In 2018, 2109 wildland fires burned nearly 900,000 acres in the U.S. state of Oregon (National Interagency Fire Center, 2018). Major fires included the Klondike Fire, which began with a lightning strike on July 15, 2018, burned over 175,000 acres, and resulted in an estimated cost of US$ 104.5 million and the Substation Fire, which started on private land on July 17, 2018, destroyed a historic home, and resulted in one death directly attributed to the fire (Gabbert, 2018; National Interagency Coordination Center, 2018). Both the size and the frequency of wildfires such as these challenge public health and safety officials to mobilize firefighting resources, alert the public of the dangers of wildfires and smoke, and put in place evidence-based strategies to keep community members safe from fire, smoke, ash, and debris.

The impacts of experiencing, responding to, and recovering from wildfires are wide-ranging. Wildfires can cause burns and smoke-related injuries that can range from mild to fatal (Shusterman et al., 1993; Sonoma County Department of Health Services, 2018). Wildfires generate enormous quantities of smoke that can linger over areas within and near the fire perimeter or drift far from the original burn sites. Respiratory health effects of wildfire smoke inhalation are well-described and include exacerbations of asthma, chronic obstructive pulmonary disease, and other respiratory health conditions (Shusterman et al., 1993; Cascio, 2018; Rice et al., 2021; Reid et al., 2016; Liu et al., 2015; Chen et al., 2021). Exposure to wildfire smoke has also been negatively associated with attention (Cleland et al., 2022) and positively associated with adverse pregnancy and birth outcomes (Abdo et al., 2019; Heft-Neal et al., 2022; Evans et al., 2022) and sleep-related symptoms (Rodney et al., 2021).

Epidemiologic studies suggest that the impact on mental health of having survived a wildfire is substantial (Shusterman et al., 1993; Rodney et al., 2021; Eisenman et al., 2021; Zhang et al., 2022; Moosavi et al., 2019; Mao et al., 2022; To et al., 2021). In 2018, investigators assessed the health and well-being of residents one year after a major wildfire in Sonoma County, California and found that 37% of residents with a history of depression or other emotional health problem reported that the conditions worsened in the year since the wildfires (Sonoma County Department of Health Services, 2018). Similarly, Rodney et al. (2021) conducted a cross-sectional survey in the Canberra region of Australia following the 2019–2020 bushfires and found that respondents reported anxiety, depression, and disrupted or poor sleep attributed to bushfire smoke (Rodney et al., 2021). In light of findings such as these, we conducted this study to assess associations between the duration of wildfire smoke exposure and the prevalence of self-reported symptoms affecting mental health among adults living in Oregon, a U.S. state regularly affected by wildfires, in 2018.

Characteristics of the 5807 BRFSS respondents and the weighted population estimate are shown in Table 1. Respondents came from each of Oregon’s 36 counties. Overall, over a quarter of adults in Oregon reported ever having been told that they have a depressive disorder (26.1%; 95% CI: 24.7, 27.5). This percentage was notably lower in adults 65+ years (18.9%; 95% CI: 16.7, 21.1), those who graduated from college or technical school (21.4%; 95% CI: 19.4, 23.5), adults employed for wages or self-employed (22.0%; 95% CI: 20.4, 23.6), and men (18.9%; 95% CI: 17.2, 20.7). The highest percentage of adults with depressive disorders occurred among adults who reported being out of work or unable to work (52.7%; 95% CI: 47.9, 57.5).

Percentages of adults in Oregon who self-reported experiencing specific symptoms more than half the days of the past two weeks ranged from 7.7% (95% CI: 6.8, 8.6) reporting feeling down, depressed, or hopeless to 10.4% (95% CI: 9.4, 11.4) reporting feeling nervous, anxious, or on edge (Table 2). Percentages who reported depressive disorders were considerably higher among adults who reported the four symptoms more than half the days of the past two weeks, than among adults who reported them never or several days of the past two weeks.

Overall, adults in Oregon in 2018 experienced an average of 17.7 days (range: 6–60 days) on which the smoke plume was categorized as heavy and 31.8 days (range: 15–74 days) on which it was categorized as medium or heavy (Table 3). The mean numbers of days with each category of smoke were notably higher in nonmetropolitan counties and notably lower among adults who graduated from college or technical school, non-Hispanic black adults, and adults in metropolitan areas.

