Beyond the mean: Examining change in the BMI distribution in Belgium from 2014 to 2023

Abstract Background Body Mass Index (BMI) is a key indicator for monitoring population health in obesity prevention. Analyses that rely solely on mean BMI or obesity prevalence may overlook differences in distribution shape and skewness across population subgroups and mask disparities. This study examined how sex, age, education, and region accounted for change in BMI distribution in Belgium from 2014 to 2023. Methods Data came from two Belgian Food Consumption Survey waves, a nationally representative cross-sectional survey. Participants aged 18-64 (n = 2443) provided sociodemographic and anthropometric data at two home visits. We used quantile regression and GAMLSS models (BCCGo distribution) predicting mean, variance, and skewness. Covariates were survey year, sex, 10-year age groups, education, region, and the interactions between the 3 former variables with year. Results Among the 5% of the population with the lowest BMI, women's BMI increased while men's decreased between 2013 and 2024. At the 50th and 75th percentiles, BMI rose in 18-24-year-olds and declined in 55-64-year-olds. At the 50%, 75%, and 95% percentiles, individuals with medium education levels showed BMI increases compared to those with low education levels. At the 95% percentile, individuals with high education also showed BMI increases compared to those with low education levels. GAMLSS models showed decreasing BMI variability in women, but an increase among men. Young adults experienced growing BMI variability, unlike older adults, whose BMI variability remained stable. BMI skewness increased over time among those with medium education and remained skewed among those with low education. Conclusions Despite no change in mean BMI over time, this study highlights the value of models capturing distributional change, enabling early detection of risk beyond obesity prevalence changes. Prevention strategies should prioritize young adults and those with lower education, who showed unfavorable BMI shifts. Key messages • BMI increased at specific percentiles of the distribution, while becoming more unequal among young adults, therefore confirming the need for early, targeted prevention efforts. • Distributional regression analyses revealed specific BMI disparities; moving beyond averages enables more precise public health action among specific target groups.