Arguments have been raging in social media about the accuracy of the reports we, the public, are being given about Covid and the oncoming “‘second wave”.
We have been told by the government’s advisors that if we do nothing the second wave could be worse than the first, with infection increasing exponentially followed by the NHS under renewed pressure and a dramatically increasing death rate. One member of the advisory group, SAGE, has even suggested that the increased restrictions announced by the Prime Minister on Tuesday not only fail to go far enough, but fall so dramatically short that, we have to assume, the pandemic wave to come will be worse than anything we can imagine.
All this is supported by pages and, in the press briefings, screens, of statistics, all of which the main media, especially the BBC, have devoured with apparent glee.
But then, to quote the old saying, there are lies, damned lies and statistics.
Certainly Covid-19 as a world-wide virus exists, and certainly it is highly infectious. It is also, as with other respiratory illnesses, seasonal with the most acute periods being spring and autumn. But do the figures really require a complete shutdown of all business and social activity? And is lockdown, mask wearing and social distancing really an answer?
Back in the spring, little was known about the virus and the restrictions introduced then were designed to flatten the curve of the graph usually followed by this type of virus (Spanish flu, bird flu, swine flu are all examples) which shows a rapid steep incline to a peak as the virus takes hold, followed by a more gentle decline as it tails off.
The government said they would follow the science, and science predicted little short of Armageddon if nothing was done. The result, as we know, was lockdown, a massive blow to industry, commerce and jobs, a delay in all other medical work (itself, resulting in further deaths) and a disaster for the country’s economy which will take years to recover.
Statistics (yes, them again) put forward by a number of eminent authorities both in the UK and abroad, suggest that in fact the infection and death rate differed little from a number of previous years, following a similar pattern, and was indeed below the figures recorded 2000.
The second wave was easily predicted because that is the way a virus works, with the initial tsunami being followed by a wavelet or two until there is barely a ripple.
The reported death rate on September 22 was 37, up only marginally from the beginning of the month and with the weekly average remaining more or less constant. But the “scientists” are advising immediate precautionary measures, often going much further, and increasing the damage to our economy and general wellbeing, than the government, thank goodness, are prepared to accept at the moment.
The daily rate of infection is now a little over 6,000, up significantly from last week. But are either of these figures an accurate reflection of an increase in the virus? I would suggest that the answer is that it is not, and for the following reasons:
Research into the UK by the European Centre for Disease Prevention and Control (European CDC) which has looked at the effect of pubs re-opening since June has found that despite an estimated 450 million pub visits during which no masks have been worn and social distancing was, at best, variable, the curve of the virus overall has continued to flatten. The surge in a few areas has been found to be largely (but I accept, not entirely) due to other, mainly family-orientated social activities, rather than general mixing in the pub.
There is also the question of the death rate. Recorded deaths are reported as deaths where the deceased has tested positive for Covid-19 within the previous 28 days. So if someone tests positive and then has, for example, a medically-unconnected heart attack and dies, they are on the Covid list despite the fact that they did not die of it.
Returning to the numbers of positive tests, the more tests that are carried out, the more positive cases of the infection will be found. But then there arises the question of the accuracy of these “positive” findings.
Dr Andrew Bamji, an occasional and well informed contributor to Rye News, has written to our MP putting the case for concern about the value of positive finding statistics. With his permission I quote a section of his letter here:
“I am a lockdown agnostic and have been blogging about coronavirus since February. I have written to you previously regarding what I consider the alarming lack of focus on the treatment of the severe consequences of SARS-CoV-2 infection, namely Covid-19. However I am now alarmed by the misinformation on supposedly rising numbers of cases, which is provoking new lockdowns. This is generated by a failure to understand the statistics of testing. Put simply, if the false positive rate of a test is 0.8%, then for every 10,000 people tested there will be 80 false positives. The ONS (Office of National Statistics) figures predict a prevalence of 1 case per 1000. This would generate 10 positive tests per 10,000.
“Therefore one expects 90 positive tests, but only 10 of these are real positives. Thus there is a huge overestimation of the real numbers of cases (I would add that if you have two positive tests then the likelihood that you have a real positive jumps to over 90%).
