GH009-04
Short-term PM2.5 and cardiovascular admissions in NY State: assessing sensitivity to exposure model choice

Monday, 14 December 2020: 10:15
Virtual
Mike He1, Patrick Kinney2, Arlene M Fiore3, Vivian Do4, Xiaomeng Jin5, Silliang Liu4, Nicholas DeFelice1, Jianzhao Bi6, Yang Liu7, Tabassum Insaf8 and Marianthi-Anna Kioumourtzoglou9, (1)Icahn School of Medicine at Mount Sinai, Environmental Medicine and Public Health, New York, NY, United States, (2)Boston University School of Public Health, Boston, United States, (3)Columbia University, Palisades, NY, United States, (4)Columbia University of New York, New York, United States, (5)Columbia University in the City of New York, Lamont-Doherty Earth Observatory, New York, NY, United States, (6)Emory University, Atlanta, GA, United States, (7)Emory University, Gangarosa Department of Environmental Health, Atlanta, GA, United States, (8)New York State Department of Health, Center for Environmental Health, Albany, NY, United States, (9)Columbia University of New York, Mailman School of Public Health, New York, United States
Recorded Presentation
Abstract:
Background: In recent years, there has been an increasing use of air pollution prediction models in epidemiologic studies for exposure assessment even in areas without monitoring stations. To date, most studies have assumed that a single prediction model is correct. However, an evaluation of the sensitivity of the estimated effects to the choice of exposure model is lacking.

Methods: We obtained county-level daily cardiovascular (CVD) admission counts from the New York (NY) Statewide Planning and Resources Cooperative System (SPARCS) and four sets of fine particulate matter (PM2.5) spatio-temporal predictions (2002-2012), including the Community Multi-scale Air Quality Model (CMAQ) and data from the Centers for Disease Control and Prevention’s Wide-ranging Online Data for Epidemiologic Research (CDC WONDER), among others. We employed overdispersed Poisson models to investigate the relationship between daily PM2.5 levels and CVD admissions, adjusting for potential confounders, separately for each state-wide PM2.5 dataset. We also repeated our analyses using PM2.5 measured at monitoring stations available in 18 of 62 counties.

Results: For three out of the four PM2.5 datasets, we observed positive, significant associations between PM2.5 and CVD. Across the modeled exposure estimates, effect estimates ranged from 0.23% (95%CI: -0.06, 0.53%) to 0.88% (95%CI: 0.68, 1.08%) per 10 µg/m3 increase in daily PM2.5. In general, CMAQ PM2.5 yielded the highest effect estimates, while the CDC dataset yielded the lowest estimates. Still higher was the effect estimate using monitored concentrations 0.96% (95%CI: 0.62, 1.30%) for the subset of counties where these data were available.

Conclusions: Effect estimates varied by a factor of almost four across methods to model exposures, likely due to varying degrees of exposure measurement error. This highlights the need for improved exposure assessment that minimizes error and includes uncertainty characterization – including uncertainty due to exposure model choice – and propagation into the health models. Nonetheless, we observed a consistently harmful association between PM2.5 and CVD admissions, regardless of model choice.

<< Previous Abstract | Next Abstract >>