BACKGROUND:Individuals with bipolar disorder (BD) often present with language abnormalities, but few studies have quantified these features or examined their relationship with white matter microstructure. METHODS:We analyzed language samples from 17 Veterans with BD and 15 healthy Veteran controls elicited with an open-ended prompt about a stressful interpersonal experience. We applied singular value decomposition (SVD) to 275 linguistic features to derive low-dimensional component scores capturing the dominant covariance structure. In parallel, we conducted feature-level between-group tests and entered features showing nominal group differences (p < .05, uncorrected) into an exploratory principal components analysis (PCA) with oblique rotation to derive interpretable components. Neurite orientation dispersion and density imaging (NODDI) was used to characterize intracellular and extracellular microstructural properties within a mask of the arcuate fasciculus. RESULTS:SVD revealed a significant group difference on the first component (p = .03). Feature-level analyses identified 16 variables showing group differences. PCA of these features yielded four interpretable components: (1) model predictability and present focus, (2) agency and directedness, (3) central concept connectedness, and (4) lexical and grammatical complexity, together explaining 75% of the variance. Within the BD group, lower scores on the first component were associated with greater neurite orientation dispersion in the arcuate fasciculus. LIMITATIONS:Cross-sectional design, small sample size, and lack of correction for multiple comparisons. CONCLUSIONS:Preliminary findings indicate quantifiable language abnormalities in Veterans with BD and suggest links to altered dispersion of axons and dendrites in language-associated tracts.

Machine learning predictive modeling can support scalable prevention of suicide-related behavior (SRB). SAFEGUARD is a three-pronged universal, indicated, and clinical SRB-prevention intervention system focused on key military career touchpoints. The targeted SAFEGUARD interventions are designed to improve on the mixed results of universal interventions. Level Up uses digital tools, personalized messaging, and remote booster sessions to deliver customized universal military-focused cognitive behavioral therapy skills training designed to reduce SRBs during first duty assignments. Operation Life Force delivers remote group dialectical behavior therapy skills training with a mental toughness focus to soldiers identified during annual physicals as high-risk for SRBs. Pathfinding delivers remote wrap-around case management after psychiatric inpatient discharge to soldiers identified as high-risk for SRBs. SAFEGUARD is a data-driven system for SRB prevention that delivers targeted best-practice interventions at critical points to optimize impact and efficiently use mental health resources across the military.

BACKGROUND:The polytrauma clinical triad (PCT)-co-occurring posttraumatic stress disorder (PTSD), traumatic brain injury (TBI), and chronic pain-is common among U.S. service members and veterans. While known to result in worse clinical outcomes, the effect of the PCT on suicide in military populations remains unknown. The current study examined the effect of the PCT on suicidal ideation and attempts in U.S. veterans and service members. METHOD:In Study 1, veterans (N = 351) completed self-reports of PTSD symptoms and chronic pain and clinician interviews of TBI and suicidal ideation and attempts. In study 2, Army soldiers (N = 8558) completed self-reports of PTSD symptoms, chronic pain, TBI, and suicidal ideation and attempts across four assessments spanning from 1-month prior to deployment to 9-months post-deployment. RESULTS:The PCT was more prevalent in veterans (30.09%) than service members (2.51%), and was associated with greater odds of suicidal ideation in veterans (OR = 4.11) when compared with only two conditions. When compared with one or no conditions, the PCT was associated with greater odds of suicidal ideation (ORs = 2.06-2.94) and attempts (ORs = 4.06-4.40) in service members. PTSD seemed to drive this effect. CONCLUSION:Findings support the PCT as a potentially important risk factor for suicide in both current and former military members. Future research should attempt to understand the impact of the PCT on mechanisms of suicide risk and suicide risk screening should be prioritized in polytrauma clinics where patients with the PCT are likely to be seen.

