Do temporal trends of associations between short-term exposure to fine particulate matter (PM2.5) and risk of hospitalizations differ by sub-populations and urbanicity-a study of 968 U.S. counties and the Medicare population

Chen Chen,Alisha Chan,Francesca Dominici,Roger D Peng,Ben Sabath,Qian Di,Joel Schwartz,Michelle L Bell
DOI: https://doi.org/10.1016/j.envres.2021.112271
2022-04-15
Abstract:While associations between short-term exposure to fine particulate matter (PM2.5) and risk of hospitalization are well documented and evidence suggests that such associations change over time, it is unclear whether these temporal changes exist in understudied less-urban areas or differ by sub-population. We analyzed daily time-series data of 968 continental U.S. counties for 2000-2016, with cause-specific hospitalization from Medicare claims and population-weighted PM2.5 concentrations originally estimated at 1km × 1 km from a hybrid model. Circulatory and respiratory hospitalizations were categorized based on primary diagnosis codes at discharge. Using modified Bayesian hierarchical modelling, we evaluated the temporal trend in association between PM2.5 and hospitalizations and whether disparities in this trend exist across individual-level characteristics (e.g., sex, age, race, and Medicaid eligibility as a proxy for socio-economic status) and urbanicity. Urbanicity was categorized into three levels by county-specific percentage of urban population based on urban rural delineation from the U.S. Census. In this cohort with understudied less-urban areas without regulatory monitors, we still found positive association between circulatory and respiratory hospitalization and short-term exposure to PM2.5, with higher effect estimates towards the end of study period. Consistent with current literature, we identified significant disparity in associations by race, socioeconomic status and urbanicity. We found that the percentage change in circulatory hospitalization rate per 10 μg/m3 increase in PM2.5 was higher in the 2008-2016 time period compared to the 2000-2007 period by 0.33% (95% posterior credible interval 0.22, 0.44%), 0.52% (0.33, 0.69%), and 0.67% (0.53, 0.83%) for low, medium and high tertiles of urban areas, respectively. We also observed significant differences in temporal trends of associations across socioeconomic status, sex, and age, indicating a possible widening in disparity of PM2.5-related health burden. This study raises the importance of considering environmental justice issues in PM2.5-related health impacts with respect to how associations may change over time.
What problem does this paper attempt to address?