News & Events

Biostatistics professors contribute to ground-breaking air pollution research

A recent article published in JAMA by the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air) research team found that air pollution can accelerate the progression of emphysema of the lung as much as smoking a pack of cigarettes a day.

UW Associate Professor of Biostatistics Adam Szpiro, a study co-author, talks about the role he and his colleague Professor of Biostatistics Lianne Sheppard played in this ground-breaking research.

Adam Szpiro and Lianne Sheppard
Adam Szpiro and Lianne Sheppard

What was your area of focus within this study and what challenges did you face?

Our primary focus was on exposure modeling and statistical analysis. One of the biggest challenges in studying the health effects of air pollution exposure is that we cannot measure each study subject’s exposure directly.  In MESA Air, we have a unique opportunity to learn about differences in exposure based on people’s locations, down to a resolution of tens of meters. 

We do this by building state-of-the art spatio-temporal statistical models, that is models that consider not only the locations but the times at which measurements are taken. These models incorporate the Environmental Protection Agency’s (EPA’s) regulatory air pollution monitoring data and additional monitoring data collected as part of MESA Air with the specific goal of better characterizing pollution levels at the kinds of locations where people spend time.  Ozone presents particularly difficult challenges since its levels vary dramatically by season, so we leveraged EPA’s dispersion models to further improve our spatio-temporal model. 

Another challenge is that air pollution exposure levels vary in time and we were interested in changes in lung function over multiple repeated scans. So in order to isolate the effect of the air pollution exposure, we used random effects regression models that allowed for separate short-term (i.e., transient) and long-term (i.e., cross-sectional and longitudinal) associations with lung function.

What’s it like to work with the MESA Air project?

I’ve been involved in the MESA Air project since 2006 and one of the things that makes it such a rewarding experience is that we have a cohesive interdisciplinary team that works together throughout all aspects of the project – from designing the study, through collecting the data, to finally analyzing the data and interpreting the results.