OBJECTIVE:Death by suicide among U.S. servicemembers is a significant public health concern - firearms are the primary method of suicide for active-duty servicemembers. The Overwatch Project provides peer-based firearms lethal means safety training for servicemembers to mitigate future risk. This study reports on the effectiveness of this training. METHODS:This longitudinal, quasi-experimental and matched-cohort study evaluated the effectiveness of the Overwatch Project by comparing pre-, post-, and two-month follow-up self-report data for active-duty servicemembers who attended the training (Arm1/Overwatch Project, n = 451) and those who did not (Arm2/Comparison Group, n = 451). RESULTS:Participation in the Overwatch Project training was associated with a significant increase in servicemembers meeting the minimum score on a firearms lethal means safety knowledge assessment (within group; pre-training: 8% vs. post-training: 85%), readiness in executing proactive firearms lethal means safety tasks (between group), creation of a proactive firearms lethal means safety plan for themselves (within group; pre-training: 26% vs. two-month follow-up: 68%), and creation of a similar plan with a peer (within group; pre-training: 26% vs. two-month follow-up: 37%). CONCLUSION:Preliminary findings suggest that the Overwatch Project peer-based firearms lethal means safety training is a promising step towards reducing firearm suicides among servicemembers.

BACKGROUND:Veterans experiencing criminal-legal involvement and/or homelessness (VCLI/H) are at elevated risk for suicide attempt and death relative to the broader Veteran population. Characterizing suicide-related outcomes in these Veteran groups is vital for Veteran suicide prevention efforts. METHODS:Using data from clinician-administered suicide risk assessments gathered from Veteran electronic medical records, this study examined rates and correlates of suicide-related outcomes in 249,011 VCLI/H. RESULTS:At assessment, 1 in 7 VCLI/H endorsed suicidal ideation, and 1 in 32 were designated by clinicians as high acute risk for suicide. Within 12 months following assessment, 1 in 37 VCLI/H had a documented suicide attempt. These outcomes were more common among Veterans experiencing both criminal-legal involvement and homelessness than Veterans experiencing only one of these. Suicide attempts were most common among Veterans with bipolar disorders, psychotic disorders other than schizophrenia, and personality disorders compared to other psychiatric disorders. Clinician-identified risk demonstrated high specificity but low sensitivity in predicting subsequent suicide attempts of VCLI/H. CONCLUSIONS:Findings reiterate the exceptionally high prevalence of suicide-related outcomes among VCLI/H and suggest a potential need to augment current risk assessment practices to increase sensitivity to better predict suicide-related outcomes among members of these populations.

Suicide prevention is a top priority of the Department of Veterans Affairs (VA), and improving the effectiveness of prevention efforts requires more precise identification of those at risk for suicide. Recent research identifies aggression as a risk factor for veteran suicide, yet little is known about moderators of the aggression-suicide association. This study examines demographic, clinical, and military characteristics as potential moderators of the association between aggression and suicide in a sample of veterans at high risk for suicide. This cross-sectional study used baseline data from a randomized clinical trial across four VA sites. Participants (n = 207veterans; 82.13% men) completed a clinical interview of suicide attempts and clinical characteristics and self-report measures of aggression, demographics, and military characteristics. We fit zero-inflated negative binomial models with interaction terms between aggression and each characteristic to test for moderators of the aggression-suicide association. None of the veteran characteristics had a significant main effect on number of suicide attempts. Across all four aggression subscales, gender and deployment history moderated the aggression-suicide association such that veteran women and those without a history of deployment had stronger positive associations between aggression and suicide. The present study adds nuance to our understanding of the relationship between aggression and suicide in veterans. Findings suggest that intervention efforts for high-risk veterans should focus on the role of aggression for veteran women and those without a history of deployment. Further research is warranted to validate findings and develop more precise suicide prevention efforts.

