how emotional experiences in everyday life are associated with time spent at home, and whether this is influenced by depression or social anxiety symptoms. OBJECTIVE: We hypothesized that more time spent at home would be associated with more negative and less positive affect. METHODS: We recruited 72 undergraduate participants from a southeast university in the United States. We assessed depression and social anxiety symptoms using self-report instruments at baseline. An app (Sensus) installed on participants' personal mobile phones repeatedly collected in situ self-reported state affect and GPS location data for up to 2 weeks. Time spent at home was a proxy for social isolation. RESULTS: We tested separate models examining the relations between state affect and time spent at home, with levels of depression and social anxiety as moderators. Models differed only in the temporal links examined. One model focused on associations between changes in affect and time spent at home within short, 4-hour time windows. The other 3 models focused on associations between mean-level affect within a day and time spent at home (1) the same day, (2) the following day, and (3) the previous day. Overall, we obtained many of the expected main effects (although there were some null effects), in which higher social anxiety was associated with more time or greater likelihood of spending time at home, and more negative or less positive affect was linked to longer homestay. Interactions indicated that, among individuals higher in social anxiety, higher negative affect and lower positive affect within a day was associated with greater likelihood of spending time at home the following day. CONCLUSIONS: Results demonstrate the feasibility and utility of modeling the relationship between affect and homestay using fine-grained GPS data. Although these findings must be replicated in a larger study and with clinical samples, they suggest that integrating repeated state affect assessments in situ with continuous GPS data can increase understanding of how actual homestay is related to affect in everyday life and to symptoms of anxiety and depression.

Spanish-Language Consumer Health Information Technology Interventions: A Systematic Review.

BACKGROUND: As consumer health information technology (IT) becomes more thoroughly integrated into patient care, it is critical that these tools are appropriate for the diverse patient populations whom they are intended to serve. Cultural differences associated with ethnicity are one aspect of diversity that may play a role in user-technology interactions. OBJECTIVE: Our aim was to evaluate the current scope of consumer health IT interventions targeted to the US Spanish-speaking Latino population and to characterize these interventions in terms of technological attributes, health domains, cultural tailoring, and evaluation metrics. METHODS: A narrative synthesis was conducted of existing Spanish-language consumer health IT interventions indexed within health and computer science databases. Database searches were limited to English-language articles published between January 1990 and September 2015. Studies were included if they detailed an assessment of a patient-centered electronic technology intervention targeting health within the US Spanish-speaking Latino population. Included studies were required to have a majority Latino population sample. The following were extracted from articles: first author's last name, publication year, population characteristics, journal domain, health domain, technology platform and functionality, available languages of intervention, US region, cultural tailoring, intervention delivery location, study design, and evaluation metrics. RESULTS: We included 42 studies in the review. Most of the studies were published between 2009 and 2015 and had a majority percentage of female study participants. The mean age of participants ranged from 15 to 68. Interventions most commonly focused on urban population centers and within the western region of the United States. Of articles specifying a technology domain, computer was found to be most common; however, a fairly even distribution across all technologies was noted. Cancer, diabetes, and child, infant, or maternal health were the most common health domains targeted by consumer health IT interventions. More than half of the interventions were culturally tailored. The most frequently used evaluation metric was behavior/attitude change, followed by usability and knowledge retention. CONCLUSIONS: This study characterizes the existing body of research exploring consumer health IT interventions for the US Spanish-speaking Latino population. In doing so, it reveals three primary needs within the field. First, while the increase in studies targeting the Latino population in the last decade is a promising advancement, future research is needed that focuses on Latino subpopulations previously overlooked. Second, preliminary steps have been taken to culturally tailor consumer health IT interventions for the US Spanish-speaking Latino population; however, focus must expand beyond intervention content. Finally, the field should work to promote long-term evaluation of technology efficacy, moving beyond intermediary measures toward measures of health outcomes.

Rough set theory based prognostic classification models for hospice referral.

