Our assessment examined employer-based health insurance claims for asthma for a period between June 1st and September 30th, 2015, covering 11 states in the western U.S. We linked approximately 450,000 asthma medication refills and 85,000 outpatient visit records with population-weighted daily average smoke concentrations estimated from a geographically weighted ridge regression (GWR) model. The GWR model blends information from surface monitors, a chemical-weather model, and satellite data products to estimate smoke particulate matter ≤2.5 microns in diameter (PM2.5). A conditional quasi-Poisson regression with month/day/place strata in combination with a distributed lag non-linear model (DLNM), was used to measure the overall cumulative effect of smoke concentrations. This regression framework explored how the associations were distributed over a 7-day lag period and included additional predictors, such as, cross-basis terms for daily measures of ozone, relative humidity, and temperature.
The estimated (95% Confidence Interval [CI]) risk ratios [RRs]) associated with a 10 µg/m3 increase in smoke PM2.5 exposure for asthma medication refills and outpatient visits were 1.066 (1.047 – 1.085) and 1.108 (1.065 – 1.152), respectively. RRs estimated on the day of smoke exposure were the highest, but delayed effects of smoke on these health outcomes were statistically significant and persisted until 5 days following the smoke exposure.
Asthma medication refills and outpatient visits are important morbidity measures to consider during wildfires. A comprehensive strategy to mitigate adverse health impacts associated with wildfire disasters necessitates a thorough understanding of population-level exposures and health impacts associated with surface smoke PM2.5 concentrations.