BACKGROUND:Among individuals with neuropsychiatric disorders, those with bipolar disorder (BD) have one of the highest rates of suicide with this risk even further elevated in the Veteran population. The assessment of suicide risk is clinically challenging, but one of the best predictors of a future suicide attempt is having a prior attempt. Although prior studies have implicated proteomic abnormalities separately in BD and among those with a suicide attempt history, little is known regarding their combined contribution to suicide risk in BD and/or their relationship to the brain white matter. METHODS:We analyzed 368 proteins targeting disease and pathway specific biomarkers using the OLINK platform in Veterans with BD either having (BD/SA+) or not having (BD/SA-) a past suicide attempt (SA) history and healthy control (HC) Veterans. We used neurite orientation and dispersion density imaging to derive average indices of fiber coherence (i.e., orientation dispersion index, ODI), axon and dendrite density (i.e., neurite density, NDI) and extracellular free-water indicative of diffusion (isotropic volume fraction, ISOVF) within the brain white matter. RESULTS:Three proteins (NEP, KYNU and IL-12B) had significantly lower expression in both the BD/SA+ and BD/SA- groups compared to HC. In the combined group of individuals with BD, lower expression of these 3 proteins correlated significantly with lower neurite density across the brain white matter. Two proteins had significantly higher (i.e., AZU1 and CEACAM8) and one significantly lower (i.e., MMP7) expression in the BD/SA+ group compared to both the BD/SA- and HC groups. In the BD/SA+ group, higher protein expression of AZU1 and CEACAM8 was associated with a higher ISOVF in the brain white matter. Pathway analysis indicated enrichment of innate immune, neutrophil degranulation, and cytokine signaling pathways among the differentially expressed proteins. DISCUSSION:In this exploratory study we found that among all individuals with BD, proteomic abnormalities were associated with an index of neurite density (i.e., axonal and dendritic packing) within brain white matter, whereas in BD/SA+ abnormal protein expression was associated with a higher ISOVF, potentially reflecting inflammation, edema and/or atrophy. Immune-related biology may represent a mechanistic link between peripheral inflammatory signals and alterations in brain microstructure among individuals with BD at heightened risk of suicide, but should be considered preliminary until replicated in larger cohorts.

Some veterans are haunted by memories of action they have taken or betrayals they have experienced that violated deeply held moral beliefs; these experiences can lead to moral injury. We have developed a depth-oriented group psychotherapy for U.S. combat veterans, to address moral injury. Depth psychotherapy is an evidence-based form of psychoanalysis; the treatment we have developed is based on Relational psychoanalysis. The aim is for the group members to each develop an organized narrative about morally injurious events and their impact on their current lives to facilitate psychosocial recovery. The hypothesized change agents of this treatment are, in order of their use in the sessions: (1) warm-up team-building activities such as exercises from the improv and psychodrama/sociometry traditions; (2) reflective listening and speaking; (3) sharing moral injury event narratives with trusted others. The clinical model we have developed for treating moral injury emphasizes that veterans will be asked to describe, to the extent that they are able, the feelings, sensations, and fragmentary thoughts that are initially hard to articulate and sometimes difficult to recall. The goal of this article is to describe relevant depth psychology theory, its application to the moral injury context, the relevance of depth-oriented group psychotherapy for moral injury and, further, the depth-oriented group psychotherapy approach we have derived from these ideas.

