VISIT STATE HEALTH COMPARE
Blog & News
New Brief Explores State Approaches to Risk Adjustment Based on Social Factors and Strategies for Filling Data Gaps
August 2020:
The role that social factors, such as housing instability, food insecurity, and access to transportation, have in influencing health has been a growing focus in health policy and health care. However, the concept of “social determinants of health” (SDOH), the idea that individuals have unique circumstances that can affect their health care outcomes, is not new to health care. In fact, the concept of “social determinants of health” (SDOH)—the idea that individual’s social circumstances can affect their health—is being used in some states to augment existing health care risk-adjustment tools, which are often used in setting payment rates or calculating performance on quality measures.
A new State Health & Value Strategies (SHVS) brief, authored by SHADAC researchers, details three approaches taken by organizations in two states, Minnesota and Massachusetts, to solving a key challenge to risk adjusting based on social factors: Where to find data on social risk factors? The brief’s profiled examples employed three approaches to filling the data gap—using survey data, using claims and administrative data, and using a combination of both.
Minnesota Medicaid ACO
In 2018, Minnesota launched an update to the state Medicaid program’s accountable care organization (ACO) model, known as “Integrated Health Partnerships” (IHPs). Among other changes, the updated IHP program introduced a new population-based payment that is now adjusted based on the social risk factors (in addition to medical conditions) of IHP program beneficiaries. The rationale for the change was to support activities typically not reimbursed by Medicaid, recognizing that some beneficiaries have greater risk factors—including social risk factors.
To create its new risk-adjustment methodology, Minnesota leaned on earlier research conducted by state Medicaid staff that had used claims and other administrative data to study the relationship between social risk factors and health outcomes. Informed by that research, the state developed a risk-adjustment approach that used beneficiary-level claims and administrative records that would adjust payments based on records of individual beneficiaries’ own social risk factors, such as an income level that would be considered as “deep poverty” (based on Medicaid eligibility data) and homelessness (based on self-reported information or analysis of beneficiary addresses).
Minnesota Community Measurement
Another entity in the same state, Minnesota Community Measurement (MNCM), is a non-profit organization that collects, analyzes, and publishes health care quality and cost data. MNCM undertook an effort to incorporate social risk factors into its risk-adjustment approach based on recommendations from the organization’s board of directors and another stakeholder committee.
Unlike Medicaid agencies, MNCM does not have access to detailed data on individual patients beyond basic demographics. However, because health care providers report ZIP codes for individual patients to MNCM, they were able to develop a geographically based index measure drawn from the U.S. Census Bureau’s American Community Survey (ACS), using the social risk factors associated with the ZIP codes where individual patients live as a proxy for their own social risk factors. Although it is not a Medicaid agency, the approach used by MNCM to risk adjust based on social risk factors could be readily duplicated by a state Medicaid agency because it uses publically available data.
Massachusetts Medicaid ACO and MCO
In Massachusetts, the state’s Medicaid program has incorporated measures of beneficiaries’ social risk factors into its methodology for risk adjusting payments to Medicaid Managed Care Organizations (MCOs) since 2016 and Accountable Care Organizations (ACOs) since 2018. Part of the state’s rationale for accounting for social risk factors in its payments to MCOs and ACOs was to “mitigate the incentive (they) might otherwise have to limit care or avoid members with greater health care needs,” recognizing that individuals with social risk factors may face additional needs and greater challenges to optimal health outcomes.1,2
Unlike the prior examples, Massachusetts has developed a hybrid approach using both administrative and survey data. Similar to Minnesota Medicaid, Massachusetts uses various elements of Medicaid claims data and administrative data tied to individual beneficiaries from within the Medicaid agency and other departments, such as data on disability status and housing instability. And similar to MNCM, Massachusetts created a geographically based index of other social risk factors using data from the ACS. By using those dual approaches, Massachusetts was able to fill more gaps in social risk factor data than either single data source was able to address on its own.
Massachusetts also took an initial step toward collecting new data on individual beneficiaries’ own social risk factors. The state began using ICD-10 Z-codes for homelessness as part of its methodology for risk adjusting based on social factors. Although Z-codes for homelessness already existed, using them as part of the state’s risk adjustment methodology—which informs payments to health care providers—gives providers an incentive for reliably screening for and documenting whether individual patients are experiencing homelessness.
Key takeaways
A key challenge to risk adjusting based on social factors is obtaining data regarding patients’ relevant social risk factors. To address that limitation, the three examples profiled in this brief highlight possible approaches to filling those gaps that other states’ Medicaid programs could reproduce or build from:
- Using survey data as a proxy for individuals’ own social risk factors;
- Repurposing existing administrative data pertaining to social risk factors;
- And taking initial steps toward collecting data on social risk factors, such as through systematic use of billing Z-codes.
1 Seifert, R.W., & Love, K.A. (2018). What to know about ACOs: An Introduction to MassHealth Accountable Care Organizations. Retrieved from https://www.bluecrossmafoundation.org/sites/default/files/download/publication/ACO_Primer_July2018_Final.pdf
2 Massachusetts’ revised risk adjustment methodology also includes some measures that are not necessarily social risk factors, such as interaction measures specific to children—ages 0-14 and 15-20—between age and medical risk scores, to account for the particular health care costs for children with complex medical situations compared to healthy children.
