Pilot study of publicly available data to evaluate the relationship between forest fires and emergency department visits due to asthma in the state of California

Introduction

Forest fires are devastating events causing widespread damage and contributing enormous environmental pollution^1,2. According to the National Interagency Fire Center, in 2017 there were 9,988 fires within the state of California. While the immediate effects of forest fires are well known, the lingering implications are less apparent. As some of the main by-products of forest fires are CO₂ and PM_2.5^3,4, questions remain with regards to how forest fires contribute to the incidence of respiratory diseases.

Asthma is a chronic respiratory illness characterized by the constriction of the bronchi, leading to difficult breathing and sometimes irreversible respiratory tissue remodeling⁵. According to the US Center of Disease and Control and Prevention, there is an asthma prevalence of 7.7% among the population in California. Because asthma symptoms can be exacerbated by airborne pollutants, such as CO₂ and PM_2.5^6,7, we hypothesized that an increase in acres of forest burned would be linked to increased emergency department (ED) visits due to asthma.

Methods

Statistical analysis

Statistical analysis was performed with R (version 3.5.0, https://www.r-project.org/). California’s counties were grouped into regions (Figure 1). Given the size of the state and no present guidelines for grouping of each county, we grouped counties by proximity into North, Coastal, Central, Motherload, and South regions. This classification was carried out within the R script included in Supplementary File 1. A pairwise t-test, with Bonferroni adjustment, was implemented to compare both the acres burned and the asthma rate, respectively, per region. A linear regression model, adjusted by year and region, was elaborated to evaluate the relationship between acres burned and ED asthma visits. p < 0.05 was considered statistically significant.

Figure 1. Counties of California grouped into the regions used by the present study.

Results

Each county was assigned to a region based on geography and proximity. As existing maps were unable to distribute California’s counties evenly, we used a breakdown of regions (Figure 1). Table 1 provides the mean acres burned and mean asthma rates per region. The ED asthma visits rates were above 50% except for the South and Motherload regions. A pairwise t-test with Bonferroni adjustment did not detect statistically significant differences in acres burned between the regions (p = 0.38 to 1.0; Table 2).

On the contrary, a statistically significant difference (p = 0.003) was found between the South (48.8%) and the Motherload (43.0%) region with regards to asthma rates. A linear regression model, adjusting for years and region, was employed to evaluate the magnitude of relationship between acres burned and ED asthma visits (Figure 2). In this model, and although statistically significant (p = 0.005), forest fires only explained less than 5% of the variability of the ED asthma visits between the years 2006 to 2015 (r² = 0.047). These results suggest that although regions within California may differ in ED asthma visits, forest fires did not explain these differences.

Figure 2. Asthma rate per acres burned in California counties between 2006 and 2015.

Discussion

In this study, data on forest fires within the state of California was compiled to address the potential relationship with ED asthma visits. As per our linear regression model, there was no link between the ED asthma visits and acres of forest burned. Future research should account for the additional variables related to asthma as well as forest fires.

In the current study, forest fires barely explained the ED visits from asthma during the years 2006 to 2015. It is possible that these ED asthma visits may have resulted from other stimuli, such as biological and non-biological indoor and outdoor air pollution besides forest fires^8–10. Also, California participates in different programs that may reduce ED visits like the National Asthma Control Initiative, which was developed by the US National Institute of Health.

Interestingly, asthma rates varied across all regions. The Motherload region had significantly lower mean asthma rates than the North region. This finding further warrants additional investigations into the asthma rates within states susceptible to natural disasters and diverse geographic and topography, such as the state of California. Future studies may shed light on the interplay of forest fires, environmental variables, and chronic respiratory diseases, such as asthma.

Limitations came in several forms. First, fires are known to often cross county lines, which makes it difficult to assign appropriate values per county. Next, the Alpine county was not included in the dataset due to the lack of data relating to asthma. News coverage on forest fires may warn the population of areas to avoid, which potentially reduces the number of ED visits and could explain the decline in asthma rates as acres of forest burned increase (Figure 2) because some counties did not report forest fires above the 300 acres threshold. Also, the data on asthma rates was available by year, not by month, thus reducing our ability to match forest fires and asthma rates monthly. Finally, it is possible that ED asthma visits may also be modified by another variable; therefore, tracking the sale and consumption of asthma medication could provide more reliable information of asthma exacerbation rates as the result of forest fires.

In summary, findings from the current study warrant examination of additional variables that could potentially contribute to environmental air pollution, besides forest fires, and are also linked to exacerbations of asthma. These include sociodemographics, medication sales due to asthma, and others that could shed light on the activities that the people of California implement during forest fires events.

Data availability

F1000Research Dataset 1: Forest fires acres burned and emergency department visits due to asthma for California, January-December 2005–2015. DOI, 10.5256/f1000research.15839.d213397¹¹

Data for emergency department visits due to asthma: http://www.cehtp.org/page/asthma/query

Data for forest fires: http://www.fire.ca.gov/fire_protection/fire_protection_fire_info_redbooks

The full dataset is available as Dataset 1.

R code for performing the analysis is available in Supplementary File 1.