QuestionDoes Project Life Force (PLF), a manualized suicide safety planning group intervention augmented with skills training, reduce suicidal behavior among veterans at high risk for suicide over 12 months?FindingsIn this randomized clinical trial of 207 veterans at high risk for suicide, no difference was found in time to first suicidal behavior between participants in the PLF arm compared with those in the treatment as usual (TAU) arm. However, PLF participants demonstrated a significantly lower number of actual attempts and improved suicide-related coping than TAU participants.MeaningThese findings suggest that PLF may lower suicide risk among veterans at high risk for suicide. This randomized clinical trial evaluates Project Life Force, a group-based suicide safety planning and skills training intervention for veterans at high risk for suicide, comparing its effects to standard treatment over a 12-month period. ImportanceNovel evidence-based suicide-specific treatments are critical to addressing rising rates of suicide among veterans.ObjectiveTo determine whether Project Life Force (PLF), a group intervention that augments suicide safety planning (SSP) with skills training, plus treatment as usual (TAU) delays or decreases suicidal behavior compared with TAU alone.Design, Setting, and ParticipantsThis multisite randomized clinical trial, using intention-to-treat analyses, included veterans at high risk for suicide. Participants were recruited after psychiatric inpatient discharge or referral from 4 Veterans Health Administration outpatient care teams in hospitals in Pennsylvania, New York, and Texas between March 2018 and February 2024. Data were analyzed from April 2024 to October 2025.InterventionPLF is a manualized, 10- to 12-session (75-90 minutes per session) group intervention combining SSP with distress tolerance, emotion regulation, and interpersonal skills training. Sessions cover each of the 6 SSP steps, health management, reasons for living, SSP accessibility, and a recap, with up to 4 optional booster sessions offered. All participants received TAU, constituting individual SSP and standard outpatient mental health care.Main Outcomes and MeasuresThe primary outcome was time in days to first suicidal behavior (aborted, interrupted, or actual attempts or suicide death) over 12 months, assessed using the Columbia Suicide Severity Rating Scale and chart review. Secondary outcomes over 12 months included self-report measures of hopelessness, depression, treatment use, attitudes toward help-seeking, and suicide-related coping.ResultsA total of 207 participants (mean [SD] age, 46.1 [13.9] years; 178 male [86%]) were assessed at baseline and randomly assigned to PLF plus TAU (101 participants) or TAU alone (106 participants). Cox proportional hazards models showed no significant between-group difference in time to suicidal behavior; however, compared with TAU, PLF participants had a lower hazard of actual suicide attempts over 1 year (HR, 0.49; 95% CI, 0.26-0.92; P = .03). Secondary outcome analyses included data from 132 participants (64%) at posttreatment, 125 (60%) at month 6, and 130 (62%) at month 12. Treatment-by-time effects ranged from Cohen d = -0.03 to 0.35. Notably, PLF significantly improved suicide-related coping at posttreatment (estimate, 4.06; Cohen d, 0.21; P = .04) and month 6 (estimate, 5.36; Cohen d, 0.27; P = .007). No study-related adverse events occurred.Conclusions and RelevanceIn this randomized clinical trial of 207 participants, PLF plus TAU did not change time to suicidal behavior but did increase time to actual suicide attempts and improve suicide-related coping, thereby enhancing SSP impact.Trial RegistrationClinicalTrials.gov Identifier: NCT03653637

[This corrects the article DOI: 10.1016/j.conctc.2025.101576.].

OBJECTIVES:Personal recovery following a suicidal episode is multifaceted and understudied. Network analysis of the Post Acute Suicidal Episode (PASE), the recovery period that follows an acute episode of suicidal consideration, planning, or attempt, may clarify the multidimensional structure of this recovery and identify clinical intervention points. METHODS:An online sample with a history of a suicidal episode (N = 940) completed the Recovery Evaluation and Suicide Support Tool (RESST). We applied exploratory graph analysis (EGA) to identify recovery clusters, expected influence and bridge centrality to identify central features, and network comparison tests (NCTs) to compare network structure by suicide attempt history. Dimensional stability was additionally tested across a three-week interval. RESULTS:EGA identified four stable clusters mirroring the RESST domains, and dimensional structure was consistent across the three-week interval. Central features included life worth, positive self-regard, feeling valued, and connectedness. The strongest bridge items included feeling that life is worth living and having meaning and purpose. NCTs did not identify structural divergence by suicide attempt history. CONCLUSION:PASE recovery showed a stable multidimensional structure, with no divergence detected by suicide attempt history or across a three-week interval. Central and bridge features (life worth, meaning, connectedness) are promising candidate intervention targets that need prospective testing. Findings support aftercare that addresses multiple recovery domains rather than a single symptom target.

