247 – 247 - Causal Inference and Statistical Learning of Intervention and Policy Effects
A Propensity-Matched Study of Obesity and Diabetes Risk: A COVID-19 Perspective
Claire Nurse
SUNY Downstate School of Public Health
The COVID-19 pandemic spurred widespread weight gain, suggesting that type-2 diabetes may also increase. However, past research finding an association between obesity and type 2 diabetes is subject to potential confounding on socioeconomic and other variables. Using data from the nationally representative longitudinal Midlife in the US study, we evaluated the association between obesity and diabetes 18 years later using robust regression methods. To reduce potential confounding, we used nearest-neighbor propensity score matching on demographics (age, race, gender), socioeconomic status (educational attainment, household income), and health risks (physical activity, smoking status) to match adults with obesity (n=450) and without obesity (n=617), which resulted in covariate balance assessed by standardized differences. In Poisson regression in the matched sample, individuals with obesity had 2.3 times the risk of diabetes in 18 years than those without obesity (ARR , 95% confidence interval (1.62, 3.50)). COVID-19 pandemic recovery must include population-based obesity interventions (e.g., optimizing agricultural subsidies to target obesity) and type-2 diabetes screening.