Virtually screening adults for depression, anxiety, and suicide risk using machine learning and language from an open-ended interview.

Background: Current depression, anxiety, and suicide screening techniques rely on retrospective patient reported symptoms to standardized scales. A qualitative approach to screening combined with the innovation of natural language processing (NLP) and machine learning (ML) methods have shown promise to enhance person-centeredness while detecting depression, anxiety, and suicide risk from in-the-moment patient language derived from an open-ended brief interview. Objective: To evaluate the performance of NLP/ML models to identify depression, anxiety, and suicide risk from a single 5-10-min semi-structured interview with a large, national sample. Method: Two thousand four hundred sixteen interviews were conducted with 1,433 participants over a teleconference platform, with 861 (35.6%), 863 (35.7%), and 838 (34.7%) sessions screening positive for depression, anxiety, and suicide risk, respectively. Participants completed an interview over a teleconference platform to collect language about the participants' feelings and emotional state. Logistic regression (LR), support vector machine (SVM), and extreme gradient boosting (XGB) models were trained for each condition using term frequency-inverse document frequency features from the participants' language. Models were primarily evaluated with the area under the receiver operating characteristic curve (AUC). Results: The best discriminative ability was found when identifying depression with an SVM model (AUC = 0.77; 95% CI = 0.75-0.79), followed by anxiety with an LR model (AUC = 0.74; 95% CI = 0.72-0.76), and an SVM for suicide risk (AUC = 0.70; 95% CI = 0.68-0.72). Model performance was generally best with more severe depression, anxiety, or suicide risk. Performance improved when individuals with lifetime but no suicide risk in the past 3 months were considered controls. Conclusion: It is feasible to use a virtual platform to simultaneously screen for depression, anxiety, and suicide risk using a 5-to-10-min interview. The NLP/ML models performed with good discrimination in the identification of depression, anxiety, and suicide risk. Although the utility of suicide risk classification in clinical settings is still undetermined and suicide risk classification had the lowest performance, the result taken together with the qualitative responses from the interview can better inform clinical decision-making by providing additional drivers associated with suicide risk.

User Feedback on the Use of a Natural Language Processing Application to Screen for Suicide Risk in the Emergency Department.

Suicide is the 10th leading cause of death in the USA and globally. Despite decades of research, the ability to predict who will die by suicide is still no better than 50%. Traditional screening instruments have helped identify risk factors for suicide, but they have not provided accurate predictive power for reducing death rates. Over the past decade, natural language processing (NLP), a form of machine learning (ML), has been used to identify suicide risk by analyzing language data. Recent work has demonstrated the successful integration of a suicide risk screening interview to collect language data for NLP analysis from patients in two emergency departments (ED) of a large healthcare system. Results indicated that ML/NLP models performed well identifying patients that came to the ED for suicide risk. However, little is known about the clinician's perspective of how a qualitative brief interview suicide risk screening tool to collect language data for NLP integrates into an ED workflow. This report highlights the feedback and observations of patient experiences obtained from clinicians using brief suicide screening interviews. The investigator used an open-ended, narrative interview approach to inquire about the qualitative interview process. Three overarching themes were identified: behavioral health workflow, clinical implications of interview probes, and integration of an application into provider patient experience. Results suggest a brief, qualitative interview method was feasible, person-centered, and useful as a suicide risk detection approach.

Suicidal ideation in veterans enrolled in evidence-based psychotherapy for posttraumatic stress disorder.

