All-cause mortality following Veterans Affairs and community mental health residential treatment

Abstract: Importance: The US Department of Veterans Affairs (VA) reports high demand for mental health residential treatment, which is partially met by private sector care paid for by VA. Little is known about the clinical quality of community programs relative to VA residential treatment. Objective: To assess whether mental health residential care in VA hospitals provides higher-quality care than non-VA programs by examining posttreatment all-cause mortality in matched veterans. Design, Setting, and Participants: Retrospective administrative data on 26464 veterans discharged from their first mental health residential stay between October 1, 2022, and September 30, 2023, were extracted and propensity score matched using an observational/quasi-experimental case-control design to compare VA residential care with VA-paid community residential care. Main Outcomes and Measures: All-cause mortality outcomes of matched veterans postdischarge by setting, VA residential treatment compared with VA-paid community residential treatment, using propensity score matching. Results: The study sample consisted of matched veterans discharged from VA-paid community residential treatment (n = 7143; median [SD] age, 51.23 [13.29] years; 10.5% female) compared with similar veterans discharged from VA mental health residential treatment (n = 19321; median [SD] age, 53.57 [12.96] years; 9.1% female) on age, emergency department use, and medical and mental health diagnoses in the year prior to admission. Propensity score-weighted Cox proportional hazards analyses found that, if veterans attending VA-paid community residential treatment had instead discharged from VA residential treatment, they would be estimated to have relatively lower postdischarge mortality rates at 9 months (b = -0.22; robust SE = 0.09; exponentiated value of the slope coefficient [Exp (b)] = 0.81; 95% CI, 0.86-0.95; z = -2.51; P =.01) and 12 months (b = -0.32; robust SE = 0.07; Exp [b] = 0.73; 95% CI, 0.63-0.84; z = -4.29; P <.001). Conclusions and Relevance: Case-control analyses indicate that VA mental health residential treatment demonstrates higher clinical quality compared with VA-paid community residential treatment as evidenced by lower posttreatment all-cause mortality in next year. Additional assessment of cost-benefit tradeoffs is needed.