Small scale analysis of urban air quality and public health in London, UK, provides insights for targeted interventions

Understanding the impact of scale and zonation is critical for accurately assessing population health in relation to air quality and demographic data. Using existing census geographies, we analyse spatial clustering and statistical associations across different census scales, focussing on vulnerable sociodemographic groups and temporal exposure patterns. This study aims to assess how different spatial resolutions affect the strength and interpretation of associations between air pollution, sociodemographic variables, and self-reported health. It also evaluates whether finer-scale, health-need-based zoning provides a more accurate basis for public health analysis. Exposure to higher PM 2.5 concentrations, and having an undrlying disability, or long-term illness shows robust associations with poor health (p < 0.001 at all census scales). However, as spatial resolution becomes coarser, the explanatory power of demographic variables weakens, underscoring the risk of ecological fallacy and misinterpretation when relying on aggregated data. Notably, demographic variables become less significant with coarser spatial resolution, supporting the need for scale-sensitive approaches in population health studies. Multiple linear regression analyses demonstrate that explanatory power in health models strengthens at coarser scales, with potential overfitting at ward levels due to high R-squared values. Whilst a stronger model fit is observed at higher levels of aggregation, this may mask within-area heterogeneity and obscure critical local disparities. Our findings suggest that effective public health policies benefit from granular and contextually aligned zoning strategies, which enhance the accuracy and relevance of health assessments. The study highlights the value of fine-scale, health-need-based geographic units in capturing the spatial nuances of population health and provides evidence supporting their use in equitable resource allocation and intervention design. The findings provide a framework for evaluating environmental and demographic factors at appropriate geographic scales to support targeted, equitable health interventions.