Counting the dead

For nearly a year after Hurricane María made landfall, the official position of Puerto Rico's government was that 64 people had died as a result of the storm. That number came from the government of Puerto Rico and was treated, for months, as the official record, even as reporters and researchers on the ground raised doubts almost immediately.

The New York Times did its own independent investigation in the months that followed and found evidence pointing to as many as 1,000 potential hurricane-related deaths: a figure many times larger than the government's count, though still an estimate rather than a confirmed tally. In February 2018, CNN and Puerto Rico's Centro de Periodismo Investigativo sued the government for its mortality records. A court ordered the data released, and they showed 1,427 more deaths in 2017 than the average of the previous four years, though not all of those deaths could be directly attributed to the hurricane.

Under growing pressure, the government had already commissioned an independent study from the Milken Institute School of Public Health at George Washington University, contracted in December 2017. On August 28, 2018 — nearly a year after María made landfall — the Government of Puerto Rico officially revised its death count to 2,975, based on that study's findings. The revision made María one of the deadliest hurricanes in U.S. history, a title the original count of 64 had entirely hidden.

The Harvard and NEJM household survey is 4,645. The George Washington University figure is 2,975. Do not attach the survey team to the other study's total.

Second: Kishore and colleagues, in the New England Journal of Medicine, May 29, 2018. A household survey estimated 4,645 excess deaths, with a 95 percent confidence interval of 793 to 8,498, for the window of September 20 to December 31, 2017. Excess means deaths above a baseline, not only drowning and trauma. The interval is wide because the sample is a survey, not a census of death certificates.

It matters to be precise about what that figure of 2,975 is and is not. It is sometimes cited as 2,982 in broader counts of the storm that include other islands, and it remains a statistical estimate of excess mortality: a comparison of how many more people died than would have been expected under normal conditions, not a name-by-name count of confirmed hurricane deaths. The real number is not precisely known. Three layers matter, and they are not one number: the government's original count of 64; the NEJM household survey of 4,645 excess deaths; and the revised estimate the government adopted, 2,975. Together they are evidence of how badly that first count failed the people it was supposed to account for.

Sources: Controversia sobre el numero de muertos del huracan Maria; Huracan Maria; Kishore et al., Mortality in Puerto Rico after Hurricane Maria, New England Journal of Medicine (2018). Freeze 2026-08-22.