OBJECTIVES: Social determinants of health (SDOH) is important in addressing health outcomes and disparities across population groups by extracting data from Electronic Health Records. This research aims to evaluate the association of SDOH with Type 2 Diabetes (T2DM) treatment selection and disease severity at treatment initiation.

METHODS: De-identified data from the multi-institutional University of California Health Data Warehouse (UCHDW) were extracted for patients > 18 y.o with a T2DM diagnosis or a Hemoglobin A1c (HbA1c) > 5.7 % and received an antihyperglycemic medication from January 2012-December 2021. T2DM disease severity was determined by patients' HbA1c levels. As an initial view into patients’ SDOH, Area Deprivation Index (ADI) was used to rank individuals from 1-10 (least to most disadvantaged) using census tract geolocation. Statistical comparisons were performed using ordinal and multinomial logistic regression.

RESULTS: The inclusion criteria led to 124,364 patients extracted for data analysis from six University of California academic health centers, with 49.2% starting on Metformin, 25.6% Sulfonylureas, 16.1% SGLT2 inhibitors, 4.5% DPP4 inhibitors, 2.8% GLP-1 agonists, 1.6% TZD, 0.2% AGI’s. In multinomial regression, ADI 10 was most strongly associated with DPP4 inhibitors treatment selection in comparison to ADI 1 (p = 0.008). In ordinal regression, T2DM disease severity was significantly associated with ADI 1 being less severe than ADI 10 at treatment initiation (p=0.001).

CONCLUSIONS: This study demonstrates significant associations between SDOH and health outcomes of T2DM, including differences in treatment selection and disease severity at treatment initiation. More disadvantaged patients were more likely to have severe disease at treatment initiation and receive DPP4 inhibitors compared to least disadvantaged patients. Incorporating SDOH factors into clinical data collection and analyses is critical to improving health equity.