Written Evidence from the Royal Statistical Society
This is the submission from the Royal Statistical Society (RSS) to the Justice Select Committee’s inquiry into ‘The Coroner Service’.
We are focusing on Question 4, since it links most closely to the Society’s work on official statistics and our longstanding concerns around death registration.
Indeed, in the aftermath of the swine-flu pandemic in 2009-10, Professor Sheila M. Bird FRSE led for the RSS on the need for legislation to end the late registration of deaths in England, Wales and Northern Ireland.
The RSS’s National Statistics Advisory Group (NSAG) routinely engages with the UK Statistics Authority (UKSA) on issues around official statistics - which include the recording of deaths - while our Covid-19 Task Force (led by our President-elect Professor Sylvia Richardson and past-President Professor Sir David Spiegelhalter FRS) has been active in addressing a wide range of statistical issues during the Covid-19 pandemic.
Question 4. How the Coroners Service has dealt with COVID 19
The RSS has a long record of campaigning on the issue of death registration and the need for legislative change. Below are a few examples of our work in this area:
The central problem is that when deaths are referred to the coroner for investigation (in England, Wales and Northern Ireland), the fact of death is not registered with the Office for National Statistics (ONS) until cause of death has been determined. This means that deaths can be registered months later than they have occurred.
For example, recently published ONS provisional death registration statistics show that a Covid-19 death that had occurred on 4th April 2020 was not registered until after 14th August - a registration delay of nearly 19 weeks.
The RSS’s concerns are two-fold. First, demographical and computational failure by three of the UK’s four nations to register fact-of-death promptly in the 21st century is astonishing and needs to be addressed urgently.
Recourse to reporting deaths by registration-week rather than by occurrence-week disguised the problem rather than solved it and, despite improved labelling of tables, even the best journalists can mis-report - for example, by writing “suicides in 2019”, rather than “suicides registered in 2019” as only half the suicides registered in 2019 will actually have occurred in 2019, with the other half (due to registration delays) having occurred in earlier years.
Secondly, official death counts, based on registration-week, are less useful to policy-makers and public health authorities - for example, in the event of public health emergencies when decision-makers need official statistics that are both timely and accurate. Based on registration-week, the true rate of increase in Covid-mention deaths in England was under-estimated in March and April 2020; the peak-week was misplaced; and the rate of decrease in Covid-mention deaths also under-estimated when deaths are counted by registration-week rather than by occurrence-week.  By law in Scotland, unlike in the rest of the UK, all deaths must be registered within eight days of deaths having been ascertained so that registration delays are curtailed.
Outside of Scotland, the Covid-19 pandemic has highlighted problems with the current approach which had been the subject of an RSS warning as long as ten years ago, in the aftermath of the H1N1 pandemic when Professor Bird gave written and oral evidence to the House of Commons Science & Technology Committee, as part of its 2010-11 inquiry into ‘Scientific Advice and Evidence in Emergencies’.
During the SARS-CoV-2 pandemic, the most reliable Covid-mention death figures have been reported by the ONS, with reporting both by registration-week and by occurrence-week.
However, these figures do not include those deaths which have been referred to coroners and are subject to ongoing investigation - meaning that they will have been under-counting the number of Covid-related deaths throughout the crisis. The greater the delay in registering deaths, the greater the uncertainty around the number of people who have died at a time when this information is vital.
Until registration delays of longer than eight days are legislated against, additional special measures are needed in pandemics so that a duty is placed on the Chief Coroner to ensure that suspected pandemic-mention deaths referred to any coroner are immediately notified to ONS. The Coronavirus Act 2020 failed to ensure that a suitable provision came into effect for the first wave of Covid-19 deaths. The RSS recommends that an appropriate amendment is made to the Coronavirus Act 2020 before the UK encounters a second wave of Covid-19 deaths.
Because of delays in the registration of deaths combined with pandemic-related need to know about the suspected cause of coroner-referred death (Covid-mention deaths, in particular), a timely count of the occurrence of deaths caused by Covid-19 has been inhibited. The problem is particularly acute for health and social care workers, who have been on the frontline of responding to the disease and are, therefore, more likely to have been exposed to the virus.
Based on Scottish data, Bird et al. estimated that there may be in excess of 50 Covid-related deaths of health and social care workers in England and Wales that have been referred to coroners but have yet to be registered with ONS. 
Information, including occupation, on Covid-related deaths that are under investigation by coroners is urgently required from the Chief Coroner in England to draw clear occupational-risk inferences. Preparedness for a second wave of Covid-19 should ensure that this coronial registration gap is plugged - not only in England but also in Wales and Northern Ireland - and that at least some inquests conclude before the autumn ends.
During a pandemic, an emergency system should be in place whereby the ONS is informed immediately about any death referred to coroners, so that fact-of-death is duly registered (as for all other deaths) even though the presumptive cause-of-death has yet to be confirmed. In addition, ONS should be informed if the presumptive cause-of-death is pandemic-related.
There is an urgent, pandemic-driven need for improvement, as above, from the Coroner Service, which is, of course, the focus of the Committee’s inquiry. If there is a substantial increase in Covid-19 cases over the winter, policy-makers and public health authorities need to be able to access information about the number, age, gender, region, place of death and occupation of Covid-mention deaths as quickly as possible, including for health and social care workers.
In the longer term, the RSS believes that a legislative change is needed to require that, irrespective of coroner-referral, fact-of-death is officially registered within eight days of death having been ascertained. Registration of fact-of-death is essential to enable record-linkage studies and clinical trials to know promptly when individuals under follow-up have died. Otherwise, investigators do not know whether re-incarcerations or hospitalizations (say) have ceased because of rehabilitation or due to death.
RSS strongly recommends that England, Wales and Northern Ireland should adopt a system more like that used in Scotland where, without exception, all deaths must be registered within eight days of death having been ascertained.
 See Sheila M. Bird and Bent Neilson, ‘Lethality of Covid-19 in Great Britain over 12 registration weeks’, Significance, August 2020.
 See Sheila M. Bird, Neil Pearce and Martie Van Tongeren, ‘We need accurate and rapid reporting on deaths referred to the coroner’, BMJ Opinion, 30th July 2020.