The Department of Statistics at Iowa State University hosted a seminar featuring Dr. Nathan Wikle from the University of Iowa. The seminar, titled "Causal inference for environmental health data: Estimating Causal Effects in the Presence of Spatial Interference," addressed the significant challenges inherent in drawing causal conclusions from spatial environmental data.
One pertinent application discussed was the estimation of the effectiveness of air quality interventions at coal-fired power plants in reducing adverse health outcomes. Wikle's research focused on pediatric asthma emergency department (ED) visits and Medicare all-cause mortality in Texas in 2016. The analysis incorporated a Bayesian, spatial mechanistic model for interference mapping, coupled with a flexible non-parametric outcome model to account for uncertainty in the structure of interference.
Despite evidence suggesting a potential reduction in asthma ED visits and all-cause mortality following emissions controls at upwind power plants, the seminar underscored the challenges of accounting for uncertainty in interference. The results, while suggestive, were largely inconclusive due to the inherent complexities in the spatial dynamics of pollution exposure.
Nathan Wikle, an Assistant Professor at the University of Iowa's Department of Statistics and Actuarial Science, led the seminar. With a background in spatial and spatio-temporal statistics, casual inference with interference, Bayesian hierarchical modeling, and statistical inference for dynamical systems, Wikle's research is motivated by pressing issues in environmental science, environmental health, and epidemiology. His presentation shed light on the intricate interplay between spatial dynamics and causal inference, offering insights crucial for addressing complex environmental health challenges.