The grim task of measuring the deadliness of the coronavirus is also a complicated one.
The UK government, the devolved governments, various public health agencies and national statistics bodies are all publishing different data on coronavirus deaths.
All have their strengths and weaknesses when it comes to helping us understand what is going on. Here is Factcheck’s user’s guide…
The daily UK death toll
Tells us: the latest total of deaths of people with coronavirus, mostly in hospitals
Published by: the Department of Health and Social Care and Public Health England
Covers: the whole of the UK
Limitations: daily number does not represent deaths in one 24-hour period; differences in reporting between England and the other nations
This is the number that we see most often in news headlines: we get the latest cumulative death toll from across the UK, reported every afternoon.
It counts people who have died and tested positive for the virus. This isn’t exactly the same as “deaths from coronavirus” as the small print says: “Deaths of people who have had a positive test for COVID-19 could in some cases be due to a different cause.”
There’s a time-lag in this data. The latest daily figure does not count people who died on that day but those whose deaths were reported then. The actual date of death might be several days earlier.
There can be delays of several days between someone dying in a hospital and the death being reported to the health authorities. There might be a need for a post-mortem, or hospitals may want to inform relatives first. Busy hospitals may simply experience administrative delays.
The UK government often states that the national death toll only covers those who have died in hospital, but this rule is only strictly true of England.
The governments of Scotland, Northern Ireland and Wales have all told FactCheck that they include deaths of people who have tested positive and died outside hospital – which might include private health homes – in the figures which they submit to Whitehall.
So it’s not quite right to say that all the deaths included in the UK daily total happened in hospital.
A spokesman for the Department of Health and Social Care said this is mentioned in the small print, and that we are talking about very small numbers of non-hospital cases in these figures.
At time of writing, around 90 per cent of UK coronavirus deaths have taken place in England, so we can assume that most of the people who appear in this daily death toll did indeed die in hospital.
Deaths in England by date
Tells us: when people actually died (unlike the daily death toll)
Published by: NHS England
Covers: England only, hospitals only
Limitations: time lag in reporting means most recent days are always an under-estimate
NHS England breaks down reported deaths by the day they actually took place, which should help us track trends.
The main problem is the delay between someone dying and the death being reported. This means the latest few days, in particular, will always need to be revised upward.
You can see the strength and weakness of this approach by looking at this visualization from Oxford University’s Centre for Evidence-Based Medicine (CEBM):
Counting deaths by the day when they happened allows us to follow the trend. For example, there is tentative evidence of a slow-down here in the daily increase of deaths – the flattening of the curve that we are all hoping to see.
But there is uncertainty too, because the death counts on the most recent days (in the grey box) will certainly have to be revised upwards.
Remember that these figures are only for England and only cover deaths in hospital.
What about people who are dying outside hospital, perhaps in care homes? This brings us to our next data set.
Registered deaths where coronavirus is mentioned
Tells us: about deaths that happen in hospitals and outside; not contingent on a positive test
Published by: Office for National Statistics: National Records of Scotland, Northern Ireland Statistics and Research Agency
Covers: England and Wales; Scotland; Northern Ireland
Limitations: significant time-lag; relies on doctors mentioning Covid-19 on death certificates
National statistics bodies covering the whole of the UK are now publishing data based on death certificates which mention Covid-19 (not necessarily as the main cause of death).
This includes people who die in hospital and elsewhere, so it inevitably gives us a larger number of deaths.
The system relies on doctors filling in death certificates (technically a “medical certificate of cause of death”), which raises some problems, given the lack of coronavirus testing in the community.
According to the Royal College of GPs, doctors can “register COVID-19 as the patient’s cause of death on the basis of a reasoned clinical assessment, whether or not they have been tested for the virus”.
But we can’t be sure that every coronavirus death will be correctly diagnosed and indicated on a death certificate.
The Care Quality Commission has announced that it will begin a new reporting regime covering deaths in care homes in England, which might be an improvement, but we haven’t seen the detail yet.
There’s also another time-lag in these death certificate stats. The ONS figures covering England and Wales were published on Tuesday (April 14) but only cover a period ending on April 3.
Confusingly, the different national statistics bodies are not publishing figures for the same time-period on the same day, with Scotland putting out the most up-to-date figures.
Tells us: whether more people than expected are dying, regardless of stated cause of death
Published by: Office for National Statistics
Covers: England and Wales
Limitations: hard to interpret
Experts are increasingly starting to look at “all-cause mortality” in the registered deaths stats. How many people are dying – for whatever reason – each week, and is it more than expected for the time of year?
The Chief Medical Officer, Professor Chris Whitty, said yesterday that this could become “the most important number” for measuring the whole impact of the epidemic. We are still using death certificates, but we can get around the problem of doctors under-reporting Covid-19 by looking at other causes of death.
The latest ONS figures show that 16,387 people died in England and Wales the week to April 3 – 6,082 more than the five-year average for this time of year.
But only 3,475 (21 per cent) of deaths registered in that week mentioned Covid-19. So a lot more people are dying than expected, but the coronavirus is not being mentioned in most cases.
Experts are interpreting this in a number of ways. It could be that Covid-19 deaths are being significantly under-reported. There may be some evidence of this in the fact that there has been a rise in deaths attributed to influenza and pneumonia – illnesses with similar symptoms to the coronavirus:
Or it could be that the epidemic is indirectly killing people in other ways: perhaps people are finding it harder to get medical attention for other serious illnesses as the NHS is overstretched.
Some experts preach caution, saying that things like changes in the age structure of the population year by year can make big differences in mortality.