“The argument is elegantly set out at https://lockdownsceptics.org/lies-damned-lies-and-health-statistics-the-deadly-danger-of-false-positives/
“There is significant backing for the veracity of this conclusion.
See https://www.hdruk.ac.uk/projects/false-positives/ and https://www.medrxiv.org/content/10.1101/2020.04.26.20080911v3.full.pdf
“What is worse, there has been no scientific rebuttal of this conclusion from the government’s scientific advisors. I conclude that the analysis above is correct, that estimates of new cases are grossly overstated – by a factor of nine – and that the extremely damaging effects of further lockdowns will be heightened by continuing reliance on bad science, as they are not necessary.
“The mismatch between what one expects from ONS figures and the numbers of recorded positive tests suggests that the large majority of current numbers of reported new cases from community testing are false positives. Given this, are further lockdowns justifiable? And should not repeat testing of positive subjects become mandatory.”
He adds further that it is the opinion of a group of respected physicians led from
Oxford (Professors Sunetra Gupta, Carl Heneghan and Karol Sikora) that targeted
protective measures are more appropriate than blanket ones
(https://twitter.com/ProfKarolSikora/status/1307972101463212032).
Should we therefore obey the restrictions that are being imposed on us? Yes, of course we should, if for no other reason than we will now be breaking the law, and subjected to heavy fines, if we do not.
But are the restrictions truly necessary? Will they control the virus, and does it indeed need to be controlled? Or should we treat it as just another seasonal version of flu, take the usual sensible precautions, vaccinate as and when a vaccine is available, but otherwise get on with our normal lives?
For those who want more detail on Covid-19, see Andrew Bamji’s blog.
Image Credits: CDC / Alissa Eckert, MS; Dan Higgins, MAMS https://phil.cdc.gov/Details.aspx?pid=23311 Public domain .
Thank you. A sensible discussion free from the anti lockdown/vac ranting that has clouded what should be a proper discussion. I have no medical or scientific training and represent the majority who are perplexed that the ‘experts’ seem incapable of finding common ground.
Years ago I was working on a major project in the telecom industry. A colleague in Holland was doing the same. Our conclusions were wildly different and being on different continents made it impossible to get to the root. Our boss ordered us to HQ and told us we would spend from 8am till 7pm every day locked in a conference room until we had agreed a joint approach. It took two and a half days and I look back on it as a most satisfying exercise.
I commend it to Boris. Whatever the man does will be wrong for someone and it really is time the discordant medical and scientific voices became a harmonious choir, signing a song we can all relate to.
John Minter writes: Put simply, if the false positive rate of a test is 0.8%, then for every 10,000 people tested there will be 80 false positives. The ONS (Office of National Statistics) figures predict a prevalence of 1 case per 1000. This would generate 10 positive tests per 10,000.”
That premise would result in 800 false positive cases per 100,000 population. The highest current reported infection rate in the UK is around 280 cases per 100,000. The lowest is about 4 per 100,000. I find it difficult to reconcile this with a false positive rate of 0.8% or anything remotely resembling that figure. The numbers don’t match and in addition one would expect the false positive rate to be much the same everywhere – they should be similar in Dorset (4 per 100,000 recorded cases) as in Blenau Gwent (280 per 100,000). That would suggest that at least in areas of high infection false positives play an insignificant part. Sir David Spiegelhalter from Cambridge University is quoted in the Independent recently as follows:
“Speaking to the BBC, Professor David Spiegelhalter from the University of Cambridge said that the figure touted for a false positive rate of 0.8 per cent “seems far too high” when looking at other ONS surveys.
“The ONS survey [from June] did 112,000 tests and only got 50 positive tests out of it,” he said, noting that even if all of these were false positives, the rate would be under 0.05 per cent.
He described the false positive issue as “a complete red herring” that was distracting from the actual issue of a rapidly spreading virus.”
https://www.independent.co.uk/news/uk/home-news/coronavirus-false-positives-testing-covid-19-test-b550133.html
Testing has been largely confined to those with possible Covid symptoms and those who have been in contact with them. The percentage of those tested in those circumstances who have tested positive has varied at different times and according to different reports but the lowest figure I have found has been 2.5%. With the current surge in cases, the percentage of positive tests is said to have become significantly higher.
