(B-036) Timely Disease Surveillance: Can a Large National EHR Network be Used to Estimate National and State Prevalence of Obesity and Type 2 Diabetes?

Background: United States (US) office-based physician electronic health record (EHR) adoption is nearly 90%, according to the 2021 National Electronic Health Records Survey. EHR systems that routinely collect vital signs, lab and test results, clinical diagnoses, and treatment interventions have the ability to revolutionize disease surveillance through the use of real-world data. However, there is limited data on how accurately large, national EHR networks can accurately estimate national and state-level prevalence rates.

Objectives: To evaluate the use of EHR data for disease surveillance by comparing EHR-derived obesity and type 2 diabetes (T2D) prevalence rates to published estimates.

Methods: Adults (18+) with 12+ months of activity between 2021-2023 in the Veradigm Network EHR (VNEHR) database were identified. VNEHR prevalence of obesity (body mass index of 30+) and T2D (by diagnosis code) is presented and compared to corresponding values for US adults, by demographic characteristic and state. Comparator data were derived from the National Health and Nutrition Examination Survey (NHANES) and the Behavioral Risk Factor Surveillance System (BRFSS) for the years 2017-2020

Results: This analysis included 38.2 million adults in the VNEHR between 2021-2023. The EHR population had a higher proportion of adults 65 + (36.4% vs. 21.6%) and higher proportion of females (59.5% vs 50.5%) compared to the US adult population. The overall prevalence of obesity was similar between VNEHR (42.4%) and NHANES (41.9%). Overall prevalence of T2D was 13.7% in VNEHR and 11.3% in NHANES. The prevalence of obesity among adults stratified by gender, race, or ethnicity was consistent with published prevalence estimates. The prevalence of T2D in VNEHR among adults stratified by age, gender, race and ethnicity tended to be higher than published prevalence patterns but followed similar trends. For both obesity and T2D, the geographic pattern for VNEHR values was similar to that of published values. States with higher prevalence were concentrated in the South and Appalachia, and states with lower prevalence were concentrated in the Northeast and Mountain regions.

Conclusions: Findings demonstrate strong alignment between EHR-derived prevalence and published US estimates. Large, national EHRs can provide representative, timely disease prevalence estimates for select groups.