Assessing the Impact of a Blended Payment Model on Client Termination from Substance Use Treatment

Background and Purpose: Blended payment models have been hypothesized as one approach to incentivizing service delivery that reduces premature termination of healthcare. In 2019, Vermont restructured payments for community mental health centers (CMHCs), changing fee-for-service Medicaid reimbursements for mental health services to capitated payments, while retaining fee-for-service payments for outpatient substance use disorder (SUD) treatment and intensive outpatient SUD treatment (IOP). Increasing outpatient mental health service provision at a monthly capitated rate while recouping fee-for-service SUD service payments could increase revenue for CMHCs through SUD program expansion or implementation. A blended payment model may incentivize increased CMHC provision of SUD treatment, improving the continuum of care in areas with highly fragmented treatment.

While over half of Vermont’s CMHCs receive enhanced funding to operate outpatient and IOP SUD treatment programs, many CMHCs do not have SUD programs and share catchment areas with SUD programs not providing mental health services. With more options for SUD care in mental health treatment settings, individuals with SUD may have increased access to programs matching their needs, reducing premature termination from treatment.

This study assesses a blended capitated and fee-for-service payment model’s impact on the likelihood of premature termination of Medicaid recipients from services in Vermont’s SUD treatment system.

Methods: This study employs a difference-in-differences (DID) design with a generalized two-way fixed effects model. DID designs are quasi-experimental designs utilizing pre-post data, and intervention and comparison groups, that control for unmeasured time-invariant variables and secular trends. This DID analysis used Vermont-specific Treatment Episode Data Set-Discharges (TEDS-D) data for 2017 through 2019.

Pre-intervention parallel trends were confirmed visually from 2017 through 2018. The intervention group consisted of adult Medicaid recipient treatment episodes, and the comparison group included adult non-Medicaid recipient episodes. Episodes without payor type were removed from the analysis. Missing data were managed using list-wise deletion. All tests were conducted at a significance level of α = .05.

Results: Study data (N=17602) included 5536 unique outpatient and IOP treatment episodes recorded from 2019 and 12066 unique treatment episodes recorded from 2017 through 2018. All treatment episodes involved individuals diagnosed with SUDs.

The odds ratio of termination from IOP and outpatient programs for Medicaid recipients post-intervention was OR=.720 (p=.010). Receipt of Medicaid (OR=1.19, p=.015) and post-intervention (OR=1.40, p < .001) both predicted increased chances of termination. The results support the hypotheses that the initiation of a blended payment model reduced the likelihood of dropout from IOP and outpatient SUD treatment for Medicaid recipients.

Conclusions and Implications: These results provide evidence that incentivizing co-occurring care through blended payment models may increase availability of co-occurring mental health and SUD services within CMHCs and enable transfers of care from the rest of the SUD treatment system while reducing the likelihood of termination from care. The results show a reduced likelihood of termination from SUD treatment for Medicaid recipients, despite increased odds of termination prior to payment reform and for those utilizing other payors post-payment reform. This suggests a direct impact on care quality and treatment retention resulting from payment reform.