Estimating Excess Deaths in the United States Early in the COVID-19 Pandemic
MetadataShow full item record
Background: Accurate mortality data may improve public compliance with non-pharmaceutical interventions and identify vulnerable communities in need of resources. Methods: We estimated excess all-cause, pneumonia, and influenza mortality using weekly state-level data from the Mortality Surveillance Survey covering September 27, 2015 to April 11, 2020, using a semiparametric method and a conventional method in 9 states with complete data: California, Colorado, Florida, Illinois, Massachusetts, Michigan, New Jersey, New York, and Washington. Results: We estimated 43-359 more deaths in Colorado than COVID-19 official counts (excess mortality 95% CI (317, 633) vs. 274 COVID-19 deaths); 418-1208 more deaths in Illinois than COVID-19 official counts (excess mortality 95% CI (1100, 1890) vs. 682 COVID-19 deaths); 280-810 more deaths in Massachusetts (95% CI (966, 1496) vs. 686 COVID-19 deaths); 615-1453 more deaths in Michigan (95% CI (2006, 2844), vs. 1391 COVID-19 deaths); 2689-4126 more deaths in New Jersey (95% CI (4872, 6309), vs. 2183 COVID-19 deaths); and 6687-10696 more deaths in New York State (95% CI (15314, 19323) vs. 8627 COVID-19 deaths). We conservatively found that between 32 and 4,259 unexplained deaths in New York, and 378 and 1,937 in New Jersey. Conclusions: Official COVID-19 mortality figures substantially understate actual mortality during the pandemic. Improving mortality data dissemination may improve public adherence to non-pharmaceutical interventions and reduce mortality in future pandemics.