Introduction
In the most recent Advanced Notice for Medicare Advantage (MA) payments, the Centers for Medicare & Medicaid Services (CMS) announced it is soliciting comments on whether enhancements can be made to the MA risk adjustment model to address the impacts of social determinants of health (SDoH) on MA enrollees’ health status. CMS is particularly interested in comments on incorporating additional factors that adjust for SDoH into the risk adjustment model that predicts the relative costs of MA enrollees. Below we address key questions related to SDoH and the implications of including such factors in the MA risk adjustment model.
What are social determinants of health?
Social determinants of health (SDoH) are the conditions in the environment where people are born, live, learn, work, play, worship, and age that affect a wide range of health and other quality-of-life outcomes. Specific SDoH factors can be at the individual—e.g., educational attainment, employment, and income—or community levels—e.g., access to food markets, public parks, and convenient public transportation.
How do SDoH factors affect health outcomes and health care costs?
The link between SDoH and health outcomes is complex, often working through multiple mechanisms and involving an interplay between factors. For example, higher education has been associated with better social, psychological, and interpersonal skills that improve people’s ability to cope with stress, recognize symptoms of ill health in a timely manner, and seek appropriate medical help. Education can also be interwoven with income as people with higher education often get higher paying jobs, which can lead to better access to health care.
Complicating things further is that SDoH factors operate in the present, but health outcome effects can accumulate over a person’s lifetime in a variety of ways, including epigenetic effects. For example, racial discrimination can accumulate in the form of excessive biological stress levels and lead to a lack of social confidence or trust when interacting with the health care system.
The link between SDoH and health care costs is also not straightforward. Cost and resource use could be less for socially disadvantaged patients because of an inability to access and use health care services, or more because of higher severity due to lack of preventive and early diagnostic services. However, it can generally be expected that to achieve the same level of health, it is more costly/resource-intensive to manage the care of socially disadvantaged groups relative to advantaged groups. When it comes to disadvantaged groups, besides physical or mental ailments, other non-clinical issues (e.g., poor nutrition or language barriers) may need to be overcome which require additional resources. This leads to our next question.
Why would it be important to account for SDoH in health care payments?
As insinuated above, the main argument in favor of accounting for SDoH is the premise that achieving favorable outcomes for patients with greater social risk is more difficult or requires greater resources to achieve the same level of outcomes in a more socially advantaged population. Moreover, the cost of implementing social service interventions, such as nutrition, transportation, or housing support, are generally not reimbursable under most current health care payment systems. In fact, the predominant payment system within Medicare is the fee-for-service (FFS) payment system, which uses provider reimbursement fee schedules that exclusively pay for clinical services. In contrast, capitated payment structures or even some value-based payment (VBP) structures (e.g., shared savings or bundled payments) can provide much more flexibility to fund a wider array of services, including social services, as such payments are not linked to specific services. This raises three important points related to the MA program:
- Given the more flexible payment system within the MA program, it represents an important opportunity for CMS to consider how to better support the SDoH needs of the Medicare population. This is more so the case considering that over 40% of the Medicare population is enrolled in an MA plan, with a growing proportion of enrollment being in special needs plans (SNPs) targeted to low-income Medicare beneficiaries.
- Since MA payments are based on FFS payment levels in the Traditional Medicare program (which generally do not cover social services), this creates a challenge for MA plans—particularly those that have a larger than average proportion of social disadvantaged patients—to address social issues without additional funding.
- Moreover, since MA plans are paid on a capitated basis, and there has been evidence showing that SDoH have a sizeable impact on people’s health—with some research suggesting up to a 6x greater impact than medical care—MA plans have an incentive to address their members’ SDoH to reduce unnecessary health care utilization and contain costs.
Risk adjusting Performance Measures for SDoH can lead to negative unintended consequences. Would there be similar vulnerabilities making SDoH adjustments in the payment system?
There are multiple places to potentially include SDoH factors for risk adjustment, such as performance measurement or payment.
Performance measurement. In performance measurement, such as with the Hospital Compare or MA Star Rating programs, measures are used to compare the performance of providers and payers. There has been controversy about whether to risk-adjust these measures for SDoH or not, especially when plans or providers can get financially penalized or rewarded based on their relative performance. On the one hand, as described above, since socially disadvantaged patients have worse health outcomes, risk-adjusting the performance measures for SDoH helps ensure that plans or providers that serve a large proportion of socially disadvantaged patients are not penalized for factors outside of their control. Furthermore, the plans and providers that go above and beyond to address SDoH needs are eligible to be rewarded for their efforts (potentially offsetting intervention costs).
