Valuation of Community Resilience to the Health Impacts of Extreme Weather

Background: A growing body of literature documents the health impacts of extreme weather events and a projected acceleration of climate change is expected to increase the frequency and severity of these events. Government agencies have called for enhancing citizens’ resilience as a way to prepare populations in advance of extreme weather events. However, there is little empirical evidence to link measures of resilience to extreme weather-related health outcomes, hindering full consideration of health from the evaluation of resilience strategies and policies. In late October 2012, Hurricane Sandy hit the northeastern United States coastline with devastating impacts. Within this context, we examine whether the spatial patterning of morbidity due to the storm can be explained by community-level resilience characteristics.

Methods: Emergency department (ED) visit data for 16 New Jersey counties, which were highly impacted by Sandy, was obtained from the Healthcare Cost and Utilization Project. A two-stage Bayesian hierarchical model was used to estimate each county’s increase in risk of ED visits for chronic obstructive pulmonary disease (COPD), asthma, chronic renal disease, and myocardial infarction (MI) due to Hurricane Sandy. We examined variation in county-level associations according to the county’s resilience, defined by the Baseline Resilience Indicators for Communities (BRIC) index, after adjusting for storm severity. We estimated the number of excess ED visits attributable to Sandy and the reduction in excess ED visits and healthcare costs that could be achieved with increased community resilience.

Results: In the months following the storm, across highly impacted counties, there was an increase in ED visits for all causes we examined, ranging from 5.6% (95% PI: 3.0, 8.4) for chronic renal disease to 18.9% (95% PI: 15.9, 22.0) for COPD. The effect of the storm on ED visits decreased by 3.1% (95% PI: 1.1, 5.0) for COPD and by 3.0% (95% PI: 0.6, 5.3) for chronic renal disease for a standard deviation increase in the BRIC index. County level variation in the BRIC index did not explain variation in MI or asthma ED visits. For a county with a median level of increased risk, we estimate that Hurricane Sandy resulted in 127.3 (95% PI: 101.7, 152.0) excess COPD ED visits. If this association is causal, a standard deviation increase in community resilience could have reduced the excess in COPD ED visits by 22.9 visits, corresponding to a potential healthcare cost savings of over $1 million dollars.

Conclusions: Community-level resilience measures explain some variation in health impacts due to extreme weather events. Understanding this relationship may be helpful in the valuation of community resilience investments. Future work will investigate the domains of resilience that contribute to this effect in order to understand the types of program and policy interventions that can contribute to extreme weather-related healthcare cost savings.