Introduction
The January 2025 Los Angeles (LA) wildland-urban interface (WUI) wildfires, including the Palisades and Eaton Fires, represent a significant environmental disaster, resulting in widespread evacuations and extensive structural damage within a densely populated region. The Palisades and Eaton Fires, both of which began on January 7, 2025, and were not contained until January 31, 2025, burned over 37,000 acres of land and destroyed or damaged over 18,000 structures 1 and news outlets reported that at one point almost 200,000 people were under evacuation orders.Reference Stelloh, Lenthang, Cohen and Helsel 2 Beyond the immediate physical and economic devastation, wildfires can have profound and lasting impacts on the mental well-being of affected populations.
Previous research has demonstrated a link between wildfire exposure and adverse mental health outcomes in adults. Studies indicate a notable increase in the acute and longitudinal prevalence of post-traumatic stress disorder (PTSD), depression, and anxiety disorders among those who have experienced wildfire events.Reference To, Eboreime and Agyapong 3 –Reference Eisenman and Galway 5 However, few studies collected pre-fire data on participants and thus could not control for pre-fire mental health status. Therefore, it is unclear whether the mental health outcomes identified in these studies were a result of the fires or were pre-existing. The few studies that did account for pre-fire mental health were assessed retrospectively and relied on self-report indicators, which may miss undiagnosed baseline variations in mental health.Reference Agyapong, Ritchie and Brown 6
This study investigates the impact of the January 2025 wildfires on the mental health of individuals residing in Southern California. Specifically, we examined whether the experience of evacuating influenced residents’ subsequent mental health outcomes. We hypothesized that after controlling for demographic covariates and pre-fire mental health, those who evacuated would show poorer mental health status relative to those who did not evacuate.
Methods
Participants
This study included current or former students, staff, and faculty from the University of Southern California (USC) in Los Angeles, California, who had participated in a study of COVID-19 and mental health in 2021-2022, called the Trojan Pandemic Response Initiative (TPRI).Reference Lee, Hu and Kawaguchi 7
Procedure
In March-April 2025, 6,889 previous TPRI participants were invited to participate in this study. Participants were eligible if they still lived in California in 2025. Of these, 1,039 responded (15.1%), with 856 (82.4%) meeting eligibility criteria and provided written electronic informed consent. The low response rate is likely due to outdated contact information because USC automatically disables email accounts after students and staff leave; the most recent contact with the TPRI participants was in 2022 via their university email addresses. To assess potential non-response bias, we tested for differences in demographic characteristics between respondents and non-respondents (Table S1). A total of 739 participants (86.3%) completed the online survey. Participants received a $10 gift card for their participation. The University of Southern California Institutional Review Board approved this study (UP-25-00077).
Measures
Evacuation status was self-reported with options to indicate if they did not evacuate, evacuated and returned, did not return due to home destruction, or did not return for other reasons. Those who selected “Did not evacuate” (N = 644, 79.5%) were classified accordingly, and all other responses (N = 166, 20.5%) were classified as “Evacuated.” Depressive symptoms were assessed using the Center of Epidemiologic Studies Depression Scale, 10-item version (CES-D-10), with a score of 10 or more indicating high risk for moderate to severe depression (α = 0.87).Reference Kohout, Berkman, Evans and Cornoni-Huntley 8 Generalized anxiety symptoms were assessed using the Generalized Anxiety Disorder 7-item (GAD-7), with a score of 10 or more indicating high risk for moderate to severe anxiety (α = 0.94).Reference Spitzer, Kroenke, Williams and Löwe 9 PTSD was assessed using the Primary Care PTSD Screen for DSM-5 (PC-PTSD-5), with a score of 3 or more indicating high risk for PTSD (α = 0.80).Reference Prins, Bovin and Smolenski 10 Demographic information and baseline CESD-10 and GAD-7 scores were obtained from a previous TPRI survey conducted in 2022 (the TPRI survey did not assess PTSD).
Data Analysis
Bivariate logistic regression models examined the effects of evacuation status on depression, anxiety, and PTSD. Multivariable models adjusted for demographic variables and baseline mental health indicators. All analyses were conducted in STATA 18 with an alpha level of 0.05.
Results
Of the 739 participants, 104 contained missing data for one or more key study variables, resulting in an analytical sample of n = 635. The analytical sample was primarily female (71.2%), White (34.9%), and Asian (31.5%), 18-29 age group (43.3%), and 18.7% evacuated. The distribution of survey respondents’ Zip Code Tabulation Areas (ZCTAs) across Southern California overlayed with the January 2025 wildfire perimeters are presented in Figure 1 (Figure 1). Risk of moderate to severe depression, anxiety, and PTSD were 29.6%, 15.7%, and 11.3%, respectively. Compared to the original TPRI sample, the current study sample was significantly older (mean age = 35.1 vs 29.3; P <0.001), more female or “other” gender (71.2% vs 63.8%; P <0.001), and had different racial/ethnic compositions (P = 0.001), notably a higher proportion of White and Hispanic participants (Table S1).
