AI system
- System
-
REACH VET (Recovery Engagement and Coordination for Health - Veterans Enhanced Treatment) suicide-risk predictive model and care-enhancement program
- Model
-
The original REACH VET model used 61 variables selected from 381.
- Release
-
The original model introduced in 2017.
- Application developer
-
US Department of Veterans Affairs, Office of Mental Health and Suicide Prevention
- Role
-
The model ranked patients by predicted suicide risk for clinical follow-up.
- Configuration
-
The original 2017 REACH VET model selected 61 variables from 381. The later REACH VET 2.0 announcement concerns a different version.
- Inputs and outputs
-
Not assessed
- Autonomy
-
Not assessed
- Workflow
-
A predictive model run monthly on VHA electronic health records to rank patients and select the top 0.1% risk tier. Facility REACH VET coordinators and providers who acknowledge the list, evaluate care, document safety plans and attempt outreach. A national health system with the data, dashboards and metrics to implement the program at every facility.
- Human oversight
-
Not assessed
- Relation to AI
-
The predictive model selects the monthly top 0.1% risk tier. Coordinators and providers then re-evaluate care and reach out, so the clinical response, not the model alone, produces any change. The evaluation is observational, uses a triple-difference design to handle non-parallel trends, and is authored by the VA office that built and runs REACH VET. It found no mortality association, replicated in 2025 by the same group. Independent reporting documents that the original model weighted being a white man heavily and omitted military sexual trauma.
- AI contribution
-
The predictive model selects the monthly top 0.1% risk tier. Coordinators and providers then re-evaluate care and reach out, so the clinical response, not the model alone, produces any change. The evaluation is observational, uses a triple-difference design to handle non-parallel trends, and is authored by the VA office that built and runs REACH VET. It found no mortality association, replicated in 2025 by the same group. Independent reporting documents that the original model weighted being a white man heavily and omitted military sexual trauma.
Benefits
- Outcome identifiers
-
Not assessed
- Benefit
-
documented suicide attempts (-0.05; -0.06 to -0.03)
- Category
-
Improved Physical Health
- Status
-
Not assessed
- Beneficiary groups
-
People. Organizations. Society.
- Beneficiaries
-
Not assessed
- Reach
-
Not assessed
- Distribution and access
-
Not assessed
- Duration
-
Not assessed
- Expected?
-
Yes
- Costs and harms
-
Not assessed
Measurement
- Studies
-
Not assessed
- Study design
-
Not assessed
- Measurement
-
Not assessed
- Setting and sample
-
Not assessed
- Measures
-
documented suicide attempts (-0.05; -0.06 to -0.03)
- Comparison
-
veterans in the top 0.1% risk tier during the pre-REACH VET period (March 2014 to December 2015) and subthreshold-risk cohorts, in a triple-difference design
- Results
-
Result: suicide attempts (ADT, −0.05; 95% CI, −0.06 to −0.03).
- Additional value from AI
-
Not assessed
- Limitations
-
Not assessed
- Denominators
-
Not assessed
- Materials for replication
-
Not assessed
Money
- Kind of value
-
Not assessed
- Amount
-
Not assessed
- Currency and prices
-
Not assessed
- Period and unit
-
Not assessed
- Realization
-
Not assessed
- Recipients and payers
-
Not assessed
- Gross or net
-
Not assessed
- Included costs
-
Not assessed
- Comparison
-
Not assessed
- Method
-
Not assessed
- Return and payback
-
Not assessed
People and places
- Participants
-
Developer: US Department of Veterans Affairs, Office of Mental Health and Suicide Prevention. Operator: Veterans Health Administration, all facilities.
- Jurisdictions
-
Benefit:
United States. Developer: United States. Operator: United States. - Ownership and control
-
Not assessed
- Dependencies
-
Not assessed
- Cross-border connections
-
Not assessed
- Geopolitical implications
-
Not assessed
Follow-up
- Conditions for replication
-
Providers act on the list: care evaluations rose from 46.1% to 87.5% of identified veterans between March 2017 and December 2018. The model's variables fit the population; the original model under-weighted factors relevant to women veterans, per Nextgov. Outcome ascertainment is comparable across periods; the authors note that closer attention may increase recorded attempts.
- Replication attempts
-
2 follow-up entries. Outcome unresolved: 2.
- Later findings
-
17 July 2025. Nextgov reported in July 2025 that REACH VET still scans records monthly with the 61-variable model and that VA is updating it to a 2.0 model; a VA cohort study published 8 July 2025 found no association between REACH VET inclusion and suicide or all-cause mortality, in line with the 2021 evaluation's mortality subanalyses.
