Investigating mortality heterogeneity among neighbourhoods of a highly industrialised Italian city: a meta-regression approach

OBJECTIVES: The purpose of this study was to investigate the role of various predictors to explain spatial mortality heterogeneity in Taranto. METHODS: Direct age-adjusted death rates (ADR) at a neighbourhood level for the period 1998-2010 were examined. SO2, PM10, distance from pollution sources, and socioeconomic status (SES) were tested as predictors within a meta-regression framework. We used ? (2) to quantify heterogeneity in ADR and I (2) statistic with 95 % confidence intervals to estimate the proportion of total variation across neighbourhoods attributable to the between-neighbourhood heterogeneity. RESULTS: High heterogeneity resulted for all and natural causes of death for both genders. One neighbourhood (Paolo VI) was detected as an outlier for all predictors except SO2, among males. After accounting for SES, moderate heterogeneity among residuals was observed for all-causes of death and was correlated with SO2. Higher concentrations of PM10 were observed in neighbourhoods close to the industrial site and higher concentrations of SO2 in neighbourhoods more distant from the industrial site. CONCLUSIONS: SES and air pollutants were predictors of spatial heterogeneity in ADR. Different distributions of SO2 and PM10 in the city suggested two exposure patterns.