Disease-specific Excess Mortality During the COVID-19 Pandemic

An Analysis of Weekly US Death Data for 2020

Dongshan Zhu, PhD; Akihiko Ozaki, PhD; Salim S. Virani, PhD


Am J Public Health. 2021;111(8):1518-1522. 

In This Article


We obtained weekly US death data for the country and state level from the National Center for Health Statistics (NCHS).[6,7] We used weekly death data from the first to the 52nd week of 2020. According to the NCHS, data might be incomplete because of the lag in time between when the death occurred and when the death certificate was completed. The delay can range from 1 week to 8 weeks or more, and the more recent the week, the greater the number of delays.

We analyzed deaths from all causes, from natural causes, and disease-specific classifications of deaths (including COVID-19 in 2020). Disease-specific deaths include heart diseases (International Classification of Diseases, Tenth Revision [ICD-10; Geneva, Switzerland: World Health Organization; 1992] codes I00–I09, I11, I13, I20–I51), cerebrovascular diseases (I60–I69), diabetes mellitus (E10–E14), influenza and pneumonia (J10–J18), respiratory diseases (J40–J47, J00–J06, J30–J39, J67, J70–J98), malignant neoplasms (C00–C97), and unclassified deaths (symptoms, signs, and abnormal clinical and laboratory findings, not elsewhere classified, R00–R99). We combined heart diseases and cerebrovascular diseases as cardiovascular disease (CVD). We combined influenza, pneumonia, and respiratory diseases as influenza and respiratory diseases. In this analysis, we recorded deaths with COVID-19 as either an underlying or contributing cause as COVID-19–attributable death (U071).

We chose the average weekly deaths of the previous 6 years (2014–2019) as a baseline to calculate the excess of disease-specific deaths at the country level, and then at the state level (for California, Florida, Texas, and New York—states with the highest numbers of COVID-19 infections), compared with the same week in 2020. We plotted the trajectory of excess deaths of each disease-specific classification using SAS version 9.4 (SAS Institute, Cary, NC). We also plotted the number of weekly new cases of COVID-19 to understand the progress of the pandemic over time. In addition, we traced the governments' response timeline (regulations and recommendations) to the COVID-19 pandemic using the Oxford COVID-19 Government Response Tracker (released by the University of Oxford) to see the influence of the governments' response on the trajectories of excess mortality.[8,9]