Suicidal ideation (SI) is a highly prevalent public health issue in the veteran population and is increasingly common in veterans who are diagnosed with other mental health conditions, such as posttraumatic stress disorder (PTSD; U.S. Department of Veterans Affairs, 2020). The present study has an initial aim of examining changes in SI over treatment, and it is hypothesized that SI will decrease across PTSD treatments. A second aim is to examine the association of SI status with PTSD symptoms across treatment, and it is hypothesized that PTSD symptomatology will decrease at similar rates over the course of treatment for those who did and did not endorse SI at pretreatment. Participants included 717 (86.3% male) veterans who participated in outpatient treatment within a Veterans Affairs Post Traumatic Stress Disorder (VA PTSD) specialty clinic between July 2014 and December 2017. Descriptive analyses found that 37.2% of veterans endorsed SI at pretreatment, while 18.6% endorsed SI at posttreatment. The relationship between pre- and posttreatment SI was significant, χ²(1, N = 247) = 23.77, p < .001. A significant proportion of veterans who endorsed SI at pretreatment no longer endorsed SI at posttreatment (64.7%). There were no differences in changes in PTSD Checklist for DSM-5 (PCL-5) scores across treatment for those with and without SI at pretreatment. While those who endorsed SI at pretreatment had higher PCL-5 scores throughout treatment, they experienced a similar rate of improvement in symptoms as those without SI at pretreatment. This finding suggests that the presence of SI does not reduce the effectiveness of PTSD treatment. Limitations include the use of a single-item measure of SI, lack of adequate power to detect difference among treatments, and a cross-sectional design. Clinical and research implications are discussed. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Impact of residential PTSD treatment on suicide risk in veterans.

INTRODUCTION: In 2019, 17 veterans died by suicide every day. Various suicide prevention treatments have emerged, yet limited research has explored the impact of Post Traumatic Stress Disorder (PTSD) treatment on suicidal ideation and behaviors. METHODS: This study examines the impact of Cognitive Processing Therapy (CPT) on suicidal ideation among veterans in three residential PTSD programs (women's, men's, and PTSD/Traumatic Brain Injury). Interview and self-report data were collected from veterans (n = 446) throughout treatment. RESULTS: Over 50% of veterans reported current suicidal ideation and a history of suicide attempts prior to treatment. Variables that predicted change in suicidal ideation included prior suicide attempt (ß = 0.21, p = 0.022), change in CAPS-5 total score (ß = 0.28, p = 0.038), employment status (ß = -0.20, p = 0.035) and history of suicide attempt (ß = 0.25, p = 0.009). Those without a previous suicide attempt made greater gains in CPT treatment than those with a previous suicide attempt. CONCLUSION: Following 7 weeks of CPT residential treatment, a decrease in PTSD symptoms was significantly associated with a reduction in suicidal ideation. Implications are that CPT can reduce suicide risk in a variety of Veteran cohorts with differing trauma types.

How Veterans Health Administration Suicide Prevention Coordinators Assess Suicide Risk.

This cross-sectional study was designed to examine the suicide risk assessment practices of Suicide Prevention Coordinators (SPCs) within the Veterans Health Administration. Specifically, this study sought to (1) identify factors SPCs consider most important in assessing risk and patient priority; (2) measure the level of consistency and agreement between SPCs in assessing suicide risk and prioritizing cases; and (3) measure individual SPC consistency between cases. SPCs (n = 63) responded to online survey questions about imminent and prolonged risk for suicide in response to 30 fictional vignettes. Combinations of 12 acute and chronic suicide risk factors were systematically distributed throughout the 30 vignettes using the Fedorov () procedure. The SPCs were also asked to identify the level of priority for further assessment both disregarding and assuming current caseloads. Data were analysed using clinical judgement analysis. Suicidal plan, ß = 1.64; 95% CI (1.45, 1.82), and preparatory behaviour, ß = 1.40; 95% CI (1.23, 1.57), were considered the most important acute or imminent risk factors by the SPCs. There was less variability across clinicians in the assessment of risk when alcohol use (p = 0.02) and hopelessness (p = 0.03) were present. When considering acute or imminent risk factors, there was considerable variability between clinicians on a vignette-by-vignette basis, median SD = 0.86 (range = 0.47, 1.13), and within individual clinicians across vignettes, median R2 = 0.80 (0.49, 0.95). These findings provide insight into how this group of providers think about acute and chronic risk factors contributing to imminent suicide risk in Veterans. Copyright © 2016 John Wiley & Sons, Ltd. KEY PRACTITIONER MESSAGE: Identifies factors that practitioners consider most important in suicide risk assessment Discusses how to distinguish between chronic and acute risk for suicide Identifies factors that lead to more consistent clinical judgments.

Values as Moderators of the Association between Interpersonal-Psychological Constructs and Suicidal Ideation among Veterans.

