Asthma Risk and the Co-occurrence of Thunderstorms and Elevated Pollen: Measuring the Strength of Association, Investigating the Effects in Subgroups, and Leveraging Data Across Large Areas.

Abstract

Abstract Severe asthma has been shown to occur in the combined presence of high pollen and thunderstorms, termed thunderstorm asthma. Previous research has focused on rare ‘epidemic’ events, such as in Melbourne, Australia 2016 where emergency room usage was 900% higher during a single thunderstorm asthma event. In my dissertation, we investigate thunderstorm asthma conditions in the Twin-Cities metro region, Minnesota, U.S.A., using detailed exposure estimates from a network of weather sensors along with daily pollen records, and asthma-related emergency department (ED) visits from 2007-2018. In manuscript 1, we investigate the association between asthma-related ED visits and thunderstorm asthma conditions within a study radius of 20 miles from the Minneapolis-St. Paul (MSP) airport using a time series model approach. We evaluated risk for the entire study area combined and at the individual zip code level to investigate potential effect heterogeneity and spatial auto-correlation. We find a 1.05 (95% CI: 1.012,1.083) times higher risk on the day of a thunderstorm asthma event with no evidence of spatial autocorrelation or effect heterogeneity. In manuscript 2, we investigate relative and absolute risk disparities of thunderstorm asthma by age and sex subpopulations. We find evidence that thunderstorm asthma has impacts across the life course for men and women, with variation in risk by individual age-sex groups contrary to typical baseline patterns of severe asthma incidence. For males 18-44 the RR of severe asthma was 1.123 times higher on storm days (95% CI: 1.042, 1.211) compared to non-storm days,] with 1.098 times higher risk of incident severe asthma on storm days for females over 45 (95% CI: 1.020, 1.181) compared to non-storm days. In manuscript 3, we investigate the ability to leverage exposure information from a single pollen site in MSP and land-use covariates to estimate thunderstorm asthma associations at 19 communities across the state of Minnesota. Using meta-regression, we find a positive association between deciduous tree and grassland land cover with the thunderstorm asthma effect size, and we find an attenuation of the thunderstorm asthma risk as distance increases from the MSP pollen site.