Abstract
Hurricane Maria struck Puerto Rico on Sept 20, 2017, devastating the island. Controversy surrounded the official death toll, fuelled by estimates of excess mortality from academics and investigative journalists. We analysed all-cause excess mortality following the storm. We did a time-series analysis in Puerto Rico from September, 2017, to February, 2018. Mortality data were from the Puerto Rico Vital Statistics System. We developed two counterfactual scenarios to establish the population at risk. In the first scenario, the island's population was assumed to track the most recent census estimates. In the second scenario, we accounted for the large-scale population displacement. Expected mortality was projected for each scenario through over-dispersed log-linear regression from July, 2010, to August, 2017, taking into account changing distributions of age, sex, and municipal socioeconomic development, as well as both long-term and seasonal trends in mortality. Excess mortality was calculated as the difference between observed and expected deaths. Between September, 2017, and February, 2018, we estimated that 1191 excess deaths (95% CI 836-1544) occurred under the census scenario. Under the preferred displacement scenario, we estimated that 2975 excess deaths (95% CI 2658-3290) occurred during the same observation period. The ratio of observed to expected mortality was highest for individuals living in municipalities with the lowest socioeconomic development (1·43, 95% CI 1·39-1·46), and for men aged 65 years or older (1·33, 95% CI 1·30-1·37). Excess risk persisted in these groups throughout the observation period. Analysis of all-cause mortality with vital registration data allows for unbiased estimation of the impact of disasters associated with natural hazards and is useful for public health surveillance. It does not depend on certified cause of death, the basis for the official death toll in Puerto Rico. Although all sectors of Puerto Rican society were affected, recovery varied by municipal socioeconomic development and age groups. This finding calls for equitable disaster preparedness and response to protect vulnerable populations in disasters. Forensic Science Bureau, Department of Public Safety, and Milken Institute School of Public Health of The George Washington University (Washington, DC, USA).
Highlights
On Sept 20, 2017, category 4 Hurricane Maria affected every region of Puerto Rico, devastating its infrastruc ture and disrupting its organisations
Between September, 2017, and February, 2018, we estimated that 1191 excess deaths occurred under the census scenario
Under the preferred displacement scenario, we estimated that 2975 excess deaths occurred during the same observation period
Summary
On Sept 20, 2017, category 4 Hurricane Maria affected every region of Puerto Rico, devastating its infrastruc ture and disrupting its organisations. The intensity of hurricanes has continually increased globally with an even larger destructive potential per storm.[1] In the after math of Maria, a controversy erupted over hurricanerelated mortality. By December, 2017, the Puerto Rico Government certified 64 deaths related to Hurricane Maria. Public perception was that mortality was higher than that in the government’s report.[2] In ensuing months, research by academics and investigative journalists estimated mortality to be much higher than the official count, between 800 and 100 excess deaths.[2,3,4,5,6,7] Survey-based research provided the highest estimate of 4645 deaths, fuelling public debate and uncertainty.[8]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
