Background: The World Health Organization (WHO) kept track of COVID-19 data at country level daily during the pandemic that included the number of tests, infected cases and fatalities. This daily record was susceptible to change depending on the time and place and impacted by underreporting. In addition to reporting cases of excess COVID-19-related deaths, the WHO also provided estimates of excess mortality based on mathematical models.
Objective: To evaluate the WHO reported and model-based estimate of excess deaths to determine the degree of agreement and universality.
Methodology: Epidemiological data gathered from nine different countries between April 2020 and December 2021 are used in this study. These countries are India, Indonesia, Italy, Russia, United Kingdom, Mexico, the United States, Brazil and Peru and each of them recorded more than 1.5 million deaths from COVID-19 during these months. Statistical tools including correlation, linear regression, intraclass correlation and Bland-Altman plots are used to assess the degree of agreement between reported and model-based estimates of excess deaths.
Results: The WHO-derived mathematical model for estimating excess deaths due to COVID-19 was found to be appropriate for only four of the nine chosen countries, namely Italy, United Kingdom, the United States and Brazil. The other countries showed proportional biases and significantly high regression coefficients.
Conclusion: The study revealed that, for some of the chosen nations, the mathematical model proposed by the WHO is practical and capable of estimating the number of excess deaths brought on by COVID-19. However, the derived approach cannot be applied globally.
Keywords: Bland-Altman plots; World Health Organization; excess deaths; intraclass correlation.
© 2023 John Wiley & Sons Ltd.