Objective:

To assess how suicide rates and hazards after military separation differed among Army veterans according to use and timing of Veterans Health Administration (VHA) services.

Approach:

• Design and cohort: Cohort study of 649,811 Army soldiers who returned from Afghanistan or Iraq deployments during fiscal years 2008–2014 and separated from service by September 2017.
• Follow-up and outcome: Suicide deaths were identified using national mortality records and ICD-10 codes X60–X84 and Y87.0. Follow-up ran from October 1, 2007, through December 31, 2018, with up to 11.3 years of observation.
• Analysis: Researchers estimated age-adjusted annual suicide mortality rates and time-dependent hazards by VHA use, timing of first care, and type of service, including outpatient and inpatient care.

Key Findings:

• There were 1,850 suicide deaths (0.28%) in the cohort.
• The age-adjusted average annual suicide rate was 28.77 per 100,000 person-years among VHA users and 35.92 among nonusers; the rate ratio was 0.80 (95% CI, 0.68–0.94; P = .007).
• Time-dependent hazard rates were higher among nonusers during the first 2 years after separation, showed no difference in year 3, were higher among users in year 4, and showed no subsequent difference.
• Among VHA users, hazard-rate point estimates were higher for those initiating care within 12 months than for those whose first use was later in several periods. First-year VHA inpatient care was associated with markedly higher point estimates, including during years 5–11. The estimate for delayed care was based on fewer than 16 events.

Interpretation:

Age-adjusted suicide rates varied by VHA service use and were lower among veterans using VHA services closer to separation. Rates nevertheless remained high among VHA users. Interpretation is constrained by selection bias, suicide misclassification, less complete clinical information for nonusers, and residual confounding.

Limitations:

• Potential selection bias and suicide misclassification.
• Clinical information was less complete for nonusers.
• Residual confounding may relate to sex, race and ethnicity, baseline clinical risk, prior suicidal behavior, and indication for VHA care.
• The cohort is historical, and care patterns and suicide-prevention initiatives have changed.
• Some estimates for delayed care were based on fewer than 16 events and warrant cautious interpretation.

Conclusion:

The authors report that suicide risk after Army service differed over time by VHA service use. They state that lower suicide rates among veterans receiving VHA services during the years immediately after separation support enhanced suicide-prevention care delivered by the VHA.