“Provider Payment Reform: Right Course, Wrong Students?,” a recent blog post authored by Chris Koller, President of the Milbank Memorial Fund, argues that while delivering high-value care to vulnerable populations is the “right coursework” for the US health care system, the “courses” (value-based payment (VBP) programs) should be targeted toward primary care practices, as opposed to specialists and hospitals. The blog post draws on recent research on the Medicare Shared Savings Program (MSSP), which found that accountable care organizations (ACOs) led by primary care providers (PCPs) or with officially designated patient-centered medical homes (PCMHs), tend to perform better on cost and quality than those without these characteristics. Koller concludes, “Medicare and commercial payer ACOs and other provider accountability efforts should focus on enhancing the roles, expectations, and centrality of primary care teams, perhaps starting with requiring that governance and leadership of the ACOs come from the primary care community.”
Koller’s blog post presented a thoughtful discussion of Medicare ACO efforts and the evidence on the effect of primary care, yet was largely silent on Medicaid ACOs. This is not a surprise, because there is not as much tangible proof available on how Medicaid ACOs perform, their characteristics, or what attributes correlate with success. While it is logical to assume a strong primary care base would help Medicaid ACOs similar to how it helps MSSP ACOs, there remain many questions about what makes Medicaid ACOs successful.
Medicaid is often described as a 50-state “Learning Lab” that can experiment with and test innovative approaches. To truly test these experiments, gaining access to basic information at both the program- and provider-level is instrumental, not only for ensuring that Medicaid agencies and other interested stakeholders are teaching these “ACO courses to the right students,” but also that they are learning key lessons along the way.
How are Medicaid ACOs performing?
Discovering the characteristics of successful ACOs requires tracking the performance of individual ACOs. The Centers for Medicare & Medicaid Services (CMS) publicly reports annual quality and cost data for MSSP ACOs, at both the aggregate- and the ACO-level. While researchers do not always reach consensus on findings, the MSSP program can be evaluated due to its publicly available data, large number of participants, and standardization. Most states report Medicaid ACO performance on an aggregate level, e.g., through press releases or annual reports, and in many cases, results have been impressive. Performance of individual Medicaid ACOs on both cost and quality, however, is rarely reported publicly. One exception is Oregon, which publishes quality performance data for each of its 16 CCOs. Even rarer are formal evaluations of Medicaid ACO programs. While MSSP ACOs tend to have similar design elements (governance, quality metrics, payment models) that are more easily comparable, state Medicaid ACO programs vary widely, which makes comparison difficult not only across ACOs in different states, but also among ACOs within a particular state.
How are Medicaid ACOS distributing savings?
Information on how MSSP ACOs distribute earned savings can provide insight into what incentives work to motivate providers to reduce costs and improve quality. MSSP ACOs are required to develop a public-reporting webpage to outline the proportion of shared savings: (1) invested in ACO infrastructure; (2) invested in redesigned care processes/resources; and (3) distributed to ACO participants. Researchers found that of the ACOs reporting allocations, the average amount of savings allocated to PCPs was 49 percent, but with a very wide range of 6.25 percent to 100 percent. Of note, ACOs planning to give at least 50 percent of savings to their PCPs and specialists were more likely to have generated savings. Although there may be an exception, in general, states do not require Medicaid ACOs to publicly report such information, thus it is not readily clear which Medicaid ACO programs are financially incentivizing primary care.
What is the relationship between advanced primary care and Medicaid ACOs?
Advanced primary care can be a key contributor to successful ACOs. However, this relationship can also work in reverse, as ACOs can help drive advanced primary care. Medicaid ACOs may have certain structural advantages over their MSSP counterparts in terms of prioritizing primary care, thanks to flexibility afforded to state Medicaid agencies in designing these programs. For example, some states require their Medicaid ACOs to fulfill primary care governance requirements, similar to how Rhode Island requires its Accountable Entities to demonstrate primary care capacity. Oregon requires its Coordinated Care Organizations (CCOs) to spend at least 12 percent of total medical expenditures on primary care by 2023, and offers CCO quality payments contingent upon the percentage of members enrolled in Patient Centered Primary Care Homes. Several pioneering states also require or incentivize Medicaid ACOs to partner with community-based organizations to address social determinants of health, recognizing that paying primary care teams to provide better care inside their four walls is not enough.
As states (re)define expectations for Medicaid ACOs, it may also be helpful to better understand which organizations — including primary care systems, federally qualified health centers, specialists, hospitals, care coordinators, and community-based organizations — participate in each ACO, and to articulate how ACOs are expected to impact delivery of primary care and utilization of health care services more broadly. An analysis of the move to CCOs in Oregon, for example, found a reduction in primary care visits and speculated that shifting to nontraditional supports may have been one of the reasons for this decline.
How do different risk adjustment strategies impact Medicaid ACO programs?
Most ACO programs use some form of risk adjustment to account for differences in health and demographics of attributed populations. For example, Medicare uses a single risk-adjustment approach — Hierarchical Condition Category (HCC) scores — to adjust cost benchmarks based on differences in the clinical complexity of populations attributed to MSSP ACOs. While HCC has its downsides, it is a transparent and standardized approach. For Medicaid ACO programs, however, the type and rigor of risk adjustment varies broadly. Massachusetts Medicaid and the University of Massachusetts, for example, developed a risk adjustment methodology that uses SDOH data. Risk adjustment methodologies incorporating social factors may need to be more widely adopted across Medicaid ACOs, given the high socio-economic disadvantages of the Medicaid population, as well as research demonstrating that inadequate risk adjustment in VBP programs can actually exacerbate health disparities. Better understanding risk adjustment approaches used in Medicaid ACOs is essential to uncover how these programs are performing, especially if attempting to draw comparisons across states.
Learning more about Medicaid ACOs
Much of what CHCS knows about Medicaid ACO programs to date has been tracked and published as part of CHCS’ Medicaid ACO Learning Collaborative, a nearly six-year effort supported by The Commonwealth Fund. While there is clearly much more to learn about the performance of Medicaid ACOs, the information available suggests that many Medicaid ACOs are performing well. For this reason, there is a great opportunity to study what is and is not working to improve quality and lower costs in Medicaid ACOs, especially through the lens of primary care. This type of research could be approached in various ways, such as informal collaboration among researchers conducting evaluations of Medicaid ACOs or informal affinity groups where states come together to review performance on common metrics and share feedback on why Medicaid ACOs may or may not be doing well. Findings from this type of work may be useful not only to improve the performance of Medicaid ACOs, but also to glean insights for Medicare or commercial ACOs. To make sure the experiments in the “Medicaid Learning Lab” produce actionable information about which innovative payment and delivery system reforms are most effective, more and better data is needed.