State governments that are beginning to marry social determinant of health efforts with value-based payment programs in Medicaid accountable care organizations are encountering a number of challenges. A new report from the Milbank Memorial Fund highlights some of these issues, which include implementing processes for integrating social risk screening and referral into practice, achieving provider buy-in, calculating provider risk when incorporating social determinants, and measuring outcomes.
The Milbank report is called “Marrying Value-Based Payment and the Social Determinants of Health Through Medicaid ACOs: Implications for Policy and Practice.” Healthcare Innovation recently interviewed co-author Elizabeth Tobin-Tyler, J.D., M.A., an associate professor of family medicine and medical science at the Warren Alpert Medical School, and of health services, policy, and practice at the School of Public Health at Brown University. Her book, Essentials of Health Justice (Jones and Bartlett, 2019), examines the legal, structural, and justice issues underlying health disparities.
HCI: Your report looks at the marriage of social determinants of health and value-based care. Are these both new areas in the Medicaid space that are now intersecting? Are there pioneering efforts in many states involving Medicaid waivers and each taking a slightly different approach?
Tobin-Tyler: This is a relatively new approach for Medicaid programs. Medicaid managed care programs have been experimenting for some time with how to pay for different types of services to address social determinants of health, and I think the ACO movement in Medicare has generated a lot of interest in states around how to structure value-based payment models in the Medicaid space. Because of the population served in Medicaid programs, marrying these two things is something that makes a lot of sense— how can we use value-based payment structures to facilitate integration of social services and address social determinants of health? Can this facilitate reduction of Medicaid costs and help improve health outcomes?
Is each state having to figure out how to go about this on their own? There is not a uniform model they can all deploy, right?
There were some states that were early adopters of the ACO model that were experimenting with new ways to deliver services and rein in costs. As there has been more attention to social determinants of health and more focus on value-based payment to incentivize delivery reform, different states have been trying this out. There is no one-size-fits-all, which makes it distinct from Medicare ACOs, which are guided by federal rules. The Center for Health Care Strategies has been providing technical assistance to states to help them share learning across states about how to do this well. But states are doing this in a variety of ways. What we try to capture in our report is the particular ways states are leveraging requirements within their ACO structures to incentivize or mandate that their ACO participants screen for social determinants of health and build social services partnerships.
What kinds of data are Medicaid ACOs seeking to gather so far? Are they looking at things like number of people being screened and services prescribed? Or trying to assess financial impact like emergency department usage or rehospitalization reductions? Can they get to metrics that would tell whether patients are experiencing better health – or is that very hard to do?
That is very hard to do. Demonstrating better health outcomes as a result of these new requirements and guidelines is the holy grail for measuring outcomes! Initially, as providers start to screen and build social service partnerships the primary measurement seems to be through process metrics: how many people have been screened and how many partnerships have been developed. The next step is to demonstrate changes in healthcare utilization. For instance, if, through this process, the ACO, through its social service partnerships are reducing substandard housing conditions that their patients are living with, is that going to reduce the number of ER visits for asthma attacks in the population served? Ideally, the ACO could then move even more upstream to look longitudinally at their population to measure specific health outcomes. I don’t think anybody is there yet.
Your report asks how screening requirements and expectations for referral and follow-up could be structured so as not to overburden already stressed providers. Did you find some good examples of the kinds of supports some programs have put in place that have worked well?
I think this is a question providers in the ACOs are struggling with, and I see it from my own work here in Rhode Island. Providers are being asked to screen patients for social needs, but this raises many questions, such as what type of screening protocol should they use? Some states are mandating a particular screening protocol, while others are more flexible. At the practice level, who is going to ask the screening questions? How is the practice going to use the information taken from the screening? I think those are the kind of conversations happening at the provider level once the state has laid out the requirements for participation in an ACO. In some states, there are conversations occurring across ACOs to share best practices for screening and referral.