Figure 1. Distribution of survey respondents’ Zip Code Tabulation Areas (ZCTAs) across Southern California, overlayed with the January 2025 wildfire perimeters. *Stars indicate location of USC Campuses (University Park Campus and Health Science Campus).
In bivariate analyses, evacuation status was associated with higher risk of depression (odds ratios (OR) = 1.62 [1.10-2.38]) and PTSD (OR = 2.08 [1.26-3.43]), but not anxiety (OR = 0.99 [0.60-1.63]). Evacuation remained associated with higher risk of depression (adjusted odds ratios (AOR) = 1.75 [1.08-2.85]) and PTSD (AOR = 2.44 [1.36-4.35]) after controlling for demographic covariates and baseline mental health (Table 1). These results indicate participants who evacuated had 75% higher odds of screening positive for moderate to severe depression and 144% higher odds of screening positive for high-risk PTSD compared to non-evacuated participants, after controlling for demographics and pre-fire mental health status. A sensitivity analysis was conducted by treating depression, anxiety, and PTSD scores as continuous outcomes and the key results remained stable (Table S2).
Table 1. Associations of evacuation status and demographics with depression, anxiety, and PTSD following the 2025 Los Angeles Wildfires
* Note: Group size (n) and row percentage (%) are reported. Unadjusted odds ratios (OR) and 95% confidence intervals are reported for bivariate logistic regression models. Adjusted odds ratios (AOR) and 95% confidence intervals are reported for multivariable logistic regression models examining evacuation status on depression, anxiety, and PTSD, controlling for gender, race/ethnicity, age-group, and pre-fire mental health. ORs and AORs with 95% confidence intervals that do not include one are bolded. Depression was measured using the Center for Epidemiological Studies Depression 10-item version (CESD-10) and using a cutoff of 10. Anxiety was measured by the Generalized Anxiety Disorder 7-item version (GAD-7) and using a cutoff of 10. PTSD was measured by the Primary Care PTSD Screen for DSM-5 (PC-PTSD-5), with a cutoff of 3. “Other” gender includes transgender, non-binary, and other and was grouped with Woman due to small sample size. “Other” race/ethnicity includes African American, American Indian or Alaskan Native, Native Hawaiian or Pacific Islander, Middle Eastern, North African, Arab/Arab American, multi-racial, or other, and were grouped together due to small sample sizes. Pre-fire depression (measured by CESD-10 and using a cutoff of 10) and pre-fire anxiety (measured by GAD-7 and using a cutoff of 10) are from previous surveys conducted in 2022. Pre-fire PTSD was not measured so pre-fire anxiety was used as a pre-fire mental health indicator in the multivariable PTSD model; CESD-10 was not included in this model because of collinearity.
Limitations
This study presents emerging findings on the mental health impacts of the January 2025 LA wildfires. Due to the rapid initiation of this study following the wildfires, this study employed convenience sampling leveraging a previous cohort of university-affiliated students, staff, and faculty who were still living in the greater Los Angeles area during the 2025 LA wildfires. Therefore, findings may not generalize to non-university-affiliated populations, including children, retired older adults, or low-SES populations. This study also does not primarily focus on the residents who were most impacted by the fires (i.e., from Altadena and Palisades). Of the 119 participants who reported evacuating, 101 (84.9%) returned home by the time of the survey, 14 (11.8%) did not return home because of property damage, and 4 (3.4%) did not return home for another reason, suggesting few participants from our study experienced the most extreme impacts from the wildfires and we did not have enough participants who reported property damage to examine their mental health effects alone. Future studies should assess the effects of wildfires on these other groups, who might be even more vulnerable. While this is an important limitation to recognize, establishing a representative cohort of Southern California residents in the immediate aftermath of a wildfire poses significant challenges, making alternative sampling strategies a necessary consideration. For this reason, the study team decided to re-recruit an existing cohort of Southern California residents who had pre-fire mental health data available and were able to be rapidly contacted post-fire. However, the pre-fire mental health data was collected in 2022, when the COVID-19 pandemic may have impacted participants’ baseline mental health levels. This context complicates interpretation of post-wildfire changes, as some observed effects may reflect lingering pandemic impacts rather than changes attributable solely to wildfire exposure. Finally, we invited previous TPRI participants to take part in our study; however, these participants had not been contacted in over 2 years, and it is likely that many had their university email address deactivated upon graduating or leaving USC. This led to a low response rate to our initial study invitation. Comparison of our study sample to the original TPRI sample (Table S1) demonstrated differences in the demographic makeup between the two samples, suggesting our results may be impacted by non-response bias.
