The influence of specific local land-use activities (continuously redistributing elements across environments) and environmental conditions (altering the chemical composition of airborne particulate matter) on the intrinsic health risk of PM2.5 exposure is sparsely reported. To fill this gap, we employed a novel integrated approach to address the influence of short-term changes in source-specific PM2.5 composition on the exposure-response risk, while controlling for weather conditions. We combine receptor-based source apportionment with conditional logistic regression in a space-time-stratified case-crossover design. This approach is different from previous studies as it: i) controls the impact of spatiotemporal variations in air pollution and human mobility using multilocation-specific fixed and disjointed space-time strata ii) addresses the spatial heterogeneity of personal exposure separating its variable effect from other predictors by allowing different baseline hazards for each space-time stratum; iii) aligns case/control periods with strong/regular episodes of source-specific PM-multipollutant fingerprint contributions rather than health outcomes. This enabled comprehensive examination of the association between source-specific PM2.5-bound species and cardiorespiratory disease hospitalizations. The epidemiological findings were that primary anthropogenic emissions [industrial (ORs 2.5 – 4.8)] were associated with higher 1-day moving average PM-induced risks. Natural-related sources [fresh / aged sea salt aerosol, dust, soil resuspension] and secondary sulfate formation were consistently associated with higher health risks (ORs 1.0 – 1.54) after 1 to 5-days since exposure. The results emphasize the importance of source-specific air quality management in complex areas and our research provides an adaptable universal tool to support targeted place-based policy interventions to mitigate air pollution impacts on health.
Read full abstract