How many more? Under-reporting of the COVID-19 deaths in Brazil in 2020.

ABSTRACT

OBJECTIVE: To evaluate the magnitude of under-reporting the number of deaths due to COVID-19 in Brazil in 2020, previously shown to occur due to low rate of laboratory testing for SARS-CoV-2, reporting delay, inadequate access to medical care, and its poor quality, leading to the low sensitivity of epidemiological surveillance and poor outcomes, often without laboratory confirmation of the cause of death. METHODS: Excess mortality due to COVID-19 was estimated directly based on various data sources, and indirectly, based on the difference between the observed and expected number of deaths from serious acute respiratory infection (SARI) and all-natural causes in 2020 had there been no COVID-19. The absence of laboratory testing for SARS-CoV-2 was adjusted based on the proportion of those who tested positive among the tested individuals whose death was attributed to COVID-19. Least absolute shrinkage and selection operator (lasso) were used to improve prediction of likely mortality without COVID-19 in 2020. RESULTS: Under-reporting of COVID-19 deaths was 22.62%, with a corresponding mortality rate per 100 000 inhabitants of 115 by the direct method, 71-76 by the indirect methods based on the excess SARI mortality and 95-104 by excess mortality due to natural causes. COVID-19 was the third cause of mortality that contributed directly with 18%, and indirectly with additional 10-11% to all deaths in Brazil in 2020. CONCLUSIONS: Underestimation of COVID-19 mortality between 15 and 14 is likely its lower bound. Timely and accurate surveillance of death causes is of the essence to evaluate the COVID-19 burden.