Over- and under-estimation of COVID-19 deaths

Eur J Epidemiol. 2021 Jun;36(6):581-588. doi: 10.1007/s10654-021-00787-9. Epub 2021 Jul 28.


The ratio of COVID-19-attributable deaths versus "true" COVID-19 deaths depends on the synchronicity of the epidemic wave with population mortality; duration of test positivity, diagnostic time window, and testing practices close to and at death; infection prevalence; the extent of diagnosing without testing documentation; and the ratio of overall (all-cause) population mortality rate and infection fatality rate. A nomogram is offered to assess the potential extent of over- and under-counting in different situations. COVID-19 deaths were apparently under-counted early in the pandemic and continue to be under-counted in several countries, especially in Africa, while over-counting probably currently exists for several other countries, especially those with intensive testing and high sensitization and/or incentives for COVID-19 diagnoses. Death attribution in a syndemic like COVID-19 needs great caution. Finally, excess death estimates are subject to substantial annual variability and include also indirect effects of the pandemic and the effects of measures taken.

Keywords: COVID-19; Death certificates; Diagnostic testing; Excess deaths; Mortality.

Publication types

  • Review

MeSH terms

  • COVID-19 / mortality*
  • Diagnostic Errors / statistics & numerical data*
  • Humans
  • Internationality*
  • Pandemics / statistics & numerical data*
  • Reproducibility of Results
  • SARS-CoV-2