OBJECTIVES: Social determinants of health (SDOH) are estimated to drive up to 80% of health outcomes. This analysis aimed to link patient/household level SDOH characteristics from consumer data to select closed claims (CC) disease cohorts for inclusion in real-world data analysis.

METHODS: CC from commercially insured enrollees (01/01/2016-12/31/2021) and 2022 SDOH data including demographics, socioeconomic, and household information, were utilized(CHRONOS. 2017-2023. Forian, Inc., Newtown, PA. https://forian.com). Both data sources are HIPAA compliant and linked by a unique anonymized patient identifier. Patients aged 18+ with evidence of HIV, chronic kidney disease (CKD), heart failure (HF), type 2 diabetes (DM2), and metastatic prostate cancer (mPC) were identified using ICD-10-CM diagnosis codes in CC before linking to SDOH data. Patients were classified as overlapped if there was ≥1 record in SDOH data and ≥1 claim with a diagnosis of interest in CC. Descriptive statistics were evaluated for age, sex, race, and custom-defined composite measures for household economic status (economic stability indicator (ESI), household income). ESI ranges from 0-30 with higher numbers indicating less economic stability.

RESULTS: In total 2,056,890 CC patients (6.8%) had a linkable SDOH, ranging from 30.2% of mPC to 40.4% of HIV patients. HIV patients had the most racial diversity, while mPC and HF patients had the least. Although all patients were commercially insured, 75% lived in households with annual incomes below the US median ($75,000). Of those with household incomes above the median, 58.1% HIV, 37.9% DM2, 36.3% HF, 33.9% CKD, and 25.0% mPC patients had ESI values >10, indicating low economic stability relative to household income.

CONCLUSIONS: SDOH measures provide insight into disease-specific patient cohorts beyond standard demographic data. Composite measures and interactions may provide deeper insights into study populations. Including patient/household level rather than geographic level SDOH measures may remove variability and bias when measuring health outcomes and costs.