Importance: Hurricanes are associated with a wide range of adverse health effects in the general population and are increasing in frequency and severity due to global climate change. Due to prior military exposures and distinct sociodemographic characteristics, US veterans may be more vulnerable than the general population to negative health effects of hurricanes.
Objective: To evaluate whether acute care mental health visits among US veterans were associated with exposure to hurricanes.
Design, setting, and participants: This retrospective cohort study consisted of US veterans enrolled in Veteran Health Administration (VHA) primary care and whose home addresses were in regions that were affected by Hurricanes Sandy or Harvey. The Hurricane Sandy cohort included data from 960 394 veterans between October 29, 2011, and October 28, 2016, and the Hurricane Harvey cohort included data from 795 746 veterans between August 25, 2016, and August 24, 2021. Affected regions were determined from Federal Emergency Management Agency designations; patient-level clinical data were obtained through VHA electronic health records. Statistical analysis was conducted from October 2022 to November 2023.
Main outcomes and measures: The outcome of interest was acute care mental health visits, which included urgent care or emergency department visits or having an inpatient admission. The association between the outcome and exposure to hurricanes was estimated using Cox proportional hazards regression models with adjustments for multiple demographic and neighborhood characteristics, including neighborhood disadvantage, assessed using the Area Deprivation Index, and prior health status, assessed using the Care Assessments Need (CAN) score.
Results: Of the 960 394 veterans in the Hurricane Sandy cohort, the mean (SD) age was 63 (16) years, and 895 726 (93.3%) were men; of the 795 746 veterans in the Hurricane Harvey cohort, the mean (SD) age was 59 (16) years, and 715 460 (89.9%) were men. For both hurricane cohorts, increasing neighborhood disadvantage was associated with an increased hazard of having a subsequent acute care mental health visit; this association was more pronounced among healthier veterans (Sandy cohort: 1-year hazard ratio [HR], 1.62 [95% CI, 1.53-1.71]; Harvey cohort: 1-year HR, 1.64 [95% CI, 1.54-1.74]) compared with veterans who had severe illness (Sandy cohort: HR, 1.22 [95% CI, 1.17-1.26]; Harvey cohort: HR, 1.21 [95% CI, 1.13-1.30]). After adjusting for baseline data, the association between hurricane exposure and our outcomes of interest was largely null. For the Hurricane Sandy cohort, those who were flooded and received individual assistance demonstrated a 1-year HR of 1.31 (95% CI, 0.81-2.12) if their CAN score was 80 or above and 0.86 (95% CI, 0.60-1.22) for a CAN score below 80. Similarly, those affected by Hurricane Harvey who were flooded and received individual assistance had a 1-year HR of 1.06 (95% CI, 0.99-1.14) for a CAN score of 80 or above and a 1-year HR of 0.98 (95% CI, 0.91-1.05) for a CAN score less than 80.
Conclusions and relevance: This study used a novel and highly granular approach to examine the associations of hurricanes with mental health outcomes. These findings are the first to show that hurricane exposure was not associated with adverse mental health outcomes among US veterans and illustrate the importance of considering preexisting regional differences when assessing the associations of hurricanes with mental health outcomes. The results suggest that neighborhood characteristics, rather than exposure to a hurricane, are the dominant determinants of mental health outcomes.