How many deaths occurred in England and Wales last week? We don’t know the answer, and that’s understandable. There hasn’t been time for all last week’s deaths to be registered and processed.
 
But let’s ask an easier question. How many deaths occurred in the week beginning 7 June 2010? That was two months ago, and a death in England & Wales must normally be registered within five days. The death registers are computerised, so it ought to be possible for the basic data to be available quickly – name, date of birth, sex, usual place of residence, plus the date of death.
 
What stands in the way is the long-established system that prevents any death referred to a coroner being registered until the cause of that death has been determined. Almost half of all deaths (46 per cent in 2009, a total of 229,000) are referred to coroners, according to the Ministry of Justice’s statistics.
 
The proportion of deaths reported to coroners has been rising steadily, from 37 per cent in 1995 to 46 per cent in 2009, a trend attributed to greater concern for due process in the wake of the Shipman murders, and the increasing use of deputising services by GPs for handling out-of-hours care.
 
Of the 229,000 deaths referred to coroners in 2009, 31,000 (13 per cent) were deemed to require an inquest. Post-mortems were conducted in the great majority (91 per cent) of these cases, and there were an additional 77,100 cases where post-mortems were conducted but there was no inquest.
 
No death can be registered without a cause of death. If an inquest is needed, the Ministry of Justice estimates it takes an average of 26 weeks (up from 22 weeks in 2004) and can take a year or more in some cases. During that period the dead person is not, officially, dead. It is perfectly reasonable to wait some time to establish the cause of death in each case – but is it necessary to wait so long for the simple fact and date of death to be recorded?
 
Scotland shows that it is not. There, deaths must normally be recorded within eight days, whether or not they have been referred to the local Procurator Fiscal. In Scotland, unlike in England & Wales, you do not need a cause of death in order to register it. 
 
The Coroners’ system in England is ancient, but that does not justify such long delays in recording simple facts. Many research projects are handicapped as a result, and some that could be undertaken are not, for the lack of timely data.
 
During the swine flu pandemic, the then Chief Medical Officer, Sir Liam Donaldson, had to conduct his own confidential inquiry into H1N1-related deaths, some of which had been referred to the coroner and were hence unreported.
 
Changes in drug-related deaths – after, for example, the introduction of a new drug of abuse such as mephedrone – are hard to follow because most such deaths are reported to coroners and go unrecorded until a cause of death has been established. Sheila Bird reported here on the difficulties this caused when she attempted to investigate whether changes in drug-taking behaviour detected among soldiers were mirrored in young people generally.
 
The soldiers showed a sharp drop in cocaine use, and it is reasonable to surmise that they had switched to the then-legal mephedrone. If others outside the Army had done the same, then a drop in drug-related deaths might have occurred because mephedrone is less lethal than cocaine. But the long and potentially lengthening delay between the quarter in which a death occurred and the one in which it was registered made such an investigation impossible.
 
Or take statistics on re-offending. Prisoners who die soon release from jail after will be treated as  reformed characters because the system assumes they are still alive, and committing no more crimes. In fact, they are dead, but unregistered, because an inquest has not yet been held.
 
This delay also makes it difficult or impossible to run trials designed to discover the best way to prevent prisoners with a history of heroin injection dying of an overdose within a few weeks of release, as a substantial number do. A randomised trial in which half are allocated to an intervention and half are not requires monitoring in a timely fashion to be ethically acceptable. If a particular intervention proves highly effective, or causes unexpected harm, then the volunteers have a right to know quickly. The present system makes that impossible. 
 
In 2007 the Office for National Statistics undertook a consultation on its mortality statistics. Before, it had produced a range of products that included both deaths registered in a particular year, and deaths that occurred in a particular year. Statistics on registered deaths appeared in the spring of the following year, but those on deaths occurring in a particular year did not appear until the autumn of the following year. Even then, not all the deaths were included because, the ONS admitted, there was evidence of increasing delays in registering deaths.
 
Because of these delays, the ONS suggested that more timely data could be provided if the statistics were based on death registrations, not occurrences, and that is now the case for the majority of mortality statistics. But the consultation, which attracted only ten responses, did not address the issue of the timely recording of fact and time of death.
 
In law, there is already provision for coroners to issue an “interim certificate as to fact of death” in cases where an inquest is to be held. This is done as a courtesy to family and friends, to assist in the administration of the deceased’s estate. If this were done as a matter of course in every case, and the data collected and published, many of the objections to the present system might be removed.
 
The Academy of Medical Sciences is conducting an investigation into the regulation and governance of medical research, chaired by Sir Michael Rawlins. Here is an issue it might usefully include in its deliberations. It seems hardly credible that in the 21st Century we can’t even monitor in a timely fashion whether people are alive or dead.