Abstract
OBJECTIVES: Regional public health policy-makers frequently adopt obesity programs and objectives that have been established at global, provincial/state or national levels. However, the presence of substantial inter-regional disparities could render this practice inefficient. Studies that collectively assess obesity prevalence, temporal trends and their heterogeneity at the region level are rare, though they could be used to support better regional surveillance and planning. To address this gap, our study projected obesity prevalence time series to 2023 for 16 health regions in Quebec. We also compared the extent to which yearly rates of increase (or slope) versus cross-sectional prevalence drove regional heterogeneity and correlated with obesity-related sociodemographic and behavioural characteristics. METHODS: Projections were done using weighted compositional regression to fit and extrapolate obesity prevalence time series (1987-2012). Heterogeneity in obesity prevalence as a function of time and obesity slope were characterized using standard deviation. The correlation of region-level obesity prevalence and slope with 14 area-level obesity-related characteristics was assessed. RESULTS: Obesity prevalence is projected to increase in all regions. Region-level heterogeneity in prevalence in 2012 (σ = 2.2%) is projected to increase to (σ = 3.1%) by 2023. The increase in prevalence heterogeneity appeared to be driven by region-level heterogeneity in slope (β = 0.22%-0.51%/year). Obesity-related characteristics were found to be more strongly correlated with slope than with prevalence. CONCLUSION: Large area obesity trends mask substantial and increasing region-level disparities. Obesity slope appears to drive region-level heterogeneity and correlate strongly with explanatory factors, and may represent a pertinent metric for public health monitoring.