Even at the figure of 2.5%, allowing for 0.8% false positives and no false negatives, two thirds of the positive tests have been correct, and the higher the percentage of positive tests goes, the less significant the false positives become.
Whatever the best approach to Covid may be, the arguments around the alleged 0.8% of false positives do not support the proposition that most positive results are false.
Far be it from me to dispute your very detailed argument, Michael, but neither am I, as a mere layman, in a position to tell the ONS that they don’t know what they are talking about. You quote Professor Spiegelhalter, and he may indeed be correct. The problem is, however, that you can find any number of eminent professors and scientists and they will all give you a different answer. The truth of the matter really seems to be that nobody knows, and every scientist anxious for their 15 seconds of fame on the BBC or in the national papers, has their own interpretation of the maths.
A couple of other figures to throw into the mix:
Sir Patrick Vallence (who, along with Chris Witty must surely by now be totally discredited) quoted a figure of up to 50,000 infections shortly if lockdown restrictions weren’t increased, which, although there appears to be absolutely no ‘science’ to support this, would perhaps be a good thing, since, on this basis, by the end of October we would all be immune and therefore restrictions could end. (Of course, we could also all be dead!)
The other figure: last week I understand that around 110 people died with – but not necessarily because of – Covid 19. In the same period, a little over 1,000 died of flu. On the basis of those statistics, flu would seem to be by far the greater concern and yet there is no suggestion of mass testing for this, nor are we being asked to self isolate if we happen to know someone who has contracted flu.
It is deeply worrying when the inmates appear to be in charge of the asylum.
Hello John. I tend to agree with you.
The figures from the ONS are the ones I go with.
I have heard Professor Sunetra Gupta speak and her view is that older people should shield or take precautions not to catch the virus whist the younger population need not worry so much and as a result herd immunity or something similar (I know this expression is disputed) will result. Then the older people can emerge and the likelihood of them catching the virus from the younger population will be considerably lessened. She makes the most sense of any expert I have listened to.
I do feel we should be more optimistic about the spread of this virus and decide for ourselves how we wish to lead our lives during the final stages of this pandemic.
If only the media and news outlets would report some positivity and real facts then the nation’s spirits might rise.
But well done and thank you for your report. A start!
John, I wouldn’t dream of telling ONS they don’t know what they are talking about. Having checked what ONS does say, it is apparent that the story about 0.8% false positives is an urban myth. What ONS does say is:
“Test specificity
Test specificity measures how often the test correctly identifies those who do not have the virus, so a test with high specificity will not have many false-positive results.
We know the specificity of our test must be very close to 100% as the low number of positive tests in our study means that specificity would be very high even if all positives were false. For example, in the most recent six- week period (31 July to 10 September), 159 of the 208,730 total samples tested positive. Even if all these positives were false, specificity would still be 99.92%.”
https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/methodologies/covid19infectionsurveypilotmethodsandfurtherinformation#test-sensitivity-and-specificity
In other words, if by some miracle, all 208,730 people were in truth negative, the rate of false positives would be 0.08% (one tenth of the figure you suggest). ONS did not say, and does not seem to have checked, how many, if any, of the positives were in fact false, but 0.08% would be the absolute ceiling.
Flu deaths: The official figures for the last few years are in Table 4 in https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/895233/Surveillance_Influenza_and_other_respiratory_viruses_in_the_UK_2019_to_2020_FINAL.pdf#page=54
This shows annual deaths associated with influenza over the past few years as between about 4000 and 22,000 a year. As with Covid, “associated with” means that some of those people will have died with influenza rather than of it. And we know that most flu deaths are in the colder months.