Table 4 shows adjusted associations of the two metrics of exposure with ever having a depressive disorder and having any of the four symptoms more than half the days during the past two weeks. For each exposure metric, highest prevalences and PRs were generated in the longest duration smoke categories, though 95% CIs for all associations include unity. For example, compared with <2 weeks of exposure to heavy smoke in the past year, exposure to ≥4 weeks of heavy smoke was associated with a 4% increase in the prevalence of ever having a depressive disorder (PR: 1.04, 95% CI: 0.90, 1.20).

In adjusted analyses of the associations between the exposure metrics and each of the four symptoms, highest prevalences and PRs were again generated in the longest duration exposure categories (Table 5). Heavy smoke for ≥4 weeks in the past year, compared to <2 weeks in the past year, was associated with a 34% increase in the prevalence of feeling nervous, anxious, or on edge more than half the time during the past 2 weeks (PR: 1.34, 95% CI: 1.04, 1.72). Similarly, the highest categories of exposure to heavy smoke (≥4 weeks in the past year) and medium or heavy smoke (≥6 weeks in the past year) were associated with 29% and 30% increases, respectively, in the prevalence of being unable to stop or control worrying more than half the time during the past 2 weeks (heavy: PR: 1.29, 95% CI: 0.98, 1.70; medium or heavy: PR: 1.30, 95% CI: 1.03, 1.65).

Estimates generated by our first sensitivity analyses in which we restricted the exposure metrics to include only those smoke days that occurred May–October were identical or nearly identical in magnitude and precision to those generated by our main analyses (Supplemental Material Table S2). For our second sensitivity analysis in which we restricted the exposure metrics to include only those smoke days that occurred in the six months preceding the BRFSS interview date, the categories of the distributions of smoke duration were < 2, 2 to <4, and ≥ 4 weeks for both metrics. Like those of our main analysis, 95% CIs for the estimates generated by this sensitivity analysis largely included unity (Supplemental Material Table S3). Estimates generated by our third sensitivity analysis in which we dichotomized responses about symptoms affecting mental health were also similar to those generated by our main analysis (Supplemental Material Table S4). For several outcomes, associations with longer durations of exposure were attenuated and more precise. For example, the highest categories of exposure to heavy smoke and medium or heavy smoke were associated with 14% and 18% increases, respectively, in the prevalence of ever being unable to stop or control worrying during the past 2 weeks (heavy: PR: 1.14, 95% CI: 0.97, 1.33; medium or heavy: PR: 1.18, 95% CI: 1.03, 1.35).

In the U.S. state of Oregon, 4068 wildfires burned over 1.6 million acres in 2017–2018 (National Interagency Fire Center, 2018; National Interagency Fire Center, 2017). We conducted this analysis to assess the associations between exposure to wildfire smoke and symptoms affecting mental health among adults in Oregon in 2018 and found that 100% of respondents, representing 100% of the weighted population estimate of adults in Oregon, lived in a location in which medium or heavy smoke plumes were detected for at least 2 weeks in the past year. We also found that the highest duration categories of exposure to medium or heavy smoke were associated with selected symptoms during the past two weeks. These findings provide valuable insights into the burden of symptoms affecting mental health among adults living in wildfire-affected areas.

People impacted by wildfires include those living in neighborhoods directly impacted by wildfires, some of whom had their homes, work-places, or other places in their communities burned by wildfires or were advised to evacuate because of encroaching wildfires. Even people not directly threatened by wildfires can be impacted by wildfire smoke, as residents might be advised to cancel outdoor events, reduce the time they spend outdoors, and prepare themselves and their families for the possibility of evacuation. Community infrastructure can also be impacted, such as when public buildings are used as cleaner air shelters or when vehicle or airport traffic is hindered by wildfire smoke.