- Availability
-
Not assessed
- Institutional decision
-
Not assessed
- Grounds
-
Not assessed
- Action and result
-
Not assessed
- Success criterion
-
Not assessed
Evidence
- Source numbers
-
AIBD-0310-source-0001–0006
- Citations
-
Numbered references below.
- Language
-
Not assessed
- Source relationships
-
Developer, Independent
- Original evidence
-
Not assessed
- Access and versions
-
Not assessed
- Supported findings
-
Not assessed
- Evidence links
-
Not assessed
- Evidence strength
-
Not assessed
- Missing evidence
-
Not assessed
Grades
- Assessed benefits
-
People. Organizations. Society.
- Depth
-
Not assessed
- Reasons
-
People: The 2021 study reports population-level reductions in documented suicide attempts, admissions and ED visits for identified veterans. No individual veteran's outcome is documented, and mortality did not differ in 2021 or 2025. Organizations: Care evaluations rose from 46.1% to 87.5% of identified veterans and 98% were contacted by October 2020, per VA-authored papers. No independent audit was found. Society: Nextgov notes the top 0.1% tier covers only 2–3% of VA suicides and the model under-identified women; no population suicide-rate effect is shown.
- Observed or intended
-
Not assessed
- Rubric
-
Not assessed
- Assessor and date
-
Not assessed
- Alternative assessments
-
Not assessed
- Fill
-
Independently reported
Entry details
- Kind
-
Not assessed
- Related entries
-
Not assessed
- Field
-
Suicide prevention / mental health care. Health-system predictive risk modeling and outreach.
- Purpose
-
Built to identify veterans in the top 0.1% of predicted suicide risk each month so that clinicians re-evaluate and enhance their care. The measured associations are that purpose.
- Dates
-
Date basis unassessed: 2021 (precision unassessed). Updated: 21 October 2026.
- Stage
-
Not assessed
- Version
-
SHA-256: a48d242a2265…
References
- Other identifiers
-
Not assessed
- Category mappings
-
Not assessed
- Brief
-
Not assessed
- Revisions
-
6 revisions. Latest: 21 October 2026.
Sources
-
Peer-reviewed study Developer
McCarthy JF, Cooper SA, Dent KR, Eagan AE, Matarazzo BB, Hannemann CM, Reger MA, Landes SJ, Trafton JA, Schoenbaum M, Katz IR (2021). Evaluation of the Recovery Engagement and Coordination for Health-Veterans Enhanced Treatment Suicide Risk Modeling Clinical Program in the Veterans Health Administration. JAMA Network Open 4(10):e2129900 (PMC8524305). Authors at the VA office that runs the program.
REACH VET implementation was associated with greater treatment engagement and new safety plan documentation and fewer mental health admissions, emergency department visits, and suicide attempts.
-
Peer-reviewed study Developer
Dent KR, Cooper S, McCarthy JF (2025). The REACH VET Program and Mortality Outcomes Among Veterans at High Risk of Suicide. JAMA Network Open, 8 July 2025 (PMC12238888). VA-authored.
inclusion in the REACH VET program was not associated with suicide mortality (6-month hazards ratio [HR], 1.18 [95% CI, 0.83-1.69]
-
Peer-reviewed study Developer
Matarazzo BB, Eagan A, Landes SJ, et al. (2022/2023). The Veterans Health Administration REACH VET Program: Suicide Predictive Modeling in Practice. Psychiatric Services (first published 30 August 2022). VA-authored.
By October 2020, 98% of veterans identified (N=6,579) were contacted by providers and had their care evaluated.
-
News report Independent author
Graham E (2025). Inside VA's yearslong AI effort to uncover veterans at high risk of suicide. Nextgov/FCW, 17 July 2025.
It's going to go something like, 'you've been identified as being at elevated risk for suicide at the present time; we'd like to offer you resources.'
-
News report Independent author
Graham E (2024). VA is updating its AI suicide risk model to reach more women. Nextgov/FCW, 18 October 2024.
considered being a white man more of an indicator of potential self-harm than other factors
-
Press release Developer
US Department of Veterans Affairs (2017). VA REACH VET Initiative Helps Save Veterans Lives: Program Signals When More Help Is Needed for At-risk Veterans. Press release, 3 April 2017.
Using a new predictive model, REACH VET analyzes existing data from Veterans' health records to identify those at a statistically elevated risk for suicide, hospitalization, illness or other adverse outcomes.