On the other hand, including SDoH factors as an adjustment to performance measures could be viewed as setting lower standards of quality for plans and providers with more disadvantaged patient populations; perhaps even obscuring poor-quality care being provided to socially disadvantaged groups. For example, a growing body of literature documents how implicit racial biases by providers are associated with less accurate diagnoses, curtailed treatment options, less pain management, and worse clinical outcomes, such as increased severe maternal morbidity, for minority patients. Hence, adjusting for SDoH in performance measurement would make it harder to identify providers or plans that need to improve their quality of care.
Payment Adjustment. Adjusting for SDoH directly in the payment model as in the MA risk adjustment model, avoids the potential unintended adverse effect of lowering the standards of care for vulnerable populations. This is because MA risk adjustment is not used for determining the relative performance of plans. Rather, it is used to ensure plans have sufficient funds to manage the health care needs of their enrollees. If the SDoH factors were similarly included in the model as the currently included demographic and clinical factors, all MA plans would have equivalent access to additional funds to address SDoH needs of their membership.
What type of social risk factors should be used?
A key operational issue involved in controlling for SDoH in risk adjustment is making sure that appropriate measures are being used in the model. The measures should be feasible to collect, be predictive of costs, and provide a consistent depiction of SDoH needs across plans.
In the advanced notice, CMS explicitly requests comments on what types of SDoH factors could be included in the MA risk adjustment model. One option they note is the use of geographic data based on the residence of plan enrollees, such as ZIP codes. Studies have shown geographic and other neighborhood-level indicators of SDoH can be useful predicters of health outcomes and costs at the individual level. For example, the Massachusetts Medicaid Managed Care Model utilizes a neighborhood stress score to risk-adjust payments based on area-level metrics such as education, employment, and income levels. The algorithm resulted in slight improvements over a diagnosis-based model in predicting costs, explaining most of the spending variation in the managed care population and reducing underpayments for several vulnerable populations. A similar area-based metric has already been created for MA plans, known as the Health Equity Summary Score, but has yet to be put into practice.
Geographic-level factors have the benefit of not requiring the collection of additional data from patients or other sources, such as social service organizations. Any data exchange required between health and social service sectors would likely create legal and regulatory issues as they may be governed by different privacy laws. MA plans could also aim to collect individual data from enrollees through screening tools, which could be administrated during annual assessments. Many MA plans already have such screening tools, but there would need to be some sort of standardization if collected data were to be used for risk adjustment.
While there may be logistical challenges in collecting individual data, it is possible to use individual level variables to adjust for SDoH. In fact, the MA risk adjustment model currently incorporates eligibility status for Medicaid in the risk adjustment model, which helps identify lower income MA enrollees. Although dual-eligibility status has been found to be an important indicator of resource-use in the MA program, there is mixed evidence on how effective including social risk factors is in improving the predictive accuracy of risk-adjusted models. For example, a study including factors for homelessness, prior substance use, and other social risk factors did not better predict hospital utilization relative to a diagnosis-based model. More research is needed on using social risk factors, particularly with individual-level data, for predicting costs at the individual level.
How do MA plans address SDoH now?
MA plans have historically provided supplemental benefits to members that are not covered under traditional Medicare, such as dental benefits or gym memberships. These benefits were required to be primarily related to health and offered to all members. However, the Creating High-Quality Results and Outcomes Necessary to Improve Chronic (CHRONIC) Care Act of 2017, now allows MA plans greater flexibility to offer SDoH supplemental benefits that Traditional Medicare cannot pay for, including meal delivery, cooking classes, home modifications to assist with mobility, pest control, indoor air quality equipment, and others. The CHRONIC Care Act also provides flexibility for supplemental benefits to be targeted to certain enrollees that would benefit from them. Given that these are voluntary benefits, there is variation in how (and if) plans are implementing them.
It should also be noted that the CHRONIC Care Act does not provide additional funding for the supplemental benefits offered. Hence, the plans take on the risk for any additional costs incurred for the additional benefits. Including SDoH factors in the risk adjustment methodology for MA payments would reduce this risk for plans. To this end, the Association for Community Affiliated Plans (ACAP) has also proposed the Incentivizing Dual-Eligible Alignment (IDEAL) Act, which would include an SDOH adjustment factor to provide additional funds for additional SDOH services as supplemental benefits. ACAP estimates a 5% adjustment would result in $10 per member per month to support SDoH services.
Conclusion
If CMS were to move forward with SDoH adjustment, it is likely to be accompanied by strong expectations for MA plans to implement more programs to address SDoH needs. Yet, based on the early reaction to the CHRONIC Care act, it is not clear to many plans which interventions work best. There will also be challenges in coordinating benefits with social services organizations. Thus, it will be important for plans to document and evaluate their efforts and for CMS to help share best practices. Including SDoH factors in risk adjustment, as well as in the MA Star Rating program, can help ease ROI concerns plans have while also creating incentives to make meaningful investments.