This study examined whether valuing relationships, achievement, and security moderated the association between interpersonal-psychological constructs (Joiner, 2005) and suicidal ideation (SI). A total of 122 veterans completed the Interpersonal Needs Questionnaire, Survey of Life Principles, and Beck Scale for Suicide Ideation. Valuing relationships moderated the association between thwarted belongingness and SI. Specifically, thwarted belongingness predicted SI among veterans who reported moderate and high, but not low, levels of valuing relationships. The estimated impact of perceived burdensomeness on SI was stronger at higher levels of valuing relationships, but only approached statistical significance. Valuing achievement and security did not moderate the association between perceived burdensomeness and SI. Future research should continue to examine specific values as they relate to interpersonal-psychological constructs and suicidal behavior.

Military Service and Suicidal Thoughts and Behaviors in a National Sample of College Students.

Background: As a result of the post-9/11 GI Bill, increasing numbers of veterans are enrolling in college. However, little is known regarding suicidal outcomes among this group. In prior research, college student veterans reported high rates of suicidal ideation and attempt. Nonetheless, no research has examined whether military service is associated with increased suicide risk among college students. Aims: Our primary aims were to examine whether a history of military service was related to past-year suicidal ideation, plan, and attempt among college students. On the basis of previous research with college students, we hypothesized that students with a history of military service (i.e., current or prior) would report a higher percentage of past-year suicidal ideation, plan, and attempt. Our secondary aims were to examine the associations between military service and major depression and nonsuicidal self-injury. Method: Our sample included 3,290 college students with and without a history of military service who participated in the Healthy Minds Study in 2011 and 2012. Results: Military service was not significantly associated with past-year suicidal ideation, plan, or attempt. Students without a history of military service were more likely to report nonsuicidal self-injury. There was no significant difference in screening positive for major depression. Conclusions: These findings conflict with previous research that identified student veterans as being at elevated risk.

Suicide risk among lesbian, gay, bisexual, and transgender military personnel and veterans: what does the literature tell us?

Research suggests that both the military and veteran and the lesbian, gay, bisexual, and transgender (LGBT) populations may be at increased risk for suicide. A literature review was conducted to identify research related to suicide risk in the LGBT military and veteran populations. Despite the paucity of research directly addressing this issue, themes are discussed evident in the literature on LGBT identity and suicide risk as well as LGBT military service members and veterans. Factors such as social support and victimization appear to be particularly relevant. Suggestions are made with respect to future research that is needed on this very important and timely topic.

Sleep problems outperform depression and hopelessness as cross-sectional and longitudinal predictors of suicidal ideation and behavior in young adults in the military.

BACKGROUND: Sleep problems appear to represent an underappreciated and important warning sign and risk factor for suicidal behaviors. Given past research indicating that disturbed sleep may confer such risk independent of depressed mood, in the present report we compared self-reported insomnia symptoms to several more traditional, well-established suicide risk factors: depression severity, hopelessness, PTSD diagnosis, as well as anxiety, drug abuse, and alcohol abuse symptoms. METHODS: Using multiple regression, we examined the cross-sectional and longitudinal relationships between insomnia symptoms and suicidal ideation and behavior, controlling for depressive symptom severity, hopelessness, PTSD diagnosis, anxiety symptoms, and drug and alcohol abuse symptoms in a sample of military personnel (N=311). RESULTS: In support of a priori hypotheses, self-reported insomnia symptoms were cross-sectionally associated with suicidal ideation, even after accounting for symptoms of depression, hopelessness, PTSD diagnosis, anxiety symptoms and drug and alcohol abuse. Self-reported insomnia symptoms also predicted suicide attempts prospectively at one-month follow up at the level of a non-significant trend, when controlling for baseline self-reported insomnia symptoms, depression, hopelessness, PTSD diagnosis and anxiety, drug and alcohol abuse symptoms. Insomnia symptoms were unique predictors of suicide attempt longitudinally when only baseline self-reported insomnia symptoms, depressive symptoms and hopelessness were controlled. LIMITATIONS: The assessment of insomnia symptoms consisted of only three self-report items. Findings may not generalize outside of populations at severe suicide risk. CONCLUSIONS: These findings suggest that insomnia symptoms may be an important target for suicide risk assessment and the treatment development of interventions to prevent suicide.