OBJECTIVE:The authors aimed to assess the proportions of patients at risk for suicide in the Veterans Health Administration (VHA) who were offered and completed the Safety Planning Intervention (SPI). METHODS:Data on SPI completion, sociodemographic characteristics, and settings were extracted from electronic health records (December 2021-February 2023). Suicide risk was identified with the Columbia-Suicide Severity Rating Scale Screener. RESULTS:Among patients with recent suicidal ideation or behavior (N=86,913), 44.2% completed an initial SPI within 30 days of screening positive; an additional 7.2% reviewed a previously completed SPI and made no changes. Of note, 69.4% of veterans were offered SPI (i.e., they completed it within 30 days of screening positive, reviewed a previously completed SPI and made no changes, completed an SPI in the previous year, or declined to complete it). Of patients mandated by VHA policy to complete SPI, 97.9% did so within the required time frame. Patients at higher risk for suicide were more likely to complete SPI, and most completed it in mental health outpatient (52.4%) or inpatient or residential (27.0%) settings. Younger, middle-aged, Black, female, Hispanic or Latine, and urban-residing patients were more likely to complete SPI. CONCLUSIONS:This study is the first to examine SPI delivery in VHA. Relatively high SPI completion rates and higher completion among patients at greater suicide risk indicate widespread adoption of the intervention, particularly in mental health care settings.

The rate of death by suicide is a significant health concern for veterans and is significantly higher than in nonveteran populations, likely due to the relatively higher proportion of veterans using firearms as lethal means of suicide. Family members and other supporters are often highly aware of a veteran’s suicidal thoughts and of their firearm ownership; thus, they may play a role in suicide prevention. This paper describes the development and feasibility testing of a lethal means safety strategy with family members and close friends of veterans. A telephone-based coaching strategy was developed using an educational website, worriedaboutaveteran.org, that presents information about suicide, guidelines for communication, and other information about securing lethal means. This coaching strategy is innovative in its approach of working with a concerned family member or friend without the direct involvement of the veteran, while encouraging collaboration with the veteran around reducing access to lethal means. Family members of 23 veterans at risk for suicide were engaged in a telephone-based coaching clinical project to encourage the use of a lethal means safety strategy to reduce suicide risk. Results suggested the intervention is feasible and that most participants accessed and integrated information from the website with coaching support and roleplay practice of recommended communication methods. These results provide initial support that family members can learn and use the communication strategies and that it can result in a reduction of veterans’ immediate access to lethal means.

Importance:Novel evidence-based suicide-specific treatments are critical to addressing rising rates of suicide among veterans. Objective:To determine whether Project Life Force (PLF), a group intervention that augments suicide safety planning (SSP) with skills training, plus treatment as usual (TAU) delays or decreases suicidal behavior compared with TAU alone. Design, Setting, and Participants:This multisite randomized clinical trial, using intention-to-treat analyses, included veterans at high risk for suicide. Participants were recruited after psychiatric inpatient discharge or referral from 4 Veterans Health Administration outpatient care teams in hospitals in Pennsylvania, New York, and Texas between March 2018 and February 2024. Data were analyzed from April 2024 to October 2025. Intervention:PLF is a manualized, 10- to 12-session (75-90 minutes per session) group intervention combining SSP with distress tolerance, emotion regulation, and interpersonal skills training. Sessions cover each of the 6 SSP steps, health management, reasons for living, SSP accessibility, and a recap, with up to 4 optional booster sessions offered. All participants received TAU, constituting individual SSP and standard outpatient mental health care. Main Outcomes and Measures:The primary outcome was time in days to first suicidal behavior (aborted, interrupted, or actual attempts or suicide death) over 12 months, assessed using the Columbia Suicide Severity Rating Scale and chart review. Secondary outcomes over 12 months included self-report measures of hopelessness, depression, treatment use, attitudes toward help-seeking, and suicide-related coping. Results:A total of 207 participants (mean [SD] age, 46.1 [13.9] years; 178 male [86%]) were assessed at baseline and randomly assigned to PLF plus TAU (101 participants) or TAU alone (106 participants). Cox proportional hazards models showed no significant between-group difference in time to suicidal behavior; however, compared with TAU, PLF participants had a lower hazard of actual suicide attempts over 1 year (HR, 0.49; 95% CI, 0.26-0.92; P = .03). Secondary outcome analyses included data from 132 participants (64%) at posttreatment, 125 (60%) at month 6, and 130 (62%) at month 12. Treatment-by-time effects ranged from Cohen d = -0.03 to 0.35. Notably, PLF significantly improved suicide-related coping at posttreatment (estimate, 4.06; Cohen d, 0.21; P = .04) and month 6 (estimate, 5.36; Cohen d, 0.27; P = .007). No study-related adverse events occurred. Conclusions and Relevance:In this randomized clinical trial of 207 participants, PLF plus TAU did not change time to suicidal behavior but did increase time to actual suicide attempts and improve suicide-related coping, thereby enhancing SSP impact. Trial Registration:ClinicalTrials.gov Identifier: NCT03653637.

