Behavioral Health Integration With Primary Care: Implementation Experience and Impacts From the State Innovation Model Round 1 States

Policy Points:

• Individuals with behavioral health (BH) conditions comprise a medically complex population with high costs and high health care needs. Considering national shortages of BH providers, primary care providers serve a critical role in identifying and treating BH conditions and making referrals to BH providers.
• States are increasingly seeking ways to address BH conditions among their residents. States funded by the Centers for Medicare and Medicaid Services under the first round of the State Innovation Models (SIM) Initiative all invested in BH integration. States found sharing data among providers, bridging professional divides, and overcoming BH provider shortages were key barriers.
• Nonetheless, states made significant strides in integrating BH care. Beyond payment models, a key catalyst for change was facilitating informal relationships between BH providers and primary care physicians. Infrastructure investments such as promoting data sharing by connecting BH providers to a health information exchange and providing tailored technical assistance for both BH and primary care providers were also important in improving integration of BH care.

Context: Increasing numbers of states are looking for ways to address behavioral health (BH) conditions among their residents. The first round of the State Innovation Models (SIM) Initiative provided financial and technical support to six states since 2013 to test the ability of state governments to lead health care system transformation. All six SIM states invested in integration of BH and primary care services. This study summarizes states’ progress, challenges, and lessons learned on BH integration. Additionally, the study reports impacts on expenditure, utilization, and quality-of-care outcomes for persons with BH conditions across four SIM states.

Methods: We use a mixed-methods design, drawing on focus groups and key informant interviews to reach conclusions on implementation and quantitative analysis using Medicaid claims data to assess impact. For three Medicaid accountable care organization (ACO) models funded under SIM, we used a difference-in-differences regression model to compare outcomes for model participants with BH conditions and an in-state comparison group before-and after model implementation. For the behavioral health home (BHH) model in Maine, we used a pre-post design to assess how outcomes for model participants changed over time.

Findings: Informal relationship building, tailored technical assistance, and the promotion of data sharing were key factors in making progress. After three years of implementation, the growth in total expenditures was less than the comparison group by $128 (−$253, −$3; p < 0.10) and $62 (−$87, −$36; p < 0.001) per beneficiary per month for beneficiaries with BH conditions attributed to an ACO in Minnesota and Vermont, respectively. Likewise, there were reductions in emergency department use for ACO participants in all three states after two to four years of implementation. However, there was no improvement in BH-related quality metrics for ACO beneficiaries in all three states. Although participants in the BHH model had increased expenditures after two years of implementation, use of primary care and specialty care services increased by 3% and 8%, respectively, and antidepressant medication adherence also improved.

Conclusions: The SIM Round 1 states made considerable progress in integrating BH and primary care services, and there were promising findings for all models. Taken together, there is some evidence that Medicaid payment models can improve patterns of care for beneficiaries with BH conditions.

Keywords: behavioral medicine/organization and administration, delivery of health care/organization and administration, delivery of health care/integrated, health policy.