Data and the scandal of the UK’s Covid-19 survival rate

I have held off writing much that is overtly critical of the UK government’s handling of the Covid-19 pandemic, but can do so no longer. We have known for a long time that data published by governments across the world about infections is highly unreliable, although figures on deaths are somewhat more representative of reality.  The UK governments’s lack of transparency, though, about its Covid-19 data is deeply worrying, and suggests deliberate deceipt.  The following observations may be noted about the figures that are currently being published, and the ways in which official (and social) media use them.

• Official infection rates are very unreliable and largely reflect the number of tests being done.  These figures are so meaningless that they should be ignored in public announcements and media coverage because they give the public completely the wrong impression.  Countries such as Germany are believed to be able to produce up to 500,000 tests a week (although their aim is to do 200,000 tests a day), whereas by 7th April there had only been 218,500 tests in total in the UK since the start of January. The UK government aims to achieve 100,000 tests a day by the end of April, but seems highly unlikely to meet this target; a figure of more than 10,000 tests per day in the UK was only first achieved on 1st April.  The official reported number of infected cases in Germany at 119,624 on 10th April is  likely to be somewhat nearer reality than the paltry 73,758 reported cases in the UK (Source: thebaselab, 10th April).  In practice, it seems that most of the UK figures actually refer to those who are tested in hospital as suspected cases, since there is negligible testing of the public in general to get an idea of how extensive the spread really is.  By keeping this figure apparently low, the UK government seems to be deceiving the population into believing that Covid-19 might be less extensive than in reality it is.
• Figures for the number of deaths should be more reliable, but are also opaque.  Even with figures for deaths there is increasing cause for doubt, not least because of differences between countries reporting whether someone has died “from” or “with” Covid-19.  In practice, it is even more complex than this, since some countries (such as the UK), are publishing immediate data only on those who die in hospital.  Those who die in the community are only added into the total official figures at a later date.  By manipulating when these figures are officially added, governments can again deceive their citizens that the deaths may in the short-term be lower than they are in reality.  A good analysis of the situation in the UK has recently (8th April) been produced by Jason Oke and Carl Heneghan for The Centre for Evidence-Based Medicine (CEBM), which highlights the considerable discrepancies between data made available by the National Health Service (NHS) and Public Health England (PHE).  Not only does this make it difficult in the short-term for modellers and policy makers to know what is really happening, but it also gives a distorted picture to the public.  As this report also concludes “The media should be wary of reporting daily deaths without understanding the limitations and variations in different sources”.
• Hugely unreliable mortality rates.  Combining published figures for infections and deaths gives rise to figures for mortality rates.  These figures are also therefore very unreliable.  Because of the low levels of testing, and yet the high number of deaths in the UK (8,958; Source: thebaselab, 10th April), the UK mortality rate is reportedly the second highest in the world at 12.15%.  This can be compared with Germany’s 2.18% (undoubtedly a much more accurate figure), Italy’s 12.77% (the highest in the world), and a global average of 6.06%.  As I have argued previously, though, these figures are largely meaningless, and the figures that really matter are the total number of deaths divided by the total population of a country.  Accordingly, to date, China has had only 0.23 deaths per 100,000 people, whereas Spain has had 33.88, Italy 30.23, France 18.80 and the UK currently 11.75 deaths per 100,000 (Source: derived from thebaselab, 10th April).  Put another way, the UK figure is 51 times more than the Chinese figure.  Such figures are far more meaningful than official mortality rates, and should always be used by the media (preferably using choropleth maps rather than proportional circles for total deaths).
• Extraordinarily depressing recovery rates.  The UK’s current “recovery rate” is by any standards appalling.  As of 9th April reported figures for the number of people who have recovered from Covid-19 in the UK were between 135 (by the baselab, and worldometers) and 351 (by Johns Hopkins University).   This suggests a “recovery rate” of possibly only 0.18% in the UK (Source:  thebaselab, 10th April), in contrast with China’s 94.56%, Spain’s 35.45% and a global average of 22.2%.  In part this is again a result of data problems.  We simply don’t know how many people have been infected mildly, and how many have survived without even knowing they have had it.  It also reflects the fact that it takes time to recover, and many people are still in hospital who may yet recover.  However, the UK’s figures is the worst in the world for countries where there have been more than 50 cases of Covid-19.

