Age- and sex-specific total mortality impacts of the early weeks of the Covid-19 pandemic in England and Wales: Application of a Bayesian model ensemble to mortality statistics

  1. Vasilis Kontis
  2. James E Bennett
  3. Robbie M Parks
  4. Theo Rashid
  5. Jonathan Pearson-Stuttard
  6. Perviz Asaria
  7. Michel Guillot
  8. Marta Blangiardo
  9. Majid Ezzati

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Background

The Covid-19 pandemic affects mortality directly through infection as well as through changes in the social, environmental and healthcare determinants of health. The impacts on mortality are likely to vary, in both magnitude and timing, by age and sex. Our aim was to estimate the total mortality impacts of the pandemic, by sex, age group and week.

Methods

We developed an ensemble of 16 Bayesian models that probabilistically estimate the weekly number of deaths that would be expected had the Covid-19 pandemic not occurred. The models account for seasonality of death rates, medium-long-term trends in death rates, the impact of temperature on death rates, association of death rates in each week on those in preceding week(s), and the impact of bank holidays. We used data from January 2010 through mid-February 2020 (i.e., week starting 15 th February 2020) to estimate the parameters of each model, which was then used to predict the number of deaths for subsequent weeks as estimates of death rates if the pandemic had not occurred. We subtracted these estimates from the actual reported number of deaths to measure the total mortality impact of the pandemic.

Results

In the week that began on 21 st March, the same week that a national lockdown was put in place, there was a >92% probability that there were more deaths in men and women aged ≥45 years than would occur in the absence of the pandemic; the probability was 100% from the subsequent week. Taken over the entire period from mid-February to 8 th May 2020, there were an estimated ~ 49,200 (44,700-53,300) or 43% (37-48) more deaths than would be expected had the pandemic not taken place. 22,900 (19,300-26,100) of these deaths were in females (40% (32-48) higher than if there had not been a pandemic), and 26,300 (23,800-28,700) in males (46% (40-52) higher). The largest number of excess deaths occurred among women aged >85 years (12,400; 9,300-15,300), followed by men aged >85 years (9,600; 7,800-11,300) and 75-84 years (9,000; 7,500-10,300).

The cause of death assigned to the majority (37,295) of these excess deaths was Covid-19. There was nonetheless a >99.99% probability that there has been an increase in deaths assigned to other causes in those aged ≥45 years. However, by the 8 th of May, the all-cause excess mortality had become virtually equal to deaths assigned to Covid-19, and non-Covid excess deaths had diminished to close to zero, or possibly become negative, in all age-sex groups.

Interpretation

The death toll of Covid-19 pandemic, in middle and older ages, is substantially larger than the number of deaths reported as a result of confirmed infection, and was visible in vital statistics when the national lockdown was put in place. When all-cause mortality is considered, the mortality impact of the pandemic on men and women is more similar than when comparing deaths assigned to Covid-19 as underlying cause of death.

Article activity feed

  1. ScreenIT

    SciScore for 10.1101/2020.05.20.20107680: (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

    No key resources detected.

    Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).

    Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:
    The main limitation of our work is that we did not have data on underlying cause of death beyond the distinction between Covid-19 and non-Covid deaths. Having a breakdown of deaths by underlying cause will help develop cause-specific models and understand which causes have exceeded or fallen below the levels expected. We also could not access age-specific data by region because the ONS only releases aggregate numbers for regions. Nor did we have data on total mortality by socio-demographic status to understand inequalities in the impacts of the pandemic beyond deaths assigned to Covid-19 as the underlying cause of death. Releasing these data will allow more granular analysis of the impacts of the pandemic, which can in turn inform resource allocation and a more targeted approach to mitigating both the direct and indirect effects of Covid-19, now and for future waves of the pandemic. Further, weekly mortality files from the ONS cover deaths registered in any given week. These include some deaths from prior weeks and leave out some deaths in the reporting week.26 However, the approach is consistent over time and does not affect year to year comparisons, including for 2020 as the lags in registration of deaths assigned to Covid-19 seem to be the same as those from other causes.27 It is likely that some of the apparently non-Covid excess deaths are due to undetected Covid-19 infections.28 An example of such deaths are the likely Covid-19 deaths in care homes.29,30 Other such deat...

    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.

  2. Version published to 10.1101/2020.05.20.20107680v1 on medRxiv