News

SARS-CoV-2 is a novel virus and we gain new information on its transmission almost daily, with an incredible amount of new research on COVID-19 being published each week. Consequently, as our understanding of COVID-19 changes over time, we aim to respond accordingly and update our predictions.

Here, we detail the major version changes to the reports produced in this website (minor changes, such as typo fixes and small edits to plots and text are not documented here). These major changes represent either a new methodology or new information being presented. We hope the following overview will help users understand the changes implemented throughout the pandemic.

Lastly, as we switch from one version to the next, there is likely to be a small delay in keeping the website up to date. We apologise for the delays.

Version 11: 2022-06-27

  • Major overhaul to fitting procedure, please see parameters and methods page.
  • Model calibration now based upon excess-mortality.
  • Incorporated parameter uncertainty for variant adjustments, IFR, and vaccine efficacy.
  • Underlying ODE model updated to account for booster vaccinations.

Version 10: 2022-02-28

  • Adjustments for Delta and Omicron Variants are now added to all countries. For more information see updates to the model parameters and methods page.

Version 9: 2021-11-04

  • For certain countries the impact of the Delta variant is now included in the model. Timings for the delta variant are derived from CoVariants. For more information see updates to the model parameters and methods page

  • Update to the scenario projections. Rather than assuming 50% increases and decreases in effective reproduction number, the scenarios now use Rt values based on estimated trends in the data. For more information see the country pages.

Version 8: 2021-05-10

  • Roll out of vaccines globally is modelled. This is included both in model fitting and scenario projections based on continued roll out speed and current understanding of global vaccine supply. For more information see updates to the model parameters and methods page


#### Version 7: 2021-01-20

  • Waning Immunity included. Mean immunity protection now assumed to be equal to 1 year and described by an Erlang distribution.
  • ECDC stopped daily reporting. We now use JHU CSSE COVID-19 Data. For countries that are not covered by JHU or have delayed reporting patterns, we use the worldometers data set for deaths.


Version 6: 2020-11-11

  • Update to model parameters - see Parameters section for more details.
  • A scenario for countries that are yet to document any COVID-19 deaths - see Parameters
  • For countries that are not covered by ECDC, we use the worldometers data set for deaths.


Version 5 (additional changes to outputs only so same Version): 2020-10-21

  • Outputs now include estimates of active infections (infection prevalence).
  • For countries without deaths, we provide forward scenario projections that consider if importation of new cases occurred. These scenarios use R0 values comparable to countries from the same income strata.
  • Updated data schema (viewable at https://github.com/mrc-ide/global-lmic-reports/blob/master/data/schema.md), which includes these changes.
  • Due to relying on fitting to death data, our estimate of \(R_t\) in the final weeks is uninformed due to the delay from infection to death. In response, for the final month, we estimate Rt, by fitting the rate of change in model reported infections by comparison against the rate of change in official reported cases.


Version 5: 2020-08-08

  • 2-week random walk to capture mobility independent human behaviour changes after the post lockdown shift and response to additional waves.
  • We now present the effective reproduction number \(R_{eff}\), which incorporates the impact that immunity has on transmission as well as the impact of interventions. Previously we presented the time varying reproductive number, \(R_t\), which does not show the impact of immunity and only represents changes in transmission due to interventions.


Version 4: 2020-06-28

  • Mobility independent change in transmission is now also modelled (alongside mobility dependent transmission) post lockdown to reflect change in people’s behaviour that is not captured by mobility.


Version 3: 2020-06-14

  • Impact of mobility on transmision estimated both pre and post lockdown for each country, allowing mobility and transmission to be able to decouple after lockdowns are relaxed if the data supports this.


Version 2: 2020-05-31

  • Data on mobility from the Google Community Mobility Reports to inform changes in transmission.
  • Estimates of \(R_t\) included in country reports
  • Projections extended to 28 days
  • Impact of countries nearing health system capacity explored
  • Regional overviews provided
  • Surged healthcare capacity explored (2020-06-02)
  • Fix for looking at healthcare capacity being surged (2020-06-09)
  • Scenario projections provided in zipped files and update historically for each of pipelines (2020-06-09)


Version 1: 2020-04-23

  • Initial version released, with reports presenting information on the following:
    • The total number of COVID-19 infections
    • The expected number of deaths within the next 14 days
    • The number of individuals requiring oxygen or mechanical ventilation in the next 14 days
    • The impact of changing their current intervention policy
  • Model outputs available to download from home page