Addressing social determinants of health (SDH) is viewed as fundamental to improving care delivery and lowering healthcare costs in the US. However, information on SDH is difficult to obtain and existing data sources are incomplete. Because of these limitations, researchers commonly rely on readily available aggregate proxy data (e.g., census tracts and block groups) to make inferences about individual-level SDH characteristics within a geographic area.

Avalere’s study, presented at AMCP Nexus 2018 on October 22, 2018, compares two aggregate proxy data sources—9-digit ZIP code and census block group—on their ability to predict health outcomes, specifically medication adherence, based on SDH characteristics. The results indicate that aggregating data across a smaller group of individuals (e.g., the 9-digit ZIP code) can provide a more precise measurement of sociodemographic differences, and thus would be more useful to researchers in predicting patient outcomes, risk, and costs. It also underscores the importance of using granular enough data in risk adjusting quality measures to ensure a level playing field for providers so that rewards or penalties are fairly assessed and based on true differences in performance.

To learn more about this study and Avalere’s research capabilities, connect with us.