9 July 2020
Lies, Damned Lies and Coronavirus
Death Statistics
by David Chilvers
There have been many areas where the UK government’s handling of the Coronavirus pandemic could have been better. One of these is the data made available to try and understand how the crisis developed and to help in devising mitigating strategies.
This week*, we look at the data that has been made available regarding deaths. It is important to stress that this is the data made available publicly (well reasonably publicly, as it’s quite hard to find some of it and even harder to be aware of what the data actually represents). One hopes the data available to SAGE has been better and more consistent that what is described here.
Initially, the data on deaths from COVID-19 came from those dying in hospital who had received a positive test. As the pandemic stretched into care homes, the media realised that the official death figures did not include this sector, which was starting to contribute an increasing proportion of the total deaths. As a result, official data extended to include those that had a positive test for COVID-19 and subsequently died in any setting – hospital, care home and at home. So, the numbers increased.
Many of those dying at home and a proportion in care homes never received a test for the virus and so were not included in these revised official statistics. So, the Office for National Statistics, rather belatedly, commenced analysing data from death certificates, where COVID-19 was stated as one of the main causes of death. Of necessity, this data becomes available a week or so after the death occurs, but nevertheless improves coverage, as it includes those who died from the virus but did not have a test. So, the numbers increased – again.
Even this data is not perfect, as some of those dying may have had the disease unknown to the medical practitioner signing the death certificate. So, the concept of “excess deaths” was brought into the public domain – this is the number of deaths occurring each week compared to the expected number from a five-year average. The difference could be attributed to COVID-19. So, the numbers increased – yet again.
Even excess deaths, at this point in time, is not perfect. As many of those dying from COVID-19 are either old and/or have prior medical conditions, it is quite likely that many would have died in the near future in the absence of the Coronavirus pandemic – the virus has brought their death forward, for some by a few weeks, for others longer. So as time goes by, we would expect “excess deaths” each week to be negative – those expected to die would have already done so – and the cumulative figure to decline. This has already happened in London, where the pandemic first took off and other regions are now showing similar signs.
The true figure for deaths will probably never be known, but the above series of events demonstrates how the initial data omitted key sectors and how it was that only under media pressure did the Government act and gradually extend coverage to include substantial missing elements. The extent to which these omissions affected strategy will only be known when the inevitable public enquiry takes place.
Next week, we look at testing statistics, where there are even more holes than in those for deaths.
This article is one of a series, find next week’s article “Testing Statistics” here.
