Tracking Social Determinants of Health Data in Medicare Advantage

Gathering social determinants of health data continues to be a challenge for Medicare Advantage health plans, according to a report that NORC at the University of Chicago (NORC) compiled for Better Medicare Alliance (BMA).

The NORC researchers did a literature review related to Medicare Advantage plans’ social determinants of health data programs. They also conducted qualitative interviews with over 20 experts from various stakeholders including health plans, providers, and vendors as well as four policy experts.

Payers pulled social determinants of health data from a couple of main sources to identify Medicare Advantage beneficiaries’ needs.

While medical claims might seem like a convenient place to harbor social determinants of health data, providers often do not have the ability to record social services in the EHR. Even racial and ethnicity data is hard to access in medical claims.

Although medical claims may be sparse with social determinants of health details, other internal sources can provide this data, including administrative sources.

Health risk assessments are one key way in which health plans collect social determinants of health data. These assessments typically provide an overview of lifestyle factors, demographic characteristics, and personal and family medical history but have evolved to include social determinants of health indicators as well.

These assessments are particularly useful for determining dual eligible social determinants of health. 

SCAN Health Plan, for example, stated that more than eight out of ten of their beneficiaries who are in dual eligible special needs plans, around eight in ten of their chronic condition special needs plan beneficiaries, and over a third of their non-special needs plan beneficiaries have completed the assessment.

Plans also employ direct questionnaires and surveys during home healthcare visits. Every health plan that NORC interviewed stated that care managers are one of the best sources of social determinants of health data.

Outside of the payer organization’s data collection mechanisms, health insurers turn to aggregated geographic-level data. 

The US Census or other major ranking systems can supply actionable information which payers can then incorporate into their efforts. 

However, this information is high-level and de-identified. De-identified data can pose a barrier to interventions. For example, when the government gave payers access to de-identified coronavirus vaccination data, payers argued that the information did not help them target their vaccination efforts.

“Of note, several health plans indicated that it would be helpful to receive currently unavailable information about beneficiaries who qualify for certain government assistance programs, such as the Supplemental Nutrition Assistance Program,” the researchers stated.

In the future, Medicare Advantage plans may utilize is the International Classification of Diseases, Tenth Revision (ICD-10) “Z codes” as a source of social determinants of health data. 

This new type of classification sets aside a Z code for different social determinants of health. Using this code, providers can indicate when a social determinant of health issue precipitated a medical condition.

While Z codes were slow to gain traction, the report stated that they have been making progress and specifically highlighted UnitedHealthcare’s experience encouraging providers to use Z codes. By partnering with the American Medical Association and partnering with health systems, UnitedHealthcare has seen Z code utilization grow.

Medicaid program adoption of Z codes has further bolstered the use of this classification system.

The NORC researchers put forward three policy recommendations in response to these findings.

First, they suggested that CMS should offer better guidance for Medicare social determinants of health data collection, particularly with a focus on the dual eligible special needs plan Model of Care Health Risk Assessment.

Second, the researchers urged more social determinants of health data sharing between federal and state healthcare agencies—including HHS and Medicaid programs—and payers to improve interventions in Medicare Advantage, traditional Medicare, and Medicaid managed care.

Finally, NORC researchers recommended that CMS take steps to promote Z code utilization among providers. In particular, they noted that training, guidance regarding referrals, and financial rewards might bolster providers’ usage and help increase social determinants of health data intake.