Prof Carl Heneghan and Jason Oke

A statistical tweak has just cut England’s Covid death toll by 5,000

A statistical tweak has just cut England's Covid death toll by 5,000
Getty images
Text settings
Comments

Public Health England has changed its definition of Covid deaths. Hitherto, it counted anyone who died who had previously had a laboratory-confirmed positive Covid-19 test - irrespective of how long ago that test was  The new definition introduces a time limit: the death of someone who tested positive no more than 28 days before death. The new data (here) means England has had 36,695 deaths using this definition as opposed to the previous reported 42,072 – a difference of 5,377. Here's how the figures compare:

So the difference is more marked in June, July and August. For example, under the old  system, 2,086 deaths were reported in England. With the 28-day cutoff,  this number is 574 – nearly a quarter of what was previously reported.

The last two days illustrate the difference this change makes. As opposed to 100 deaths reported on the 11th of August, 11 would now be reported, and today, whereas 72 were reported there are only 15 under the new system.

DateEver had a positive testTest with 28-day cut-off
11/08/202010011
12/08/20207215

We have illustrated this effect on the seven-day moving average for July:

The change in definitions means the current moving average, as of the 12 August, is down to approximately ten deaths per day.  But this still doesn't change the league table for total deaths: the UK (and England) remain in the same position, above Italy. When it comes to deaths per million (i.e. adjusting for the population size) it means the UK’s death rate is similar to Italy and Spain and above France, whereas England alone death rate per million remains higher.

CountryTotal deathsPer millionPopulation

UK

41,329

609

67,926,890

England

36,695

655

55,980,000

Italy

35,225

465

65,290,273

Spain

28,581

611

46,756,895

PHE uses two definitions of death in a person with Covid-19 in England, one broader measure and one measure reflecting current trends:

  1. A death in a person with a laboratory-confirmed positive Covid-19 and either: died within 60 days of the first specimen date or died more than 60 days after the first specimen date, only if Covid-19 is mentioned on the death certificate
  2. A death in a person with a laboratory-confirmed positive Covid-19 test and died within (equal to or less than) 28 days of the first positive specimen date.

In terms of the definition with a longer time frame, beyond 28 days, we consider this is important but should be used to perform in-depth analyses of the long term problems caused by Covid.

The 28 definition brings the definitions into line with the other countries in the UK and provides an accurate measure of the immediate impact of Covid on deaths. This will allow us to better determine if deaths related to Covid are trending up or down.


Written byProf Carl Heneghan and Jason Oke

Dr Jason Oke is a Senior Statistician at Oxford's Nuffield Department of Primary Care Health Sciences. Carl Heneghan is Professor of Evidence-Based Medicine and Director of Studies for the Evidence-Based Health Care Programmes at Oxford.

Comments
Topics in this articleSociety