COVID-19 deaths that have occurred but are not yet reported – update 18 May 2020.

May 18, 2020

Stavros Christofides, Jason Oke, Carl Heneghan

We have updated our Chain Ladder projections using deaths occurring up to 16 May 2020.

The overall occurred but are not yet reported (OBNR) estimates have been reducing due to two factors:
1. The significant reduction in daily deaths and
2. The gradual increase in the speed of reporting these deaths and what may be the catching up or clearing up of previous backlogs of cases.

The OBNR as projected by the Chain Ladder is now estimated at 456 cases.

The table below shows the projections covering the last three weeks of deaths and also the Bootstrap standard errors for the most recent ten days. With the exception of the most recent day where, understandably, the standard error is high, the rest are more or less in line with the estimated daily reduction in deaths. The reductions in these daily estimates are indicative of the improvements in the reporting of these deaths. This has been reflected in the latest Chain Ladder projection by assuming that all deaths are now being reported within 30 days.

The Chain Ladder projections are producing reasonable early estimates of actual deaths by date of death and show a continuing downward trend in these deaths since the high point of the 8th of April.  Over the last two weeks or so, the Chain Ladder estimates show that the daily rate of decline in these hospital deaths has reduced to around 10 deaths per day or less than half the daily rate fall observed of 23 cases per day over the period from the 8th April to the 5th  May.

The OBNR or the deaths that have occurred but are as yet not included in the daily counts released by the NHS are estimated to be 456 cases and reducing as the daily death rate has declined. Some 379 of these cases relate to the most recent two weeks of actual deaths.

See: COVID-19: The Chain Ladder Method to estimate deaths not yet reported

England,P.D. & Verrall,R.J.(1999).Analytic and bootstrap estimates of prediction errors in claims reserving Insurance : Mathematics and Economics,25,281-293.

Stavros Christofides B.Sc, M.Phil worked in Insurance as a Manager and as a Consultant and has published papers on Loss Reserving, Insurance Pricing and Dynamic Financial Analysis. He was also an Honorary Visiting Fellow at Cass Business School, City University.

Jason Oke is a Senior Statistician at the Nuffield Department of Primary Care Health Sciences and Module Coordinator for Statistical Computing with R and Stata (EBHC Med Stats), and Introduction to Statistics for Health Care Research (EBHC), as part of the Evidence-Based Health Care Programme.

Carl Heneghan is Professor of Evidence-Based Medicine, Director of the Centre for Evidence-Based Medicine and Director of Studies for the Evidence-Based Health Care Programme. (Full bio and disclosure statement here)

Disclaimer:  the article has not been peer-reviewed; it should not replace individual clinical judgement, and the sources cited should be checked. The views expressed in this commentary represent the views of the authors and not necessarily those of the host institution, the NHS, the NIHR, or the Department of Health and Social Care. The views are not a substitute for professional medical advice.