Your report notes that asking providers to take on risk for patients with multiple unmet social needs can be a hard sell. How have states sought to avoid unfairly penalizing providers who take on the highest-need, highest-cost patients? I think your report mentioned Massachusetts as a good example. What did they do?
They began experimenting with social risk adjustment within their Medicaid managed care programs even before they implemented their ACO program. They created a model for looking at social risk factors which included a neighborhood index for risk as well as an assessment of a patient’s housing instability. They have been a pioneer in social risk adjustment. It can be a concern for providers who decide to participate in an ACO that if they serve a very high-risk population, their costs will be higher and they will be less likely to see shared savings. It was a concern when Medicare ACOs began as well --- will there be providers who will be reluctant to take on more high-risk patients? In the Medicaid population, it is even more critical to address the issue since patients are low-income and likely to have many social needs.
Another issue you raise is that if these programs are successful, savings may accrue to other sectors outside of healthcare such as criminal justice and child welfare systems. Does this have implications for who should be involved in the projects and who should be sharing data?
Absolutely. ACOs are built on a model of saving the healthcare system money by incentivizing better, more efficient health care delivery. But if we really want to capture the benefits of having a more integrated, holistic approach to health and well-being, we should be looking across systems. Some states have begun to do that. I know in Rhode Island the Executive Office of Health and Human Services has begun to connect child welfare data and Medicaid data at the state level to try to track outcomes of efforts to better coordinate health care and social services.
I think your report also mentioned Minnesota was working on something similar.
That’s right. Minnesota has done a particularly good job of using health disparities data to risk-adjust in their ACOs but also to look at system-wide outcomes.
Can you explain the distinction between “health-related social needs,” which are individual patient needs, and social determinants of health, which are upstream, structural factors that cannot be addressed at the individual patient level? What are some implications for policymakers of understanding that distinction in terms of where they make investments?
Public health officials have been saying for some time that health care only accounts for 10 to 20 percent of a person’s health and what really influences a person’s health is social factors. This has changed the discussion in the health care sector about its responsibility for addressing social factors. Most now recognize that you can’t provide good healthcare if you are not acknowledging the social factors affecting a patient’s health. This has led to investments being made in healthcare to address social needs, at the same time that we are not addressing the upstream factors that are creating those social needs in the first place. So the money ends up going into the healthcare system downstream to address problems that could be better addressed upstream. Many public health experts have described this problem. For example, if we provided safe affordable housing in the first place, we wouldn’t have to address health problems that are associated with homelessness or substandard housing conditions. This is a fundamental question for policymakers in terms of where they make investments.
Your report mentions the phrase “Bridge to Nowhere,” which refers to the fact that as ACOs have begun to implement these interventions, they have had to confront the realities of a fragmented, underfunded social service system that often cannot meet demand.
Yes. One of the worries I have about this whole process of integrating social needs screening and developing health care--community partnerships, which I think is a good thing, is that there is not a lot of funding for and coordination with the social services sector, which is often very fragmented and underfunded. Asking social service organizations to simply take on more referrals from health care may just add to an already overburden system. Without careful planning and coordination, simply referring people to services will not ensure access. Some states are building pretty sophisticated database systems, which can be accessed by both health care and social service providers to support closed loop referrals (communication about outcomes to the person making the referral) and data tracking to see where there may be gaps in the system. North Carolina is one state that is making significant investments in this area.
To me, this is critical. One potentially positive outcome of this type of partnership would be to identify significant gaps in the safety net and to use that data to inform policy.
Could state government budget shortfalls caused by the pandemic have an impact on states’ willingness or ability to do this work? I saw that California delayed an ambitious Medicaid innovation package that expanded whole person care pilots.
I think that is a very real concern. We are going to see not only state budgets being slashed, making investment in these types of innovations more difficult, but also medical practices closing. One potentially positive outcome for the healthcare system may be that this will reinforce the need to move away from fee-for-service payment. It may also highlight the clear relationship between social needs and health and the need for upstream investments in the safety net.