19 October 2020
Any analysis of the currently available COVID-19 data needs to ignore case numbers for recent Specimen Dates. This is due to the delay in reporting test results. IFoA Fellows David Hindley and John Roberts have written an article explaining how a standard actuarial technique can be used to provide faster information regarding the trend of COVID-19 cases in the UK. In ‘Improving timely analysis of UK COVID-19 cases by Specimen Date’, the authors describe a particular form of data which provides powerful insights into the reporting delay by Specimen Date.
However, they comment that this form does not appear to be available at present from publicly available sources, without taking daily downloads of the data. Hence, the authors encourage the UK government to make the required data more readily available.
Using the required form of data, the article shows how a relatively simple technique which is very widely used in insurance could potentially be used to provide estimates of case numbers for the recent Specimen Dates. This would give useful and more timely information for policy makers on the latest trends. The article uses ‘old’ data merely to demonstrate the concepts; the authors plan to apply the techniques to more current data and will provide an update on the results in due course.