The Tour de France has helped to lift the lid on a statistical hole in the heart of the SARS-CoV-2 testing regime.
Unlike with SARS-CoV-1, MERS and Ebola, the World Health Organisation (WHO) and Centers for Disease Control (CDC) seem to be OK with countries declaring SARS-CoV-2 infections based on single tests of people who show no symptoms.
With SARS-CoV-1, for example, the WHO ruled that “a single test result is insufficient for the definitive diagnosis of SARS-CoV infection”.
So in those earlier epidemics, the WHO and CDC recommended that countries (a) tested only people with symptoms, and (b) tested them twice to guard against a false positive test result first time round.
They changed that in 2019, just in time for the SARS-CoV-2 pandemic.
However, as Professor Andrew N. Cohen of the Center for Research on Aquatic Bioinvasions has pointed out to cyclingnews.com:
“…when you begin testing large numbers of healthy people with no known exposure to infected people, the chances of any positive become so small that the chance of the result being falsely positive skyrockets“.
But ‘testing large numbers of healthy people’ is exactly what France, the UK and most other countries are doing. In the last fortnight I’ve had two letters and a text from Imperial College London and Ipsos Mori explaining I’ve been randomly selected to send off for a Bad Statistics Fun Pack – sorry, COVID-19 testing kit.
That seems to be part of a global strategy of blanket testing for COVID-19 in the absence of symptoms, of which Prof. Cohen is an outspoken critic.
The CyclingNews journalist, Laura Weislo, explains how a very low false-positive rate of one in 1,000 tests (0.1%) will translate into over 50% of positive tests being wrong if you test a population with a daily infection rate of one in a thousand people. She shows how, when false negatives are also taken into account, testing 100,000 people will find 80 actual COVID-19 cases but wrongly identify 100 non-infected people as “cases”. Please see the extract at the end of this entry for the explanation or better still read her article in full.
Whenever you read that this or that government minister has pledged to increase the number of tests to so many hundreds of thousands per day, the vast majority of those tests will be done on healthy people because the actual number of people who have symptoms, or who have been in contact with someone who’s tested positive and has symptoms remains very small.
That suggests that over half the new “cases” the legacy media unquestioningly report every day are really COVID-19-free. And therefore governments are enforcing masks, local lockdowns and travel quarantines based on innumeracy, not the actual risk from the virus.
For example, here is a chart showing deaths vs. “cases” in France. There’s been no correlation since late May/early June:
As for the Tour de France, false positive syndrome belatedly forced the sport’s governing body to change its rigid policy of excluding any rider on the basis of a single positive test. The move followed fiascos where positive results led to several riders, and the whole Bora-Hansgrohe team, being turfed off other races in the run-up to the Tour, only for everyone involved to re-test negative immediately afterwards.
The TdF organisers got the point. The remorseless mathematical logic of false positives when intensively testing fit, healthy people – and racing cyclists are as fit and healthy as you can get – risked whittling the 196 starters down to a comical remnant of a peloton by the time the race reached Paris. Also, France has waited for a home-grown tour winner for 35 years. Having one snatched from her grasp in 2020 by a lousy false positive on Julian Alaphilippe or Thibaut Pinot would be absolument impensable.
So, Tour riders and support staff will now only be removed from the race for a positive test if a second test independently confirms it.
This has opened up a small crack in the official pandemic facade. But only it seems inside the bubble of the Tour and a few other “too big to fuck around with” sports, including NFL in the US. Outside sport, most countries are enthusiastically pumping up the number of new “cases” by over-testing the general population and applying the WHO guidance around confirming infections on the basis of a single test.
If the public learned from the media that over half of the positives turned up by the ICL/Ipsos study-cum-fishing expedition are statistically likely to be false, they might be less afraid and less tolerant of the ongoing restrictions. But they won’t of course, because the MSM is largely a shambolic mess dominated by groupthink. You’re better off reading CyclingNews if you’re after actual journalism.
People would be even more pissed off if they understood what how testing should actually work if governments were serious about it.
From the CyclingNews article:
“If I were running [the Tour de France], I’d think about setting up two swabbing stations, a few hundred yards apart, with separate staff, for riders’ left and right nostrils (although I might check to see if there are any studies on the probability of having detectable virus in one nostril but not the other), sending the specimens by different couriers to two different labs that use two different assays and running both specimens (or if timing allows, running the second only if the first is positive). That’s about as independent as one could get,” Cohen says.
“It would be good data too. Over the multiple testings during the race, they’d likely detect at least a few false positives, thereby gathering important information on the false positive rate in actual practice.”
The TdF probably won’t take up Prof. Cohen’s suggestion. But the fact that the race organisers now require two positive tests before declaring an infection shows that there is a double standard at work. There is a rigorous standard for nationally-important exceptions like the Tour, where false positives threaten to have political consequences. And another standard – statistically sloppy and at odds with previous WHO policy – for the rest of us lemmings.
Of course we need to trace people and isolate people who have had close contact with a definite infection – i.e. one confirmed by two swab tests.
But what is the point of randomly sending self-test kits to healthy people like me? Incidentally, I’ve reported over and over again to the C-19 study that I’m perfectly well.
That will only make things worse.
Though, if the rest of the official response to coronavirus is anything to go by, that seems to be the whole point.
Extract from When is a positive not a positive? COVID-19 and the Tour de France
By Laura Weislo | CyclingNews
Take 100,000 people, and assume 100 have COVID – that’s the average new infections per day in Alpes-Maritimes. You test them all. Up to 20 per cent of the tests can miss an active infection (either because people are tested too soon after exposure or they don’t stick the swab up their nose far enough) and be a false negative.
Let’s say that the lab being used is pretty good at their job and only 1 in 1,000 tests is a false positive.
80 COVID-19-infected people will be correctly identified, 20 will not because of false negatives. Of the other 99,900 healthy people, around 100 will test positive even though they aren’t infected and 99,800 will be negative.
Just looking at the positives, there are a total of 180. 80 are correct and 100 are not! Only 44.4 per cent (80 of 180) of the positives were correct. That’s the “Positive Predictive Value (PPV)”.
But if you do a second test to confirm the positive, the chances of getting a false-positive plummets because now you have two independent tests on the same person with the same outcome. One can be nearly 100 per cent confident of two identical results, but there’s only a 13.8% chance the person is actually COVID-19 infected if they have one of each.