New Study Identifies Health Plan SDoH Data Challenges

Without data on Social Drivers of Health, there can be no intervention, measurement or improvement.

In 1624, poet John Donne wrote that gold bullion must be “coined into current money” to be of use. Four hundred years later, data is the gold that health plans must coin — extract, collect, store, share, analyze and apply — to achieve results.

Some of the most difficult gold to coin is data, specifically related to Social Drivers of Health (SDoH) data. A new qualitative study published in Health Affairs details health plans’ top three SDoH data challenges and possible solutions.

Progress and pitfalls as SDoH measure application grows

SDoH are the non-clinical factors — food insecurity, housing instability, transportation needs, difficulty paying utilities, interpersonal safety — that contribute to 80% of clinical outcomes. Healthcare has made SDoH progress. Quality measures for five SDOH domains are included or proposed in 21 federal programs, initiatives, or guidance documents:

There is also far to go. As noted in the Health Affairs study: “The ability of SDOH quality measures to improve care is predicated on health plans and other entities being able to collect and report these data.” Not only collect and report but analyze and implement. Not only improve care but make it more affordable and equitable — all in a way that is both standardized and patient centric.

The study adds: “Widespread implementation of SDOH quality measures is therefore dependent on an array of factors related to a health plan’s data system, including what SDOH-related fields are collected, how they are coded and stored, and interoperability.”

Data challenges persist as SDoH domains and measures grow. The Health Affairs study included two domains being developed by the National Committee for Quality Assurance (NCQA) — utility insecurity and social connection — and a third new measure — Social Need Screening and Intervention (SNS-E), which addresses unmet food, housing, and transportation needs.

The study included eight health plans that serve between 100,000-47 million government program members across the U.S. The findings produced three SDoH data barriers with mixed results across plans: 1) Health plan coding; 2) Health plan storage, extraction, and mapping; and 3) Systemwide considerations

Barrier 1: Health plan variable coding capabilities

To receive NCQA accreditation, health plans must submit HEDIS measures (Health Effectiveness Information Data Set). To collect and report on SDoH via HEDIS, health plans must be able to submit data via one of three standardized code sets: LOINC, SNOMED or CPT (hereafter “Code Sets). [1]

In the Health Affairs study, plans differed in their ability to extract these Code Sets:

• Only 37.5% (3 of 8) could access LOINC codes, which are associated with health observations and measurement.
• Of these 3 plans, only 2 could also pull SNOMED or CPT codes, used collectively for clinical terminology, interventions, services, and procedures.
• Most health plans could also report Z Codes, used for SDoH screenings and interventions), but they are currently not allowed by HEDIS.

Health plans reported that financial incentives could help standardize coding practices.

Barrier 2: Health plan data storage, extraction, and mapping

Health plans’ ability to map SDoH data from EHRs was also mixed. Again, only 37.5% had automation capabilities to map while four plans mapped manually. Other results pertained to the SNS-E measure. Using SNS-E, most plans could report screening indicators but not interventions. However, it was difficult to map the two or link intervention with follow-up because the SDoH data came from multiple databases and lacked the needed time stamps.

Barrier 3: Systemwide need for improvement

The above-referenced link between screenings and interventions requires standardizing screening tools, intervention definitions, and data terminology. This requires the following improvements:

• operationalizing SDOH data queries.
• strengthening interoperability between health plans and providers.
• facilitating referrals between health plans and community resources.

This is a systemwide opportunity. Health plans, as well as healthcare systems and community partners, must be able to align — with one another and with national measurement initiatives as they expand and change.

The plans also support federal policies that help facilitate, standardize and incentivize SDoH data practices. Current CMS policies designed to achieve this and updated in 2024 include:

• The Interoperability and Prior Authorization Final Rule: Requires government health plans (Medicare Advantage, Medicaid, CHIP, marketplace) to establish and maintain a provider access API for claims and encounter data sharing. The hope is that this rule, which plans must comply with by Jan. 1, 2027, will facilitate the tracking of SDOH screening and interventions, and the coding time stamps needed to link the two and support follow-up care.
•  
• Physician Fee Schedule: Incentivizes the use of specific codes for SDOH risk assessments and to improve documentation. The study notes that these codes are “an incremental step toward collecting standardized SDOH data while interoperability continues to develop.”

Collectively these three barriers and opportunities for improvement deliver early lessons in how to integrate SDoH measures into practice. The Health Affairs study concludes: “Further research is needed to explore additional codes, mechanisms for collecting SDOH data in a patient-centric manner, and ensuring that health plans, health care systems, and community partners can align with national measurement initiatives. Standardizing these data will be key to improving outcomes for all.

[1] LOINC: Logical Observation Identifiers, Names, and Codes. SNOMED: Systematized Nomenclature of Medicine, used for clinical terminology. CPT: Current Procedural Terminology, used for interventions, services, and procedures.