Longitudinal changes in BMI and obesity across COVID-19 pandemic periods among underrepresented populations: An analysis of the All of Us dataset
Abdul-Hanan Saani Inusah, Huiyi Xia, Atena Pasha, Zhenlong Li, Andrew T. Kaczynski, Xiaoming Li, Shan QiaoIntroduction
The COVID-19 pandemic significantly affected physical activity and lifestyle, potentially increasing Body Mass Index (BMI) and obesity risk. We examined changes in BMI and obesity across Pre-Pandemic, Pandemic Peak, and Peri-Pandemic periods and identified sociodemographic correlates.
Methods
We analyzed longitudinal electronic health record data from the All of Us Research Program for adults ≥18 years with at least one BMI record in each period (N = 38,632). Periods were defined as Pre-Pandemic (Jan 1, 2018–Mar 12, 2020), Pandemic Peak (Mar 13, 2020–Dec 31, 2021), and Peri-Pandemic (Jan 1, 2022–Oct 31, 2023). BMI was modeled using linear mixed-effects models and obesity (BMI ≥ 30) using GEE logit models, adjusting for age, sex, race, ethnicity, income, and employment. A sensitivity analysis was conducted by expanding eligibility to participants with BMI data in at least two periods (N = 83,698).
Results
Mean BMI increased from 30.05 (SD = 6.67) Pre-Pandemic to 30.14 (SD = 6.76) during the Pandemic Peak, then declined to 29.96 (SD = 6.66) Peri-Pandemic; obesity prevalence followed a similar pattern (43.1% → 43.4% → 42.8%). In adjusted models, the linear time effects were positive and the quadratic terms negative for both outcomes (BMI: β = 0.218 and −0.018; obesity: aORs = 1.053 and 0.996; all p < 0.001). Female, Black/African American, and Hispanic/Latino participants had higher adjusted BMI and greater odds of obesity than their respective reference groups, while higher income was associated with lower BMI and obesity odds. Significant time × subgroup interactions were observed, with age and employment status showing the most consistent effects across both outcomes.
Conclusion
BMI and obesity displayed statistically detectable but modest changes across the pandemic-defined periods. Persistent adjusted differences by sex, race, ethnicity, and income highlight the relevance of equitable access to obesity-prevention resources.