Such figures raise huge questions for the British government and people:

• Why are UK reported survival rates so low? Surely the government should want to do all it can to show the success of the NHS in treating patients and it should therefore publish the real figures?  That is unless, of course, these figures are truly bad.
• What is the balance of numbers between those dying in hospital from Covid-19 and those leaving having recovered?  The rare euphoria that greets those who leave hospital having recovered (as with 101-year-old Keith Watson who was recently discharged from a hospital in Worcestershire) suggests that very few people have actually left hospital alive having been admitted with Covid-19.  Is the government trying to hide this?  Is the grim truth that you are likely to die if you go into hospital with Covid-19?  Does this mean that people are being admitted to hospital far too late because of the advice given by the NHS and its 111 service?  Should the NHS simply stop trying to treat patients with Covid-19? (An update noted below suggests that more than half of the people going into intensive care in UK hospitals with Covid-19 die).
• Why did the government not act sooner?  Some of us had argued back in January of the threat posed by the then un-named new coronavirus (I first raised concerns on 20th January, and first posted about its extent in China on social media on 27th January).  It was very clear then (and not only with hindsight) that this posed a global threat.  Undoubtedly the WHO failed in its warnings, and did not act quickly enough to declare a pandemic, but many governments did act to get in supplies of Personal Protective Equipment (PPE), testing equipment, and ventilators.  The UK government has failed its people.  One quarter of my close family have probably already had Covid-19; many of my friends have also had it – some very seriously.  I guess therefore that between a quarter and a third of those living in the UK may already been ill with the pandemic (Update 13th April: this must be an exaggeration, as news media over Easter suggest that experts think the current figure of infections is only 10%; Update 26th April, the MRC-IDE at Imperial College modelling back from actual deaths, suggest that only some 4.36% of the UK population is infected).  They are individual human beings, and not just statistics.

These questions are hugely important now, and not just when a future review is done, because it is still not too late to act together wisely to try to limit the impact of Covid-19 in the UK.  The fact that the government has not yet been transparent and open about these issues is deeply worrying.  In trying to explain them the following scenarios seem likely.  I very much hope they are not true, and that the government can provide clear evidence that I am wrong:

1. Throughout, the government knew that the NHS would be overwhelmed by Covid-19, and has been doing all it can to cover up its own failings and to protect the NHS.  In 2016, a review called Exercise Cygnus was undertaken to simulate the impact of a major flu pandemic in the UK. The full conclusions have never been published, but sufficient evidence is in the public domain to suggest that it showed that the NHS was woefully unprepreard, with there being significant predicted shortages of intensive care beds, necessary equipment, and mortuary space.  In December 2016 the then excellent Chief Medical Officer Dame Sally Davies, conceded that “a lot of things need improving”.  It is now apparent that the government (largely including people who are still leading it) did nothing to rectify the situation, and must therefore be held in part responsible for the very high death rate in the UK.  Its failure to fund the NHS appropriately in recent years is but a wider symptom of this lack of care and attention to the needs of our health system.  I therefore find it very depressing that this government is now so adamant in asking us to protect the NHS; as shown on the cover of the document sent to all households in the UK (illustrated above), it seems to be more concerned with protecting the NHS (listed second) above saving lives (listed third).

2. The government has consigned those least likely to survive Covid-19 to death in their homes.  Despite claims that the government is caring for the most vulnerable, it seems probable that its advice to the elderly and those most at risk to stay at home was not intended primarily for their own good, but was rather to prevent the NHS from being flooded with people who were likely to die.  This is callous, calculating and contemptable.  On March 22nd, The Sunday Times published an article that stated that “At a private engagement at the end of February, Cummings [the Prime Minister’s Chief Advisor] outlined the government’s strategy. Those present say it was “herd immunity, protect the economy and if that means some pensioners die, too bad”. Downing Street swiftly denounced this report, but it remains widely accepted that even if these were not the exact words Cummings used, this was indeed the view of some of those at the top of the UK government at that time.  Subsequent evidence would support this.  Some, perhaps many, hospital trusts, for example, have clearly told their staff not to accept people who are very old and fall into the most vulnerable category.  Likewise, Care Homes have been told to care for Covid-19 patients themselves, since they may not be accepted in hospital. The British Geriatrics Society thus notes (30th March) that:

• “Care homes should work with General Practitioners, community healthcare staff and community geriatricians to review Advance Care Plans as a matter of urgency with care home residents. This should include discussions about how COVID-19 may cause residents to become critically unwell, and a clear decision about whether hospital admission would be considered in this circumstance”
• “Care homes should be aware that escalation decisions to hospital will be taken in discussion with paramedics, general practitioners and other healthcare support staff. They should be aware that transfer to hospital may not be offered if it is not likely to benefit the resident and if palliative or conservative care within the home is deemed more appropriate. Care Homes should work with healthcare providers to support families and residents through this”

This  policy incidentally (and also helpfully for the government) lowers the daily reporting death rate because such people are not counted as “dying in hospital”.

