UC San Diego

Wildfire smoke adversely impacts respiratory health as fine particles can penetrate deeply into the lungs. Epidemiological studies of differential impacts typically target population subgroups in terms of vulnerability to wildfire smoke. Such information is useful to customize smoke warnings and mitigation actions for specific groups of individuals. In addition to individual vulnerability, it is also important to assess spatial patterns of health impacts to identify vulnerable communities and tailor public health actions during wildfire smoke events.

Methods

We assess the spatiotemporal variation in respiratory hospitalizations in San Diego County during a set of major wildfires in 2007, which led to a substantial public health burden. We propose a spatial within-community matched design analysis, adapted to the study of wildfire impacts, coupled with a Bayesian Hierarchical Model, that explicitly considers the spatial variation of respiratory health associated with smoke exposure, compared to reference periods before and after wildfires. We estimate the signal-to-noise ratio to ultimately gauge the precision of the Bayesian model output.

Results

We find the highest excess hospitalizations in areas covered by smoke, mainly ZIP codes contained by and immediately downwind of wildfire perimeters, and that excess hospitalizations tend to follow the distribution of smoke plumes across space (ZIP codes) and time (days).

Conclusions

Analyzing the spatiotemporal evolution of exposure to wildfire smoke is necessary due to variations in smoke plume extent, particularly in this region where the most damaging wildfires are associated with strong wind conditions.