As the COVID-19 pandemic persists, time series prevalence data is an essential piece of information needed to track trends in the spread of infection. Across the globe daily COVID-19 viral and antibody tests continue to be administered to monitor new cases and determine if individuals had a COVID-19 infection in the past, respectively. Though these tests give us information on individuals, their suitability for tracking population level trends changes is questionable due to the limited availability of the tests and the populations that they target, those who are symptomatic or in close proximity to confirmed cases. This strategy is more likely to return positive results than a complete scan of the general population but also likely under counts asymptomatic individuals who are known to exist.
In order to get around these shortcomings a team of researchers at the University of Washington, headed by Martina Morris, have been developing a back calculation methodology which uses mortality time series data, an assumed lag time, and estimated infection fatality rates (IFR) to estimate the cumulative total infections over time. The approach relies on population patterns we should observe given that mortality lag and IFR are a particular value which is taken from a previous study. The method is not reliant on testing which is known to be of variable quality depending on the context and has the potential to enable policy makers get a better understanding of the magnitude of the population which may unknowingly carry the virus. Dr. Morris’s work is still under development but you can stay up to date with her teams work and get a more in depth overview of the methodology and data from her working group’s website.