In 2018, a rapid needs assessment was conducted in Sonoma County, California to assess the health of residents one year after the 2017 Sonoma Complex Fires (Sonoma County Department of Health Services, 2018). Compared to the year before the fires, investigators observed increased post-fire prevalences of numerous symptoms affecting mental health, including anxiety or fear and feeling depressed, hopeless, or a loss of interest in normal activities. Especially traumatic experiences reported included seeing a direct threat to the life of oneself or a family member, being separated from a family member or unaware of their location or well-being during the wildfire, being trapped or delayed in evacuation, having a home destroyed, and experiencing or witnessing a serious injury (Sonoma County Department of Health Services, 2018). Our findings support and extend what was learned about symptoms affecting mental health following the 2017 Sonoma Complex Fires and other major wildfires by describing changes in the prevalence of self-reported symptoms with increasing durations of exposure to wildfire smoke. Together, these findings can be used by public health officials and emergency responders as they consider the range of resources that might be needed in communities affected by not only wildfires, but also by plumes of wildfire smoke.

For this analysis, we used BRFSS survey data and county-level smoke plume observations to estimate county-level wildfire smoke exposures for each BRFSS respondent. We generated county-level exposure estimates using NOAA’s publicly-available data, which include categorizations of smoke plume density based on automated fire detection, digitization of smoke plume images, and expert image analyst review (National Oceanic and Atmospheric Administration, 2021). Wildfire smoke plumes typically include a mixture of particulates, gases, and other components that vary depending on the fuel burned in the fire. Because of this variation in the composition of smoke plumes, there is no single ground-level measurement that can best validate the estimates of wildfire smoke exposure used in this analysis. Supplemental Material Fig. S2 shows daily averages of 24-h average concentrations of particulate matter ≤2.5 μm in diameter (PM_2.5) (U.S. Environmental Protection Agency, 2022) and Hazard Mapping System smoke density category in four regions of Oregon (Oregon Health Authority, 2022). In each of the four regions, peaks in PM_2.5, a major component of wildfire smoke, appear to coincide with increases in smoke density. Nonetheless, county-level wildfire smoke exposures do not reflect personal exposure to wildfire smoke or to the stressors associated with wildfires. In fact, wildfires can occur at the same time as droughts, heat waves, and other environmental events and while Hazard Mapping System data provide valuable information about wildfire smoke plumes, if other environmental events coincided with periods of peak wildfire smoke, then our analyses could have generated associations that could be due in part or entirely to other hazards or to combinations of hazards. Furthermore, these data do not include information about factors that may have affected personal exposure and that could be used to assess the validity of the wildfire smoke exposure metrics we assigned or the psychological stressors they might reflect. For example, our data did not include information about the amount of time respondents spent in their county of residence or the extent to which the wildfires or wildfire smoke affected the respondents’ activities. Overall, however, we do not envision that our results were affected by a systematic error attributable to the exposure assignment approach we used.

Similarly, our analyses do not account for other factors potentially associated with mental health status, including individuals’ educational attainment or county-level measurements of ambient air pollutants, temperature, precipitation, or other weather-related factors. The relationships between individuals’ mental health status and characteristics of the communities in which they live are complex and our data sources do not include information that might be used to assess them or their roles in the associations of wildfires and wildfire smoke with symptoms affecting mental health. Most notably, our analyses do not account for evacuations due to wildfires and we have no data with which to assess the extent that the associations observed here might be affected by the experience of evacuating because of a wildfire. Finally, our analyses assessed the relationship between the number of smoke days in the past year and the frequency of symptoms in the past two weeks and do not account for potential differences in responses to smoke events that occurred in the days or weeks before the interviews were conducted compared to those that occurred closer to one year before the interviews. Nonetheless, our findings provide evidence of the lasting toll of wildfire smoke exposure on mental health and raise the possibility that survey questions soliciting information about longer time periods or about the onset of symptoms might provide additional information about the occurrence of such symptoms following periods of wildfire smoke.

In summary, we applied satellite-derived estimates of exposure to wildfire smoke to BRFSS survey data to estimate associations between wildfire smoke and symptoms affecting mental health. We used the number of days with wildfire smoke in each county as a proxy for duration of stress attributable to living in a community affected by wildfires. We found that among adults in Oregon, selected symptoms were associated with extended durations of smoke. These findings highlight the burden of symptoms affecting mental health among adults living in communities affected by wildfire smoke and can be used to plan recovery efforts in and near communities affected by wildfires.