BACKGROUND: This paper explores and evaluates the application of classical and dominance-based rough set theory (RST) for the development of data-driven prognostic classification models for hospice referral. In this work, rough set based models are compared with other data-driven methods with respect to two factors related to clinical credibility: accuracy and accessibility. Accessibility refers to the ability of the model to provide traceable, interpretable results and use data that is relevant and simple to collect. METHODS: We utilize retrospective data from 9,103 terminally ill patients to demonstrate the design and implementation RST- based models to identify potential hospice candidates. The classical rough set approach (CRSA) provides methods for knowledge acquisition, founded on the relational indiscernibility of objects in a decision table, to describe required conditions for membership in a concept class. On the other hand, the dominance-based rough set approach (DRSA) analyzes information based on the monotonic relationships between condition attributes values and their assignment to the decision class. CRSA decision rules for six-month patient survival classification were induced using the MODLEM algorithm. Dominance-based decision rules were extracted using the VC-DomLEM rule induction algorithm. RESULTS: The RST-based classifiers are compared with other predictive and rule based decision modeling techniques, namely logistic regression, support vector machines, random forests and C4.5. The RST-based classifiers demonstrate average AUC of 69.74 % with MODLEM and 71.73 % with VC-DomLEM, while the compared methods achieve average AUC of 74.21 % for logistic regression, 73.52 % for support vector machines, 74.59 % for random forests, and 70.88 % for C4.5. CONCLUSIONS: This paper contributes to the growing body of research in RST-based prognostic models. RST and its extensions posses features that enhance the accessibility of clinical decision support models. While the non-rule-based methods-logistic regression, support vector machines and random forests-were found to achieve higher AUC, the performance differential may be outweighed by the benefits of the rule-based methods, particularly in the case of VC-DomLEM. Developing prognostic models for hospice referrals is a challenging problem resulting in substandard performance for all of the evaluated classification methods.

Physician perceptions of barriers to infection prevention and control in labor and delivery.

OBJECTIVE: To learn about the perceptions of healthcare personnel (HCP) on the barriers they encounter when performing infection prevention and control (IPC) practices in labor and delivery to help inform future IPC resources tailored to this setting. DESIGN: Qualitative focus groups. SETTING: Labor and delivery units in acute-care settings. PARTICIPANTS: A convenience sample of labor and delivery HCP attending the Infectious Diseases Society for Obstetrics and Gynecology 2022 Annual Meeting. METHODS: Two focus groups, each lasting 45 minutes, were conducted by a team from the Centers for Disease Control and Prevention. A standardized script facilitated discussion around performing IPC practices during labor and delivery. Coding was performed by 3 reviewers using an immersion-crystallization technique. RESULTS: In total, 18 conference attendees participated in the focus groups: 67% obstetrician-gynecologists, 17% infectious disease physicians, 11% medical students, and 6% an obstetric anesthesiologist. Participants described the difficulty of consistently performing IPC practices in this setting because they often respond to emergencies, are an entry point to the hospital, and frequently encounter bodily fluids. They also described that IPC training and education is not specific to labor and delivery, and personal protective equipment is difficult to locate when needed. Participants observed a lack of standardization of IPC protocols in their setting and felt that healthcare for women and pregnant people is not prioritized on a larger scale and within their hospitals. CONCLUSIONS: This study identified barriers to consistently implementing IPC practices in the labor and delivery setting. These barriers should be addressed through targeted interventions and the development of obstetric-specific IPC resources.

Online cognitive bias modification for interpretation to reduce anxious thinking during the COVID-19 pandemic.

Anxiety disorders are highly prevalent, and rates increased during the COVID-19 pandemic. However, most individuals with elevated anxiety do not access treatment due to barriers such as stigma, cost, and availability. Digital mental health programs, such as cognitive bias modification for interpretation (CBM-I), hold promise in increasing access to care. Before widely disseminating CBM-I, we must rigorously test its effectiveness and determine whom it is best positioned to benefit. The present study (which is a substudy of a parent trial) compared CBM-I against psychoeducation offered through the public website MindTrails, and also tested whether baseline anxiety tied to COVID-19 influenced the rate of change in anxiety and interpretation bias during and after each intervention. Adults with moderate-to-severe anxiety symptoms were randomly assigned to complete five sessions of either CBM-I or psychoeducation as part of a larger trial, and 608 enrolled in this substudy after Session 1. As predicted (https://osf.io/2dyzr), CBM-I was superior to psychoeducation at reducing anxiety symptoms (on the OASIS but not the DASS-21-AS: d = -0.31), reducing negative interpretation bias (d range = -0.34 to -0.43), and increasing positive interpretation bias (d = 0.79) by the end of treatment. Results also indicated that individuals higher (vs. lower) in baseline COVID-19 anxiety had stronger decreases in anxiety symptoms while receiving CBM-I but weaker decreases in anxiety symptoms (on the DASS-21-AS) while receiving psychoeducation. These findings suggest that CBM-I may be a useful anxiety-reduction tool for individuals experiencing higher anxiety tied to uncertain events such as the COVID-19 pandemic.

Perceived Burdensomeness and Thwarted Belongingness in Text Messages of Suicide Attempt Survivors.