OBJECTIVE:Expectations about the future are central to theories of suicide. However, the clinical relevance of an individual's direct expectation of a future suicide attempt is not well understood. This study examined the correlates and prospective significance of expecting a future suicide attempt in a high-risk cohort. METHOD:Veterans at high risk of suicide (N = 126) were tracked for 365 days during a randomized clinical trial. Chi-square tests and Wilcoxon rank-sum tests were used for cross-sectional comparisons and Cox proportional hazards regression was used for prospective prediction of suicidal behavior. RESULTS:Veterans who expected a future suicide attempt had higher rates of preparation for suicide, recent suicidal behavior, and perceived access to a method and opportunity for suicide than those who did not expect one. Groups did not differ in hopelessness, positive and negative expectations toward the future, and perceived suicide-related coping. Veterans who expected a suicide attempt had elevated risk of suicidal behavior. CONCLUSION:Risk factors differentiated Veterans who did and did not expect a future suicide attempt, yet this expectation predicted suicidal behavior independent of these factors. Research on how individuals anticipate their own suicidal behavior may clarify the transition from suicidal thoughts to behavior and facilitate suicide prevention.

A suicide epidemic exists among young U.S. veterans, with risk especially elevated in the first year of transition for the 200,000 servicemembers exiting the military annually. The VA Veteran Sponsorship Initiative (VSI) is a public-private-partnership between federal and community partners that aims to decrease suicides by providing a VA-certified volunteer peer sponsor and connection to community services. Onward Ops is a key community-based national program that enrolls, matches and manages the relationship between servicemembers and sponsors. A prior randomized controlled trial showed that the effectiveness of community interventions can be enhanced when augmented by an Onward Ops sponsor. In preparation for national implementation, we conducted a quasi-experimental, matched-cohort pilot to evaluate the feasibility of an adapted VSI protocol and then assessed effectiveness. The adaptations were executed using the Framework for Reporting Adaptations and Modifications-Enhanced between April 2021 and April 2023. The formative results supported the feasibility of the adaptations to enable proactive enrollment on military installations and expand data infrastructure, partnerships, peer sponsors, and VA clinical services. We then assessed the effectiveness for outcomes not studied in the original VSI trial for active-duty soldiers who enrolled between April and December 2023. After nearest-neighbor matching, the sample included 551 VSI participants and 551 soldiers transitioning as usual. The point-probability contrast or risk differences from the conditional logistic regression model indicated that the VSI caused a statistically significant increase in VA primary care utilization of 0.198 and a statistically significant decrease in suicide attempts of -0.019, both assessed 10 months post-military discharge. The study demonstrated the utility of public-private-partnerships, peer-sponsorship programs and enhanced VA services to support servicemembers during transition.

