A Bayesian estimate of the early COVID-19 infection fatality ratio in Brazil based on a random seroprevalence survey

  1. Valerio Marra
  2. Miguel Quartin

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

No abstract available

Article activity feed

  1. Version published to 10.1016/j.ijid.2021.08.016
  2. ScreenIT

    SciScore for 10.1101/2020.08.18.20177626: (What is this?)

    Please note, not all rigor criteria are appropriate for all manuscripts.

    Table 1: Rigor

    NIH rigor criteria are not applicable to paper type.

    Table 2: Resources

    Antibodies
    SentencesResources
    We also make use of an empirical distribution between the first symptoms and the development of antibodies45 to estimate the mean time-delay τsa between both events.
    antibodies45
    suggested: None

    Results from OddPub: Thank you for sharing your code and data.

    Results from LimitationRecognizer: An explicit section about the limitations of the techniques employed in this study was not found. We encourage authors to address study limitations.

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We did not find any issues relating to the usage of bar graphs.

    Results from JetFighter: We did not find any issues relating to colormaps.

    Results from rtransparent:
    • Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
    • Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
    • No protocol registration statement was detected.

    About SciScore

    SciScore is an automated tool that is designed to assist expert reviewers by finding and presenting formulaic information scattered throughout a paper in a standard, easy to digest format. SciScore checks for the presence and correctness of RRIDs (research resource identifiers), and for rigor criteria such as sex and investigator blinding. For details on the theoretical underpinning of rigor criteria and the tools shown here, including references cited, please follow this link.

  3. Version published to 10.1101/2020.08.18.20177626v3 on medRxiv
  4. NCRC

    Our take

    This ecological study used patient data from three large Brazilian national datasets to calculate a national infection fatality rate (IFR) for Covid-19 of 1.05%, representing 133 cities and 36% of the population. The increase in IFR from May to June 2020 likely represents the increasing saturation of the healthcare system. Though there are a few limitations, the study represents a robust estimate of IFR, adding to the picture previously provided only by modeling.

    Study design

    ecological

    Study population and setting

    This study calculated a population level estimate for COVID-19 infection fatality rate (IFR), utilizing three datasets from Brazil: 1) EPICOVID19-BR, a seroprevalence survey which tested 89,397 individuals (as of July 1, 2020) in three stages, over five weeks, across 28 states in Brazil; 2) SIVEP-Gripe dataset, a prospectively collected respiratory infection registry of Severe Acute Respiratory Syndrome (SARS) cases across both public and private hospitals maintained by the Ministry of Health, which includes 97,924 patients (as of June 16, 2020) with SARS-CoV-2 positive RT-PCR tests; and 3) Painel Coronavírus dataset recording 42,309 Covid-19 fatalities as of July 17, 2020, in 133 Brazilian cities, the Brazilian reference to keep track of the pandemic at the federal level and provides the deaths by COVID-19 with their geographic location.

    Summary of main findings

    The study calculated a country-wide average IFR of 1.05% (95% CI: 0.96–1.17%), and found evidence for the IFR increasing starting in June 2020. This estimate is in agreement with some, but not all, of the previous world estimates. They found a jump in COVID-19 prevalence from late May to early June, and a stabilization in late June. They calculated a mean of 10.3 days from development of antibodies to death, for the whole of Brazil.

    Study strengths

    The study is based on three large Brazilian national datasets, with 42,000 individuals in the smallest, and 98,000 individuals in the largest. The sample included 133 cities around the country, accounting for over a third of the population. Their IFR calculation uses seroprevalence estimates rather than modeling, yielding a more accurate representation. They corrected for false positive and false negative test rates to more accurately estimate prevalence. They properly accounted for the estimated time delay between the development of symptoms and subsequent fatality. Their estimate is also quite precise, with a small statistical error.

    Limitations

    The EPICOVID19-BR study did not hit its target of 250 individuals per city per round, potentially underrepresenting areas with less access to services or trust of healthcare, which could lead to underestimating the IFR. Conversely, the study assumes that SARS-CoV-2 antibodies remain present in patients, which could lead to an underestimation of the percentage of Brazilians that have been infected, which would thus overestimate the IFR. Also, it may be that not all COVID-19 deaths are registered in the Painel Coronavírus dataset. This would more likely be the case for out-of-hospital fatalities and in the poorest areas with less healthcare infrastructure, but as the study focusses on 133 key cities, this bias is likely limited. The population of those cities included 35.5% of the Brazilian population, which is sizable, however the results only apply to those cities, and the IFR may be different in smaller cities and rural or poorer areas which may have less robust healthcare available.

    Value added

    This is the first IFR estimate for Brazil based on patient data rather than modeling.

  5. Version published to 10.1101/2020.08.18.20177626v2 on medRxiv
  6. Version published to 10.1101/2020.08.18.20177626v1 on medRxiv