BACKGROUND: Wildfires are increasingly common in the Western U.S. and pose a threat to human health. Studies have shown associations between wildfire smoke and respiratory outcomes, especially asthma emergency department (ED) visits. Some studies have also shown increased impact of wildfires in vulnerable populations. This study of a vulnerable population was prompted by an extensive wildfire in San Diego County in 2007, which resulted in numerous school and road closures and evacuation of 500,000 residents. We examined whether the effect of wildfire smoke on asthma ED visits is modified by neighborhood vulnerability characteristics.
METHODS: Insurance claims for asthma ED visits in the MediCal population (the Medicaid program serving low-income Californians) were collected from the California Department of Health Care Services during the wildfire period and reference periods before and after the fire, then aggregated by zip code. A smoke model developed by Michigan Tech Research Institute was used to estimate average daily particulate matter (PM2.5) to determine exposed and unexposed periods. Rate ratios of asthma ED visits in wildfire-exposed days compared to the reference days were calculated by zip code. Regression analyses were used to test effect modification using CalEnviroScreen 1.1, a vulnerability score based on environmental and social vulnerability indicators created by California’s Environmental Protection Agency.
RESULTS: 89 zip codes were included in the analysis. Zip codes were exposed to wildfire smoke for an average of 20 days out of the 160-day study period (N=645 asthma ED visits). The Mantel-Haenszel estimate of the rate ratio of asthma ED visits during the wildfire exposed period compared to reference was 1.44 (1.21, 1.72). The regression coefficient for the log of the rate ratio against the CalEnviroScreen Summary Score was -0.015 (P-value: 0.049). The CalEnviroScreen subcomponent ambient PM2.5 (non-wildfire) had a regression coefficient of -0.25 (P-value: 0.057); the subcomponent poverty had a regression coefficient of -0.010 (P-value: 0.055). A regression coefficient of 0 indicates no effect modification; negative slopes reflect decreasing asthma rate ratios with increasing vulnerability measures.
CONCLUSIONS: This study confirms a strong increase in asthma ED visits during the wildfires among a vulnerable population. Interestingly, these results suggest that MediCal recipients in more vulnerable areas appeared to use the ED less frequently for asthma during this fire than those in neighborhoods rated more positively in terms of environmental and social metrics. Further investigation of barriers to care and other factors are needed to better understand these results.