Expert comments from statisticians about the COVID-19 death data released daily by the government on their website.
Prof Carl Heneghan, Director of the Centre for Evidence-Based Medicine, and Professor of Evidence-Based Medicine, University of Oxford, said:
“We’ve updated our data on deaths in hospitals in England: https://www.cebm.net/covid-19/covid-19-death-data-in-england-update-2nd-july/
“We look at the number of deaths by date of occurrence and we find that this is showing a different trend to the data on GOV.UK (which is by date of reporting) as ours is still trending down – see the moving average and the weekday comparisons.
“The GOV.UK data is discrepant with ONS deaths data and up to the 19 June there are far fewer deaths reported on the daily death counts released on GOV.UK. We’re in the midst of trying to understand why there is such a discrepancy and whether the GOV.UK data is continually playing catch up.
“The discrepancy is about 9000 deaths compared to ONS deaths by actual date up to the 19 June.
“Using date of reporting of deaths as opposed to the date of occurrence will lead to inaccurate inferences.”
Dr Jason Oke, Senior Statistician at the Nuffield Department of Primary Care Health Sciences, University of Oxford, said:
“The deaths reported on the GOV.UK data are by reporting date and not by date of occurrence and it is unclear, when and where these deaths come from and why they fluctuate so much. For example, yesterday, 176 deaths were reported but only 50 deaths of these were reported from hospitals in England (including catchups) and 9 were from Wales and Scotland. That leaves over 100 deaths outside of the hospital setting. However, this is contrary to the ONS data on location and number of deaths. The ONS data shows that most COVID-19 deaths are in hospital (~60%) and for the last available date, 19 June, and ONS recorded 66 deaths (occurred on June 19) in all settings in England and Wales continuing a downward trend in deaths since Mid-April.
“We need to know how many of the deaths reported on GOV.UK that are outside of hospital are current to be able to judge whether the trend is plateauing.”
All our previous output on this subject can be seen at this weblink: