There is a license error on this site:
The license file "C:\Sites\SOA-11-14\License.config" does not exist
The Web site remains functional, but this message will be displayed until the license error has been corrected.

To correct this error:If you do not have a license file, please request one from EPiServer License Center.
Integration of Social Determinants of Health into Medicaid Managed Care Risk Adjustment: Considerations and Financial Impacts | SOA
Log in / My SOA

Integration of Social Determinants of Health into Medicaid Managed Care Risk Adjustment: Considerations and Financial Impacts

soa-research-societal-purpose-logo.png

June 2025

Authors

Rong Yi, PhD

Jeffrey Milton-Hall, FSA, MAAA

Nicholas R. Gersch, FSA, MAAA

Brian Merkey, PhD

Megan Nicholson, MS

Executive Summary

To evaluate the impact of SDOH risk adjustment on financial outcomes for Medicaid ACOs and managed care organizations (MCOs, which the authors refer to interchangeably with ACOs in this report), the authors calibrated risk-adjustment models with the addition of SDOH risk factors and then they simulated financial outcomes for ACOs/MCOs across synthetic state Medicaid managed care markets with and without SDOH risk adjustment.

Integration of SDOH into Medicaid risk adjustment was found to provide a small but measurable reduction in volatility and variation of financial results versus a recalibrated model without SDOH (morbidity-only). The improvement in fit, while modest, is greatest for populations with below-average morbidity risk and above-average SDOH-related risk. Prospective risk scores for these beneficiaries were too low under morbidity-only models but closer to parity after incorporating SDOH, suggesting that SDOH risk factors may help right-size revenue for populations with costs that reflect utilization of services below their healthcare needs as indicated by prior-year diagnoses.

However, the authors also found that the SDOH risk factors in their models explained much less variation in estimated risk than existing Chronic Illness and Disability Payment System with Pharmacy (CDPS+Rx) condition categories, and that the financial impacts of SDOH risk-adjustment integration were effectively undetectable under an alternate sampling method that doesn’t explicitly stratify MCOs by SDOH risk.

Several important lessons from this analysis and the study results included the following:

  • Defining and measuring SDOH is challenging, and there are many different ways to do so.
  • National individual-level SDOH data sources are hard to come by and are currently not reliably populated.
  • Empirical approaches based on historical claims data—such as risk adjustment—have limitations that make them a less-than-ideal tool for identifying and addressing the basis for differences in care, and whether those are caused by differences in access to care or underlying health needs.
  • Integration of SDOH into risk adjustment offers an opportunity for payers to address policymaker objectives of incentivizing the provision of appropriate care that enhances the health outcomes of all of its members. However, its success will depend on ensuring that SDOH risk factors have a meaningful impact on revenue allocation and align with MCOs' ability to implement effective interventions. In this focused research study, the influence of community-level SDOH as separate risk factors on revenue allocation was determined to be limited. There may be different interpretations.
    • Recent studies[1] have shown that disease prevalence rates are correlated with SDOH. The impact of SDOH on determining underlying member morbidity may have been partially captured through the condition categories, hence adding SDOH as separate risk factors would not contribute significantly relative to the condition categories. Or,
    • The methodology design used in this study does not fully capture the complex interplay between SDOH and morbidity. Or,
    • Community-level SDOH data lacks granularity. It is likely that individual-level SDOH may show more predictive value in risk adjustment.

There may well be other interpretations. Further research and additional data are needed to better understand this important topic.

  • Risk adjustment, as a research methodology, a payment mechanism, and policy instrument, is constantly evolving. SDOH risk adjustment may hold the potential to drive meaningful change, but its full impact may take time to emerge as the interplay between coding incentives, care delivery, and the data informing risk scores and capitation rates continues to evolve.
  • Rather than viewing SDOH risk adjustment as a standalone remedy, it is more prudently considered as one component of a comprehensive approach to appropriately address the needs of the entire population.

Finally, several opportunities for refinement in future research were identified including the following:

  • Integrating individual-level SDOH risk factors. As of January 2024, a new CPT code G0136 became effective to pay for administering an SDOH risk assessment. It is possible that individual-level SDOH data will become more widely available in the near future.
  • Attempting to tease out access from need through a methodology design that first predicts access to care using non-claim risk factors and then predicts expected cost conditional on receipt of care.
  • Expanding the current analysis to encompass additional data years and states to incorporate populations outside of the current cohort, such as dual eligibles, pregnant women, and recipients of long-term services and supports.
  • Improving statistical power and generalizability of the calibrated SDOH risk-adjustment models.

Material

SDOH in Medicaid Risk Models

Acknowledgments

The authors’ deepest gratitude goes to those without whose efforts this project could not have come to fruition: the volunteers who generously shared their wisdom, insights, advice, guidance, and arm’s-length review of this study prior to publication. Any opinions expressed may not reflect their opinions nor those of their employers. Any errors belong to the authors alone.

Project Oversight Group:
Joan Barrett, FSA, MAAA
Max Billings, FSA, MAAA
Craig Cartossa, ASA, MAAA
Reese Dai, FSA, MAAA
Samuel Driscoll, FSA, MAAA
Ian Duncan, FSA, FIA, FCIA, MAAA, FCA
Wendy Feng, FSA, MAAA
Thomas Lemire, ASA, MAAA
Pei Pei, ASA, PHD
Tanvi Tilloo, MPP
Rodger Yan, FSA, MAAA
Jing Zhang, FSA, MAAA

At the Society of Actuaries Research Institute:
Achilles Natsis, FSA, MAAA, FHLI, Health Research Actuary
Barbara Scott, Senior Research Administrator

Questions or Comments?

Give us your feedback! Take a short survey on this report. Take Survey

If you have comments or questions, please send an email to Research@soa.org

[1] For an example on the relationship between community-level SDOH and disease prevalence, see Benavidez, G. A., Zahnd, W. E., Hung, P., et al. (2024) Chronic Disease Prevalence in the US: Sociodemographic and Geographic Variations by Zip Code Tabulation Area. Preventing Chronic Disease, vol. 21, doi:10.5888/pcd21.230267. For an example on the relationship between individual-level SDOH and disease prevalence, see Kunnath, A. J., Sack, D. E., and Wilkins, C. H. (2024) Relative Predictive Value of Sociodemographic Factors for Chronic Diseases Among All of Us Participants: A Descriptive Analysis. BMC Public Health, vol. 24, Article no. 405, doi:10.1186/s12889-024-17834-1.