Background:Approximately 25% of combat veterans with posttraumatic stress disorder (PTSD) seek treatment for traumas involving potentially morally injurious events (PMIEs), which involve acts of commission, omission, or betrayal that deeply transgress one's sense of right and wrong. The sequelae of exposure to PMIEs, called "moral injury," are associated with functional and psychiatric impairment and disrupt veterans' sense of identity and meaning, ability to connect with and trust others, and engender disturbing guilt, shame, rage, and disgust. Currently, no first-line treatments directly address moral injury, and evidence-based treatments for PTSD may be limited because they were derived from civilian contexts, poorly fit the war zone context, and do not allow veterans to discuss the details of the PMIEs with other veterans. Objective:We propose a depth-oriented group psychotherapy for morally injured US combat veterans. Its development is guided by user-centered design principles. Methods:This research plan will use user-centered design methods that continuously gather user experiences during treatment development, with the goal of increased effectiveness and usability. Aim 1 is to discover user needs and preferences as well as treatment-engagement barriers and facilitators from the perspectives of PMIE-impacted veterans and the US Department of Veterans Affairs (VA) trauma clinicians (ie, mental health providers and chaplains) using semistructured qualitative interviews. Aim 2 is to design a treatment manual and refine it using feedback from veterans, VA clinicians and chaplains, and an expert clinical advisory board. Aim 3 is to conduct 2 rapid prototyping open trials (ie, tangibly testing treatment approaches using a prototype manual) with PMIE-impacted veterans (N=12) and iteratively revise the manual based on veteran, provider, and clinical expert panel feedback. We hypothesize that the treatment manual will meet usability, feasibility, learnability, and acceptability criteria. Results:This study was funded with a start date of November 2021. Participant recruitment for Aim 3 pilot trials began in October 2023 and ended in January 2025. We anticipate study data collection and primary data analysis to be completed in January and April 2026, respectively. Conclusions:This project aims to develop and pilot test a treatment manual for a depth-oriented group psychotherapy for moral injury. We anticipate that the manual will meet predetermined usability, feasibility, and acceptability criteria because we used user-centered design methods; if it does not, we are well-poised to revise the manual based on user feedback. Should primary outcomes be met, the next step will be to design and execute a parallel-group randomized controlled trial. We additionally plan to disseminate learnings about combat veteran and VA clinician needs, preferences, and barriers regarding moral injury group psychotherapy, which may inform treatment development efforts beyond this project and study team.

Importance:Suicide is the leading cause of death among active-duty US Army soldiers. Evidence-based preventive interventions exist but need to be targeted to be cost-effective. Objectives:To develop machine learning models using administrative data available during periodic health assessments (PHAs) to determine eligibility for a remote group dialectical behavior therapy-based skills training intervention for soldiers with elevated risk of suicide or nonfatal suicide attempt. Design, Setting, and Participants:From regular US Army soldier PHAs completed from 2015 to 2019, separate 70% training samples were created for suicides and SAs. Each training sample contained all cases and a stratified equal-probability sample of 10 times as many controls, with inverse-probability of selection weights applied to controls. Model performance was evaluated in the remaining 30% test sample. Data analysis was conducted September 2025 through January 2026. Main Outcomes and Measures:Suicides were recorded in the National Death Index for soldiers either still in or out of active service. SAs were recorded in US Army records only for soldiers still in active service. A 24-month risk horizon was used because successive PHAs, though designed to occur annually, are sometimes separated by this much time. A loss to follow-up weight was used in the SA model to adjust for leaving service within 24 months. Results:PHA data were available from 668 684 regular US Army soldiers. A total of 1 932 243 PHAs (including 85.5% by male soldiers, 46.7% by soldiers aged 28 years or older, and 58.1% by married soldiers) completed during the 2015 to 2019 period were included in analyses. Over 24 months, suicide prevalence was 66.7 per 100 000 (SE = 3.6) and SA prevalence was 704.1 per 100 000 (SE = 14.7) adjusted for loss to follow-up. Test sample area under the receiver operating characteristic curve (AUROC) for suicide was 0.72 (SE = 0.17), with Integrated Calibration Index (ICI) of 0.0003 and Brier score of 0.0007. Test sample AUROC for SA was 0.81 (SE = 0.02), with ICI of 0.0014 and Brier score of 0.0070. Suicide prevalence was meaningfully elevated (at least twice that expected by chance) only in the 5% of PHAs with highest predicted risk (sensitivity = 18.6%; SE = 2.0%). SA risk was elevated only in the 10% of PHAs with highest predicted risk (sensitivity = 46.5%; SE = 1.0%). Predicted probabilities of suicide and SA were correlated (Pearson r = 0.29; t = 233.8; P < .001). Overall, 11.9% of PHAs revealed US Army soldiers at elevated risk of either suicide or SA, and 3.1% were at elevated risk of both outcomes. Conclusions and Relevance:Results of this prognostic study suggest that as suicides and SAs have distinct predictors, an attempt to optimize resource allocation will require thoughtful postintervention considerations of costs and benefits of interventions to optimize net benefit.