I suspect the figure you mention of 1000 deaths may have been a corruption by certain elements in the popular press of the figures in https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/deathsregisteredweeklyinenglandandwalesprovisional/weekending7august2020
This suggests that there have been about 1000 deaths a week where flu or pneumonia is present, but with no breakdown between them. Given the separate annual figures for flu, it would seem probable that most of those deaths were associated with pneumonia. The 45,000 to 65,000 excess deaths last spring were clearly well in excess of normal flu deaths.
I agree that it is deeply worrying when the inmates appear to be in charge of the asylum, but I would not rank Patrick Vallence and Chris Whitty amongst them. It is also deeply worrying when those elements in the popular press who never let the facts get in the way of a good story propagate fake news, which is then innocently spread by others, on something as important as this virus and its potential effect on our health and economic well being.
I am unconvinced that the false positive debate represents fake news. If it did, the seriousness of the issue shoule have provoked an immediate response from government advisors, which it has not. There are two problems: first, that there is still no reliable figure for the number of false positive or negative tests – there is a wide spread of estimates, and second, that respectable clinicians and scientists may be interpreting data differently. The short paper quoted from Health Data Research has nothing to do with the popular press; neither does a recent paper from “The Lancet” (https://www.thelancet.com/journals/lanres/article/PIIS2213-2600(20)30453-7/fulltext) which concludes “To summarise, false-positive COVID-19 swab test results might be increasingly likely in the current epidemiological climate in the UK, with substantial consequences at the personal, health system, and societal levels.” It outlines some of the reasons, which include the possibility that some positive tests are picking up virus remnants from old infection.
That said, the Huffington Post has an article rebutting the false positive story! The bottom line is that there remains considerable uncertainty as to the meaning of positive tests, and that what really matters is whether hospitalisation rates are rising – which they are, though not by as much as the surge in positive test numbers might predict (there is of course a time lag). If a positive test doesn’t matter, because the person is at low risk of developing severe disease, then the only reason they should consider isolating is if they might transmit it to a vulnerable person.
There is no information on where the current hospital admissions are, on where the ventilated patients are, on whether the hospitals they are in are treating Covid-19 to protocol. I am trying to find this. My recent correspondence with an American hospital which is working to a protocol (described in one of my previous blogs) suggests that the case fatality rate there is around 6%, compared to an average in the western world of 21%. If coronavirus does not kill people it’s not much worse than “ordinary” flu.
Sadly, I was lost in Andrew Bamji’s first sentence which is almost as convoluted as the small intestine and ends with the phrase ‘fake news’, and since we know this phrase was popularised by Trump to mean the exact opposite (every time he says something is ‘fake news’ it is demonstrably true news – Orwellian Doublespeak) I’m left cognitively constipated! I need a phrenic laxative!
However, on another earlier comment (by John Minter) which raises a comparison of current Covid 19 deaths versus flu deaths, I can find no statistical mention of 1,000 deaths from flu in a week in mid September in the UK! I can find reference though to deaths with/from flu and pneumonia lumped together. Some people with flu die of pneumonia, but not everyone with and who dies from pneumonia has flu. Far from it. My mum and aunt both died of pneumonia (as recorded on death certificates) but neither had flu!
I assume that John refers to the Prime Minister, his back-stabber and unelected henchman when he refers to the inmates having taken over the asylum – his description of Johnson, Gove and Cummings is, as ever, very generous!
Flu killed 10 times more Britons than Coronavirus last week,for the 14 th week in a Row. A little perspective goes a long way.
I would like to throw into the mix the concern about ‘long covid’ which affects thousands of coronavirus patients, including many who were not ill enough to be hospitalised. They have been suffering for months from fatigue and a range of other symptoms, with no certainty that these issues will resolve themselves over time. If herd immunity is going to be pursued by getting on with our lives as normal, there is a risk that a substantial proportion of the working population will have this debilitating condition. This alone should justify stringent measures to suppress the virus, I would have thought.
Many of these comments have clearly been the result of careful, and doubtless time-consuming research and often make fascinating reading. If there is one thing, however, that they demonstrate most clearly, it is that statistics can prove pretty much whatever you want, which supports my view, given in my response to an earlier comment, that the number of interpretations of the figures will be roughly equal to the number of scientists who are asked for their views. Or, in other words, nobody yet really knows the true answers.
