Comparison of Air Pollution Exposures and Health Effects Associations Using 11 Different Modeling Approaches in the Women's Health Initiative Memory Study (WHIMS)

Background: Many approaches to quantifying air pollution exposures have been developed. However, the impact of choice of approach on air pollution estimates and health-effects associations remains unclear.

Objectives: Our objective is to compare particulate matter with aerodynamic diameter () concentrations and resulting health effects associations using multiple estimation approaches previously used in epidemiologic analyses.

Methods: We assigned annual exposure estimates from 1999 to 2004 derived from 11 different approaches to Women's Health Initiative Memory Study (WHIMS) participant addresses within the contiguous US. Approaches included geostatistical interpolation approaches, land-use regression or spatiotemporal models, satellite-derived approaches, air dispersion and chemical transport models, and hybrid models. We used descriptive statistics and plots to assess relative and absolute agreement among exposure estimates and examined the impact of approach on associations between and death due to natural causes, cardiovascular disease (CVD) mortality, and incident CVD events, adjusting for individual-level covariates and climate-based region.

Results: With a few exceptions, relative agreement of approach-specific exposure estimates was high for concentrations across the contiguous US. Agreement among approach-specific exposure estimates was stronger near monitors, in certain regions of the country, and in 2004 vs. 1999. Collectively, our results suggest but do not quantify lower agreement at local spatial scales for . There was no evidence of large differences in health effects associations with among estimation approaches in analyses adjusted for climate region.

Conclusions: Different estimation approaches produced similar spatial patterns of concentrations across the contiguous US and in areas with dense monitoring data, and effects associations were similar among estimation approaches. estimates and effects associations may differ more in samples drawn from smaller areas or areas without substantial monitoring data, or in analyses with finer adjustment for participant location. Our results can inform decisions about estimation approach in epidemiologic studies, as investigators balance concerns about bias, efficiency, and resource allocation. Future work is needed to understand whether these conclusions also apply in the context of other air pollutants of interest. https://doi.org/10.1289/EHP12995.