OBJECTIVE: Perceived burdensomeness and thwarted belongingness are considered interpersonal risk factors for suicide. Examining these themes in personal text messages may help identify proximal suicide risk. METHOD: Twenty-six suicide attempt survivors provided personal text messages and reported dates for past periods characterized by positive mood, depressed mood, suicidal ideation (with no attempt), or the two-week period leading up to suicide attempt(s). Texts were then classified into the applicable period based on matching dates. Texts (N = 194,083; including n = 86,705 outgoing texts) were coded for perceived burdensomeness and thwarted belongingness by masked trained raters. Multilevel models were fit to examine whether the target themes (combined into one overall interpersonal risk variable due to low base rate) were more prevalent in texts sent during higher risk episodes (e.g., suicide attempt vs. depressed mood episodes). RESULTS: 0.57% of outgoing texts contained either target theme. As hypothesized, logistic models showed participants were more likely to send texts containing the target themes during suicide attempt episodes relative to suicidal ideation (with no attempt) episodes, depressed mood episodes, and positive mood episodes, and during suicidal ideation (with no attempt) episodes relative to positive mood episodes. All contrasts were robust to post-hoc correction except for suicide attempt episodes vs. ideation (with no attempt) episodes. No other significant pairwise differences for episode type emerged. CONCLUSIONS: Despite the small sample size and low base rate of target themes in the texts, perceived burdensomeness and thwarted belongingness were associated with intra-individual suicide risk severity in personal text messages.

Burnout and staff turnover among certified nursing assistants working in acute care hospitals during the COVID-19 pandemic.

INTRODUCTION: Healthcare worker burnout is a growing problem in the United States which affects healthcare workers themselves, as well as the healthcare system as a whole. The goal of this qualitative assessment was to understand factors that may lead to healthcare worker burnout and turnover through focus groups with Certified Nursing Assistants who worked in acute care hospitals during the COVID-19 pandemic. METHODS: Eight focus group discussions lasting approximately 30 minutes each were held remotely from October 2022-January 2023 with current and former Certified Nursing Assistants who worked during the COVID-19 pandemic in acute care hospitals. Participants were recruited through various sources such as social media and outreach through professional organizations. The focus groups utilized open-ended prompts including topics such as challenges experienced during the pandemic, what could have improved their experiences working during the pandemic, and motivations for continuing or leaving their career in healthcare. The focus groups were coded using an immersion-crystallization technique and summarized using NVivo and Microsoft Excel. Participant demographic information was summarized overall and by current work status. RESULTS: The focus groups included 58 Certified Nursing Assistants; 33 (57%) were current Certified Nursing Assistants and 25 (43%) were Certified Nursing Assistants who no longer work in healthcare. Throughout the focus groups, five convergent themes emerged, including staffing challenges, respect and recognition for Certified Nursing Assistants, the physical and mental toll of the job, facility leadership support, and pay and incentives. CONCLUSIONS: Focus group discussions with Certified Nursing Assistants identified factors at individual and organizational levels that might contribute to burnout and staff turnover in healthcare settings. Suggestions from participants on improving their experiences included ensuring staff know they are valued, being included in conversations with leadership, and improving access to mental health resources.

Feasibility and acceptability testing of CommSense: A novel communication technology to enhance health equity in clinician-patient interactions.

Background: Quality patient-clinician communication is paramount to achieving safe and compassionate healthcare, but evaluating communication performance during real clinical encounters is challenging. Technology offers novel opportunities to provide clinicians with actionable feedback to enhance their communication skills. Methods: This pilot study evaluated the acceptability and feasibility of CommSense, a novel natural language processing (NLP) application designed to record and extract key metrics of communication performance and provide real-time feedback to clinicians. Metrics of communication performance were established from a review of the literature and technical feasibility verified. CommSense was deployed on a wearable (smartwatch), and participants were recruited from an academic medical center to test the technology. Participants completed a survey about their experience; results were exported to SPSS (v.28.0) for descriptive analysis. Results: Forty (n = 40) healthcare participants (nursing students, medical students, nurses, and physicians) pilot tested CommSense. Over 90% of participants "strongly agreed" or "agreed" that CommSense could improve compassionate communication (n = 38, 95%) and help healthcare organizations deliver high-quality care (n = 39, 97.5%). Most participants (n = 37, 92.5%) "strongly agreed" or "agreed" they would be willing to use CommSense in the future; 100% (n = 40) "strongly agreed" or "agreed" they were interested in seeing information analyzed by CommSense about their communication performance. Metrics of most interest were medical jargon, interruptions, and speech dominance. Conclusion: Participants perceived significant benefits of CommSense to track and improve communication skills. Future work will deploy CommSense in the clinical setting with a more diverse group of participants, validate data fidelity, and explore optimal ways to share data analyzed by CommSense with end-users.

Inferring sleep disturbance from text messages of suicide attempt survivors: A pilot study.

