Link Your Large Health Data Sets to the Area Deprivation Index, the ezADI Way.

Abstract

Increasing attention has been paid to investigations on how social determinants of health (SDOH; e.g., income, employment, education, housing, etc.) impact health outcomes. However, these variables are often not collected in routine clinical practice. As a consequence, researchers may attempt to link retrospective medical records to those datasets that can provide additional SDOH information, such as the Area Deprivation Index (ADI). However, time-consuming geographic calculations can deter these analyses. To reduce this burden, the ezADI R package performs batched geocoder mapping on inputted addresses, constructs Federal Information Processing Series (FIPS) codes, and then merges these data with ADI scores. The applicability and feasibility of this ezADI tool was tested on a sample of patients with sickle cell disease (SCD). Individuals with SCD are at risk for developing serious comorbidities; disadvantageous SDOH may increase this risk, in turn leading to higher rates of hospital utilization and longer lengths of stay on admission. In this sample of 1,105 individuals with SCD in Tennessee (53.8% female, 97.5% African American), higher ADI scores (i.e., more neighborhood disadvantage) were significantly associated with increased hospital utilization (rho = 0.093, p = 0.002) and longer lengths of stay (rho = 0.069, p = 0.021). These areas could be targeted with neighborhood-level interventions and other resources to improve SDOH. This study provides proof of concept that the ezADI tool simplifies geocoding calculations to allow researchers to link datasets with the ADI and assess associations between SDOH factors and health outcomes.