March 23, 2020

expert reaction to preprint (not a published paper) on estimates of excess mortality from COVID-19 in the UK

A preprint (not a paper) estimating the excess 1-year mortality from COVID-19 in the UK was posted on Monday 23 March 2020.  This work is not peer-reviewed.

Prof Duncan Young, Professor of Intensive Care Medicine, University of Oxford, said:

“The Intensive Care National Audit and Research Centre (ICNARC) is collecting data for all ICU admissions with COVID-19 in England, Wales and Northern Ireland in near real time.  The first set of collated data was made public on Friday 21 March – see here: https://www.icnarc.org/About/Latest-News/2020/03/22/Report-On-196-Patients-Critically-Ill-With-Covid-19.  This can be used by researchers in trying to understand the disease and who is most at risk of serious disease.  This ICNARC report is preliminary data on the first 196 patients admitted to all ICUs in England, Wales and Northern Ireland with confirmed COVID-19 infection.

“On 20th March 161 ICU beds were occupied with patients being treated for COVID-19 infections.  There are normally over 4000 adult ICU beds in England (https://www.england.nhs.uk/statistics/statistical-work-areas/critical-care-capacity/critical-care-bed-capacity-and-urgent-operations-cancelled-2019-20-data/).

“The numbers have been steadily rising for the last 10 days.

“The majority of cases were being treated in London, Thames Valley and Wessex areas.  This matches the reports of the number of positive tests (see https://www.arcgis.com/apps/opsdashboard/index.html#/f94c3c90da5b4e9f9a0b19484dd4bb14).

“71% were male, a higher proportion than has been seen in reports of positive tests.  The majority were in the 50-80 age range.

“Relatively few had severe comorbidities, but the ICNARC data collection only records the most severe comorbidities causing significant limitation to daily living.  Milder comorbidities such as COPD or type 2 diabetes are not recorded.

“75% (of 175 in whom data on this are available) required treatment with a ventilator when first admitted to the ICU.

“At the time the report was written 16 patients had died, 17 had been discharged from the ICU to a hospital ward and the remainder were still in ICU.  The commonest duration of ICU treatment was three days but this may change as more patients are discharged.

“There had been concern that COVID-19 caused acute kidney injury (acute renal failure).  In this very preliminary report only 4 (of 33 in whom full data are available) patients required renal replacement therapy (akin to dialysis) suggesting this may not be as great a problem as anticipated, but still affects enough patients to cause a significant additional workload.

“This report should be viewed as a very early snapshot of the patients admitted to an ICU with COVID-19. The majority are still being treated in ICUs and so outcome data such as treatment required, duration of stay and survival are not available.”

Prof Robert Dingwall, Professor of Sociology, Nottingham Trent University, said:

“I am not sure that the call for greater suppression efforts is fully backed up by the data in this paper.  Suppression efforts would not prevent deaths, merely delay them, and it is misleading to imply otherwise.  What the paper shows is that Covid-19 is probably going to bring forward a lot of deaths that would have occurred anyway in the next 12 months.  If we take the mid-point of a very wide confidence interval and use 2018 deaths in England as the denominator, so we are talking about roughly 50,000 deaths as a proportion of roughly 506,000, we end up with an increase of  around 10 per cent.  A proportion of these will be deaths brought forward from the next 12 months.  It should not need to be said that each of these deaths is a sad loss for families, friends, neighbours, workmates, etc.  At a population level, however, we can reasonably ask whether the collateral damage to society and the economy from more aggressive suppression is justifiable.  If the benefit is so modest, we might also wonder whether it will actually be exceeded by mortality from other causes such as suicides resulting from the mental health impact of self-isolation, cardiac events prompted by the associated inactivity or longer-term mortality resulting from unemployment and reduced living standards.”

Dr Jennifer Cole, Biological Anthropologist, Royal Holloway, University of London, said:

“This preprint clearly sets out the importance of human behaviour on the spread of the outbreak and the immense responsibility on individuals and the government to make the right decisions.  As the paper says, “the net effect of mortality of this emergency on the population is thus not only a matter of modelling an infectious disease, but modelling the mortality effects of wider medical and societal changes”.  Human behaviour is notoriously difficult to model and will be even more so in a time of social transformation, when norms, behaviours and rules will need to be reshaped and recodified as we enter a ‘new normal’.

“One of the areas in which societal attitudes need to change quickly is over the access researchers are given to NHS data.  At present, due to privacy concerns that have kept NHS data sets siloed, it is not possible to access NHS data on the COVID-19 epidemic as a nation-wide whole.  A key recommendation in this paper is the section that rightly states, “research is a key part of the COVID-19 response, and we believe that the government should act, with legislation and other means, to massively mobilise publicly accountable access to nationwide NHS health and social care data across a large number of data custodians for researchers, clinician and policy makers NHS data across the whole country”.  This is essential to ensure data is available when and where it is needed.

“In 2016, I was one of the signatories to the International Association of National Public Health Institutes (IANPHI) call for health data to be more widely shared – see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6038731/ and https://www.chathamhouse.org/sites/default/files/publications/research/2017-05-25-data-sharing-guide.pdf – and events such as this are precisely why such access to health data is essential.  This is not about invading privacy or snooping on the population, it is about saving lives.  The number of excess deaths, the paper concludes, may be dependent on being able to target preventative interventions to those at highest risk.  Researchers, policymakers and healthcare professionals cannot know which groups or individuals these are without full access to patient data.  The data analysis also then need to be shared with the public, and explained clearly so that the progression of the disease and its likely impact on individuals is fully understood.”

*Preprint (not a paper): ‘Estimating excess 1- year mortality from COVID-19 according to underlying conditions and age in England: a rapid analysis using NHS health records in 3.8 million adults’ by Amitava Banerjee et al. was be posted on Monday 23 March 2020.  This work is not peer-reviewed.

Declared interests

None received.