A dangerous combination: how co-occurring floods and blackouts increase cardiovascular hospitalizations

Extreme weather events (EWEs) are becoming more frequent, but few have assessed the joint effects of floodings and power outages (POs) on health.

We aimed to quantify the independent and joint effects of flooding and POs on hospitalizations for cardiovascular diseases (CVDs).

Associations between flooding/POs and CVD hospitalizations in New York State (NYS) were estimated using distributed lag nonlinear models (DLNMs) within a quasi-Poisson regression framework, adjusting for PM_2.5, weather factors and time-varying factors. We also evaluated length of stay, costs and comorbidities, thresholds, and compared the before and during COVID-19. Stratified analyses were conducted by CVD subtypes, demographics, season, and urbanicity.

Flooding and POs were independently associated with total and specific CVDs, but the impact was stronger when they co-occurred (highest RR = 1.09, 95% CI: 1.03-1.15). CVD risk increased when flooding persisted for >3 days and when POs exceeded 15.3%. Flooding/POs were associated with longer hospital stays and higher comorbidities. Flooding/PO joint occurrence showed larger risk of CVD hospitalization during COVID-19, although not statistically significant. Stronger joint effects were observed on subtypes of hypertension, stroke, and ischemic heart disease, people over 65 years, those with Medicaid or no insurance coverage, rural residents, and during winter/spring.

This study highlights the health impacts of co-occurring environmental hazards of flooding and POs, revealing seasonal and temporal patterns and identifying vulnerable subgroups. Targeting susceptible population and deploying preparedness efforts in critical time windows may help mitigate related excess CVD burden.

Our study provides new evidence on the impact of co-occurring flooding and POs on CVD hospitalizations. We found that their joint occurring is associated with greater increases in CVD hospitalizations than either exposure alone, disproportionately affecting older adults, socioeconomically vulnerable populations, and rural communities. Distinct seasonal, temporal, and threshold patterns were identified for flooding and POs. These findings offer useful information for emergency management agencies to reduce CVD burden during compound hazard events.