Discussion
Approximately 2-3 months after the LA wildfires, participants who evacuated their homes had significantly higher risk of depression and PTSD, relative to those who did not evacuate. These findings remained significant even after controlling for the participants’ pre-fire mental health, indicating that the experience of evacuating one’s home in and of itself was associated with higher risk of depression and PTSD, independent of prior mental health status. These results are consistent with prior research demonstrating elevated risk of depression and PTSD following wildfire exposure among adults.Reference To, Eboreime and Agyapong 3 –Reference Eisenman and Galway 5
Interestingly, we did not find significant differences in anxiety by evacuation status which is contradictory to other studies on the mental health impact of wildfires.Reference To, Eboreime and Agyapong 3 –Reference Eisenman and Galway 5 In a recent review article on the impact of wildfires on mental health, the authors note that wildfires can increase risk of anxiety by creating emotional distress due to the uncertainty that follows significant losses and damage to oneself, loved ones, property, belongings, and livelihood.Reference To, Eboreime and Agyapong 3 Our sample, however, had few participants whose homes were directly damaged or destroyed by the fires; thus our sample may not have experienced as great an impact on their risk of anxiety. Additionally, participants in our study were surveyed 2-3 months following the January wildfires, by which time most evacuees had returned home. While elevated symptoms of depression and PTSD among evacuees seem to persist after returning home, anxiety may resolve more quickly due to decreased feelings of immediate threat and uncertainty. Future research should examine how risk of anxiety changes over time following wildfire exposure and examine differences between those who experienced significant loss and those who were impacted but were able to return home.
As WUI fires increase in frequency, the number of people who are forced to evacuate during wildfire events also increases. As shown during the 2025 LA wildfires, almost 200,000 residents were under an evacuation order, highlighting the impact WUI fires have on urban and suburban communities. The findings from this study highlight the need for mental health screening and interventions after natural disasters for not only those directly impacted (i.e., with property damage or loss) but also for those living in close proximity and who went through the stress of evacuating. This information can be used by policy makers, local health departments, and clinicians in their disaster management protocols and practices to ensure mental health support is available in addition to other important resources for wildfire evacuees. As suggested by To, Eboreime, and Agyapong in their recent scoping review on the impact of wildfires on mental health, digital interventions, such as supportive text message programs, are a promising tool to help mitigate adverse mental health symptoms following disasters and could be utilized and evaluated among wildfire evacuees.Reference To, Eboreime and Agyapong 3 At a university or large institution level, these results demonstrate the importance of disaster planning and response to support members (whether employees or students) after devastating disasters. Even if the institution is not directly impacted by the disaster, institution members may still be impacted and need additional support, such as mental health counseling, financial and emotional support, and other accommodations.
Conclusions
Evacuating was significantly associated with increased risk of depression and PTSD after controlling for pre-fire mental health, underscoring the psychological impact of wildfires. Future wildfire preparedness protocols should incorporate mental health support for evacuees.
Supplementary material
To view supplementary material for this article, please visit http://doi.org/10.1017/dmp.2025.10257.
Acknowledgments
We would like to thank our participants for contributing to this project in the aftermath of the devastating wildfires, especially those who were greatly impacted.
Author contribution
Ryan Lee: Conceptualization, investigation, methodology, formal analysis, data curation, writing—original draft preparation—and project administration; Jennifer B. Unger: conceptualization, methodology, writing—reviewing and editing—supervision, and funding acquisition; Daniel W. Soto: conceptualization, methodology, writing—reviewing and editing—supervision, and funding acquisition; Eric S. Kawaguchi: conceptualization, methodology, data curation, and writing—reviewing and editing; Myles Cockburn: methodology and writing—reviewing and editing; Shuvan Paul: methodology, data curation, writing—reviewing and editing; and Frank D. Gilliland: conceptualization, methodology, writing—reviewing and editing—supervision, and funding acquisition.
Funding statement
This research was funded by The Hastings Foundation, The Houser Foundation, The Southern California Clinical and Translational Science Institute (SC CTSI), The Southern California Environmental Health Sciences Center (SCEHSC), and The USC Office of Research Innovation President’s Sustainability Award.
Competing interests
The authors declare no conflict of interest.
Open access