Last week, Sir Ian Vallance was predicting Covid cases could rise exponentially and at his suggested rate of increase they should now stand at well in excess of 20,000 a day. In fact, at the time of the Prime Minister’s press briefing on Wednesday this week, the figure was only a touch over 7,100 – up barely 1,000 from this time last week. Deaths have certainly risen – up from the low 30s last week to 71 on Wednesday, but we do not know how many of these were specifically the result of a Covid infection or how much of the increase is due to the expected seasonal fluctuation in the death rate.
The current varying restrictions that we are seeing around the country may well supress the virus to a degree, as they did during the first wave in the Spring, but they will not by themselves eliminate it.
I appreciate the comment on ‘Long Covid’ from Dominic Manning, but the greatest concern must surely be the damage that is being done to the economy, the prospects of the younger generations together with the mental and physical health and wellbeing of all of us.
All of which still leaves the question, are we better off isolating ourselves from any possible contact with the virus or, with the clear exception of those who come into the most vulnerable categories, should we be allowed to go about our normal lives, taking just common sense precautions?
Among the long list of unanswered questions is whether “Long Covid” is any different from the postviral fatigue syndrome seen in influenza, glandular fever and other infections. I suspect it may be worse – but we don’t know for sure. Neither do we know whether early treatment reduces risk. Maybe next year. The probability of eliminating SARS-CoV-2 is extremely low; no other pandemic virus has ever been removed completely, except for smallpox. So I don’t think planning should be based on that expectation.
John Minter states above that the number of interpretations of the (Covid) figures will be roughly equal to the number of scientists who are asked for their views. He then proceeds to quote Sir Ian Vallance. Having chaired BT from 1987 to 2002, Sir Iain Vallance’s views on numbers are probably best restricted to telephones.
Of course John meant Sir Patrick Vallance, so let’s draw a pleated curtain on Sir Iain and put him to bed.
But let’s look at what Sir Patrick actually said and ‘predicted’. On the morning of 21 September (with figures for 20 September being the most recently available), Sir Patrick warned: “At the moment, we think that the epidemic is doubling roughly every seven days. It could be a little bit longer, maybe a little shorter, but let’s say roughly every seven days.”
On 20th September it was 3,899 new infections while on 30 September it was 7,108. Not quite double but getting there and within Vallance’s highly qualified warnings.
The risk of 20,000 new daily infections Vallance made clear was from a starting figure of 5,000 (not 3,899) and was after an exact doubling every seven days. It isn’t a fortnight yet, so there is no reason at all, with this clear modelling, why the number should be ‘well in excess of 20,000 a day’, and let’s wait till next Monday to see if the figure is 15,600 (using 3,900 as the base figure).
Hopefully not, thanks to the millions of ordinary Britons who are doing their bit for the country and their fellow citizens and are sticking to the rules, unlike the Prime Minister’s dreary father!
Andy’s last comment has, I think, proved the point I was making earlier, that figures can be made to prove anything you want, depending on how you use them. Interestingly, some sources are now suggesting that there are signs of the latest wave beginning to flatten out (if one ignores the student population where the rise in infections should have been entirely predictable, but not a matter for serious concern in a young, largely healthy, community) and this would seem to be borne out by the figures I produced in my last comment (taken from a fews days after those Andy quoted). However, as always, time will tell and the next days and weeks will give a better indication of success, or otherwise, in fighting the pandemic. There is probably little more to be usefully said on this subject right now, but there is no doubt that it will crop up again in the future.
For the latest ONS report, published today, comparing deaths from coronavirus, pneumonia and flu see:
https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/deathsduetocoronaviruscovid19comparedwithdeathsfrominfluenzaandpneumoniaenglandandwales/deathsoccurringbetween1januaryand31august2020
A friend of mine in her 60s living in Wealden who was visiting for an outside lunch this weekend has just contacted me – distraught – her son in his 20s who lives in her household has just tested positive for covid (with none of the NHS main symptoms).
She is the main carer for her father who is in his 90s and lives alone.
Let us not forget the human face and the complexity of how the generations depend on each other.