3. The use of digital technologies may be used to identify those unlikely to be given hospital treatment.  The government quite swiftly introduced online methods by which people who think that they fall into the extremely vulnerable category could register themselves, so that they might receive help and such things as food deliveries.  Whilst aspects of this can indeed be seen as positive, it also seems likely that this register could be used to deny people access to hospital services, since they are most likely to die even with hospital treatment.  If true (and I hope it is not), this would be a very deeply worrying use of digital technologies.  Nevertheless, care homes are being forced to hold difficult discussions with those they are meant to be caring for about end-of-life wishes, and all doctors and medical professionals are increasingly having to make complex ethical decisions about who to treat (see Tim Cook’s useful 23rd March article in The Guardian).

4. The government has tried to pass the blame onto the scientists. Early on in the crisis I was appalled to see and hear government spokespeople (including the Chief Medical Officer – so beloving of systematic reviews) saying that they were acting on scientific advice.  As some of us pointed out at the time, there is no such things as unanimity in science, and so it was ridiculous for them to claim this.  However, they seem to have been doing so, and in such a co-ordinated manner, because they were seeking to shift the blame in case their policies went wrong.  Leading a country is a very tough job, and those who aspire to do so have to make tough decisions and stand by them.  Fortunately, this position by the government is no longer tenable, especially now that academics are competing visciously in trying to prove that they are right, so that they can take the credit. Nevertheless, there remains good science and bad science, and it is frightening how many academics seem to be pandering to what governments and the public might want to hear.  Tom Pike (from Imperial College), for example, predicted (against most of the prevailing evidence) in a pre-print paper with Vikas Saini on 25th March that if the UK followed China (which it clearly wasn’t doing) the total number of deaths in the UK would be around 5,700, with there being a peak of between 210 and 330 people, possibly on 3th April.  Although he retracted this a few days later when it was blatantly obvious that his model was deeply flawed, news media who wanted a good news story had been very eager to publish his suggestion that the pandemic would not be as bad as others had predicted (he certainly got lots of pictures published of himself in his lab coat).  Likewise, at the other end of the scale, the IHME in the USA predicted that the UK would have 66,314 deaths in total by 4th August, rising to a peak of 2,932 deaths a day on 17th April.  This  might have been wishful thinking, because on 7th April, UK reported deaths were only 786, which was substantially below their model prediction of around 1250.  By then, though, their research had already hit the news headlines with lots of publicity.  Subsequently (as at 11th April), they revised their predictions to a peak of “only” 1,674 deaths a day (estimated range 651-4,143) with a cumulated total of 37,494 deaths.  These differences are very substantial, and emphasise that scientists often get it wrong.  Put simply, the UK government cannot hide behind science.  They can try to take the credit, but government leaders must also admit it openly when they have been wrong with the policies that they make based on the evidence.

In conclusion, by sharing these thoughts I have sought to:

• Ask the UK government to be more open and transparent in the information that it provides about Covid-19;
• Plead with media of all sorts to use data responsibly, and to be critical of claims by governments and scientists who all have their own interests in saying what they do; and
• Encourage everyone to work together for the common good, openly and honestly in trying to respond to the Covid-19 pandemic.

Above all, I write with huge respect for the many people in our NHS who have been working in the most difficult of circumstances to try to stem the tide of Covid-19.  Too many of them have already died; too many of them have become sick.

[Update 12th April: A report in The Times notes that “The death rate of Covid-19 patients admitted to intensive care now stands at more than 51 per cent, according to a study on a sample of coronavirus patients”.  The original report is by ICNARC, which showed that “Of the 3883 patients, 871 patients have died, 818 patients have been discharged alive from critical care and 2194 patients were last reported as still receiving critical care”. I should add that this is despite the very valiant efforts of our NHS staff]

[Update 14th April: Great to see that the BBC is at last reporting more responsibly about government reported deaths (based on those in hospital) being a serious underestimate of total deaths, and comparing trends of deaths with previous years – two useful graphs included and copied herewith below

    

Thanks BBC]

Updated 14th April