Association of Alternative Payment and Delivery Models With Outcomes for Mental Health and Substance Use Disorders: A Systematic Review

IMPORTANCE: Health care spending in the United States continues to grow. Mental health and substance use disorders (MH/SUDs) are prevalent and associated with worse health outcomes and higher health care spending; alternative payment and delivery models (APMs) have the potential to facilitate higher quality, integrated, and more cost-effective MH/SUD care. OBJECTIVE: To systematically review and summarize the published literature on populations and MH/SUD conditions examined by APM evaluations and the associations of APMs with MH/SUD outcomes. EVIDENCE REVIEW: A literature search of MEDLINE, PsychInfo, Scopus, and Business Source was conducted from January 1, 1997, to May 17, 2019, for publications examining APMs for MH/SUD services, assessing at least 1 MH/SUD outcome, and having a comparison group. A total of 27 articles met these criteria, and each was classified according to the Health Care Payment Learning and Action Network’s APM framework. Strength of evidence was graded using a modified Oxford Centre for Evidence-Based Medicine framework. FINDINGS: The 27 included articles evaluated 17 APM implementations that spanned 3 Health Care Payment Learning and Action Network categories and 6 subcategories, with no single category predominating the literature. APMs varied with regard to their assessed outcomes, funding sources, target populations, and diagnostic focuses. The APMs were primarily evaluated on their associations with process-of-care measures (15 [88.2%]), followed by utilization (11 [64.7%]), spending (9 [52.9%]), and clinical outcomes (5 [29.4%]). Medicaid and publicly funded SUD programs were most common, with each representing 7 APMs (41.2%). Most APMs focused on adults (11 [64.7%]), while fewer (2 [11.8%]) targeted children or adolescents. More than half of the APMs (9 [52.9%]) targeted populations with SUD, while 4 (23.5%) targeted MH populations, and the rest targeted MH/SUD broadly defined. APMs were most commonly associated with improvements in MH/SUD process-of-care outcomes (12 of 15 [80.0%]), although they were also associated with lower spending (4 of 8 [50.0%]) and utilization (5 of 11 [45.5%]) outcomes, suggesting gains in value from APMs. However, clinical outcomes were rarely measured (5 APMs [29.4%]). A total of 8 APMs (47.1%) assessed for gaming (ie, falsification of outcomes because of APM incentives) and adverse selection, with 1 (12.5%) showing evidence of gaming and 3 (37.5%) showing evidence of adverse selection. Other than those assessing accountable care organizations, few studies included qualitative evaluations. CONCLUSIONS AND RELEVANCE: In this study, APMs were associated with improvements in process-of-care outcomes, reductions in MH/SUD utilization, and decreases in spending. However, these findings cannot fully substitute for assessments of clinical outcomes, which have rarely been evaluated in this context. Additionally, this systematic review identified some noteworthy evidence for gaming and adverse selection, although these outcomes have not always been duly measured or analyzed. Future research is needed to better understand the varied qualitative experiences across APMs, their successful components, and their associations with clinical outcomes among diverse populations and settings.