# Risk Stratified Enhancements to Clinical Care: Targeting Care for Patients Identified Through Predictive Modeling as Being at High Risk for Suicide, with the Office of Mental Health Operations

> **NIH VA I01** · CENTRAL ARKANSAS VETERANS HLTHCARE SYS · 2021 · —

## Abstract

Despite work done to strengthen Veterans Health Administration (VHA) mental health services and suicide
prevention, suicide rates in VHA have been stable [1]. These rates stand in contrast to increasing rates in other
Americans, especially middle-aged men [2,3] and in Veterans who do not utilize VHA services [4,5], suggesting
that VHA programs may have mitigated expected increases although this has not been determined.
Nevertheless, the finding that suicide rates in VHA remain high represents a strong call for action. Novel
approaches that reduce the incidence of suicide-related events are needed earlier, ideally before suicide-
related behaviors occur.
One innovative approach recently validated in VHA is predictive modeling that identifies Veterans at risk and
thus facilitates implementation of targeted prevention. The VHA predictive model has identified the top 5% of
VHA patients who were at the highest predicted risk for suicide. This model provides new information about
who is at risk; fewer than 2% of the 5% of patients identified as high risk received clinical flags for being at risk.
For those identified as high risk, VHA’s Office of Suicide Prevention has implemented a national suicide
prevention outreach program entitled Recovery Engagement and Coordination for Health – Veterans
Enhanced Treatment (REACH VET). REACH VET utilizes a dashboard to provide the names of patients
identified by the model to coordinators at each VA medical facility. REACH VET coordinators are responsible
for notifying providers of the patient’s status and prompting providers to re-evaluate care and take any
appropriate steps if they are not already occurring (e.g., contacting the patient to re-engage in care).
To further strengthen REACH VET, an effective, low-cost suicide prevention intervention, caring letters, is
being added as a REACH VET augmentation. Caring letters involves the sending of recurring brief notes to
high risk patients expressing care, concern, and offers of help if needed. It is one of the only suicide prevention
interventions that have reduced suicide mortality rates in a randomized controlled trial [8–12].
Specific Aim 1: Evaluate the impact of virtual external facilitation versus standard implementation by
conducting a formative evaluation to identify barriers and facilitators to implementation to define and refine
virtual external facilitation strategies and a summative evaluation of virtual external facilitation versus standard
implementation.
Specific Aim 2: Develop and evaluate the augmentation of REACH VET using caring letters, an evidence-
based suicide prevention intervention by conducting a formative evaluation of the augmentation of REACH
VET with caring letters and refining the caring letter intervention for scale up to the VHA-wide REACH VET
program.
The overall design will be a hybrid effectiveness-implementation controlled program evaluation using a mixed
methods approach. REACH VET is currently being implemented in VHA using a...

## Key facts

- **NIH application ID:** 9981443
- **Project number:** 5I01HX002403-03
- **Recipient organization:** CENTRAL ARKANSAS VETERANS HLTHCARE SYS
- **Principal Investigator:** Sara J Landes
- **Activity code:** I01 (R01, R21, SBIR, etc.)
- **Funding institute:** VA
- **Fiscal year:** 2021
- **Award amount:** —
- **Award type:** 5
- **Project period:** 2017-10-01 → 2021-03-30

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/9981443

## Citation

> US National Institutes of Health, RePORTER application 9981443, Risk Stratified Enhancements to Clinical Care: Targeting Care for Patients Identified Through Predictive Modeling as Being at High Risk for Suicide, with the Office of Mental Health Operations (5I01HX002403-03). Retrieved via AI Analytics 2026-05-24 from https://api.ai-analytics.org/grant/nih/9981443. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*