COVID-19 deaths in England and Wales: resolving discrepancies in deaths outside of hospital
July 28, 2020
Daniel Howdon, Carl Heneghan
We note two potentially important issues about data collected in England and Wales. Both of these suggest potential areas of concern in non-hospital settings. First, many more individuals, compared to a baseline of the weekly five-year average, appear to be dying at home. Second, a large discrepancy is apparent in COVID-19 deaths according to whether these are measured using a death certificate-based method or a positive test-based method. Explaining these issues has implications for public policy regarding COVID-19 mitigation and public health more broadly.
The Office for National Statistics (ONS) publish data on all deaths registered in England and Wales on a weekly basis. This dataset comprises a count of death certificate mentions of COVID-19, irrespective of whether it is considered to be an underlying or contributory cause of death. Such data can be termed “death certificate-based”.
All other data regarding COVID-19 related deaths published in England and Wales rely on whether an individual has had a positive test for the disease. Such data can be termed “positive test-based”.
This includes NHS England’s (NHSE) daily publication of deaths in English hospitals and Public Health Wales’ (PHW) daily publication in Welsh hospitals and care homes. Data from both of these sources form part of the announced national figures on deaths in the UK, published on the gov.uk website.
Although daily headline numbers from all of these sources are based on a count of deaths not previously announced, all such sources now provide counts by date of death.
Excess Deaths at home
While the currently registered deaths in England and Wales have fallen below the five-year average for the last five weeks, excess deaths at home remain above average and high. Over 700 excess deaths per week – 3,799 in total reported in the home setting over the past five weeks. Only 179 deaths of these have mentioned COVID-19 on the death certificate.
Figure 1: excess deaths in England and Wales by the week of registration, by setting (source: ONS).
It is not clear why there is such excess in the home. What is clear is that this represents a huge number of unexplained – and potentially avoidable deaths – particularly if they represent individuals deterred from visiting hospitals.
The number of deaths in the home setting are almost 40% higher than the total number registered with COVID-19 in any setting over the last five weeks (3,799 versus 2,582).
COVID-19 deaths outside of hospital and differences between death certificate-based and positive test-based.
Using date-of-death based figures provided by ONS, NHSE, PHW, and the all-setting series for England compiled by Public Health England (PHE), available at gov.uk there are large divergences in daily deaths according to whether counts are death certificate-based or positive test-based.
Figure 2 shows, by date of death since 1st of June, the PHE positive test-based all-settings count of COVID-19 deaths in England, and the ONS death certificate-based count of the same.
Since the 1st of June, the positive test-based count of deaths in England and Wales has exceeded the death certificate-based counts.
While the total number of deaths recorded in each method has fallen the difference has remained roughly constant at around 30 deaths per day. This represents as many as 1,424 extra positive test-based deaths compared to certificate-based deaths.
Examining the ONS’s breakdown of date-of-death data by the setting of death.
As ONS does not provide any breakdown of this setting by nation, we compare figures below for England and Wales.
Figure 3 uses a comparison of in-hospital deaths recorded by ONS and a sum of in-hospital deaths (NHSE and PHW) and care home deaths (PHW). There is an extra count in the positive test-based figures of deaths in Welsh care homes, but this would represent only a small number of deaths.
The total difference between death-certificate based and positive test-based death counts in hospitals come to only 143: only 10% of the 1,424 difference in such deaths observed in England in all settings and illustrated in Figure 2. This suggests that the discrepancy between positive test-based and death certificate-based deaths is largely arising outside of the hospital setting. We cannot provide more clarity on this issue due to the inadequacies in the current data.
Daniel Howdon is a Senior Research Fellow in health economics at Leeds Institute of Health Sciences, University of Leed. Bio here
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.