OBJECTIVE: Identifying digital markers of sleep disturbance-a known suicide risk factor-may aid in the detection of imminent suicide risk. This study examined sleep-related communication and texting patterns in personal text messages (N = 86,705) of suicide attempt survivors. METHOD: Twenty-six participants provided dates of past suicide attempts and 2-week periods of positive mood, depressed mood, or suicidal ideation. Linguistic Inquiry Word Count was used to identify sleep-related texts via a custom dictionary. Mixed effect models were fitted to test the association between suicide/mood episode type (e.g., attempt versus ideation) and three outcomes: likelihood of a text including sleep-related content, nightly count of texts sent from midnight to 5:00 AM, and sum of unique hour bins from midnight to 5:00 AM with outgoing texts. RESULTS: Analyses with a sleep dictionary that was manually revised to be more accurate (but not the original unedited dictionary) showed sleep-related communication was more likely during depressed mood episodes than positive mood episodes. Otherwise, there were no significant differences in sleep-related communication or objective texting patterns across episode type. CONCLUSIONS: Although we did not detect differences in sleep-related communication tied to suicidal thoughts or behaviors, sleep-related communication may differ as a function of within-person mood level.

Personalized State Anxiety Detection: An Empirical Study with Linguistic Biomarkers and A Machine Learning Pipeline.

Individuals high in social anxiety symptoms often exhibit elevated state anxiety in social situations. Research has shown it is possible to detect state anxiety by leveraging digital biomarkers and machine learning techniques. However, most existing work trains models on an entire group of participants, failing to capture individual differences in their psychological and behavioral responses to social contexts. To address this concern, in Study 1, we collected linguistic data from N=35 high socially anxious participants in a variety of social contexts, finding that digital linguistic biomarkers significantly differ between evaluative vs. non-evaluative social contexts and between individuals having different trait psychological symptoms, suggesting the likely importance of personalized approaches to detect state anxiety. In Study 2, we used the same data and results from Study 1 to model a multilayer personalized machine learning pipeline to detect state anxiety that considers contextual and individual differences. This personalized model outperformed the baseline's F1-score by 28.0%. Results suggest that state anxiety can be more accurately detected with personalized machine learning approaches, and that linguistic biomarkers hold promise for identifying periods of state anxiety in an unobtrusive way.

Shifting Episodic Prediction With Online Cognitive Bias Modification: A Randomized Controlled Trial.

Negative future thinking pervades emotional disorders. This hybrid efficacy-effectiveness trial tested a four-session, scalable online cognitive bias modification program for training more positive episodic prediction. 958 adults (73.3% female, 86.5% White, 83.4% from United States) were randomized to positive conditions with ambiguous future scenarios that ended positively, 50/50 conditions that ended positively or negatively, or a control condition with neutral scenarios. As hypothesized (preregistration: https://osf.io/jrst6), positive training participants improved more than control participants in negative expectancy bias (d = -0.58), positive expectancy bias (d = 0.80), and self-efficacy (d = 0.29). Positive training was also superior to 50/50 training for expectancy bias and optimism (d = 0.31). Training gains attenuated yet remained by 1-month follow-up. Unexpectedly, participants across conditions improved comparably in anxiety and depression symptoms and growth mindset. Targeting a transdiagnostic process with a scalable program may improve bias and outlook; however, further validation of outcome measures is required.

Detecting Social Contexts from Mobile Sensing Indicators in Virtual Interactions with Socially Anxious Individuals.

Mobile sensing is a ubiquitous and useful tool to make inferences about individuals' mental health based on physiology and behavior patterns. Along with sensing features directly associated with mental health, it can be valuable to detect different features of social contexts to learn about social interaction patterns over time and across different environments. This can provide insight into diverse communities' academic, work and social lives, and their social networks. We posit that passively detecting social contexts can be particularly useful for social anxiety research, as it may ultimately help identify changes in social anxiety status and patterns of social avoidance and withdrawal. To this end, we recruited a sample of highly socially anxious undergraduate students (N=46) to examine whether we could detect the presence of experimentally manipulated virtual social contexts via wristband sensors. Using a multitask machine learning pipeline, we leveraged passively sensed biobehavioral streams to detect contexts relevant to social anxiety, including (1) whether people were in a social situation, (2) size of the social group, (3) degree of social evaluation, and (4) phase of social situation (anticipating, actively experiencing, or had just participated in an experience). Results demonstrated the feasibility of detecting most virtual social contexts, with stronger predictive accuracy when detecting whether individuals were in a social situation or not and the phase of the situation, and weaker predictive accuracy when detecting the level of social evaluation. They also indicated that sensing streams are differentially important to prediction based on the context being predicted. Our findings also provide useful information regarding design elements relevant to passive context detection, including optimal sensing duration, the utility of different sensing modalities, and the need for personalization. We discuss implications of these findings for future work on context detection (e.g., just-in-time adaptive intervention development).