by James Ferguson

https://lockdownsceptics.org/?s=government+innumeracy

Matt Hancock and his closest advisors receive the latest modelling update from Prof Neil Ferguson

Are you positive you are ‘positive’?

“When the facts change, I change my mind. What do you do sir?” – John Maynard Keynes

The UK has a big problem with the false positive rate (FPR) of its COVID-19 tests. The authorities acknowledge no FPR, so positive test results are not corrected for false positives and that is a big problem.

The standard COVID-19 RT-PCR test results have a consistent positive rate of ≤ 2% which also appears to be the likely false positive rate (FPR), rendering the number of official ‘cases’ virtually meaningless. The likely low virus prevalence (~0.02%) is consistent with as few as 1% of the 6,100+ Brits now testing positive each week in the wider community (pillar 2) tests actually having the disease.

We are now asked to believe that a random, probably asymptomatic member of the public is 5x more likely to test ‘positive’ than someone tested in hospital, which seems preposterous given that ~40% of diagnosed infections originated in hospitals.

The high amplification of PCR tests requires them to be subject to black box software algorithms, which the numbers suggest are preset at a 2% positive rate. If so, we will never get ‘cases’ down until and unless we reduce, or better yet cease altogether, randomized testing. Instead the government plans to ramp them up to 10m a day at a cost of £100bn, equivalent to the entire NHS budget.

Government interventions have seriously negative political, economic and health implications yet are entirely predicated on test results that are almost entirely false. Despite the prevalence of virus in the UK having fallen to about 2-in-10,000, the chances of testing ‘positive’ stubbornly remain ~100x higher than that.

First do no harm

It may surprise you to know that in medicine, a positive test result does not often, or even usually, mean that an asymptomatic patient has the disease. The lower the prevalence of a disease compared to the false positive rate (FPR) of the test, the more inaccurate the results of the test will be. Consequently, it is often advisable that random testing in the absence of corroborating symptoms, for certain types of cancer for example, is avoided and doubly so if the treatment has non-trivial negative side-effects. In Probabilistic Reasoning in Clinical Medicine (1982), edited by Nobel laureate Daniel Kahneman and his long-time collaborator Amos Tversky, David Eddy provided physicians with the following diagnostic puzzle. Women age 40, participate in routine screening for breast cancer which has a prevalence of 1%. The mammogram test has a false negative rate of 20% and a false positive rate of 10%. What is the probability that a woman with a positive test actually has breast cancer? The correct answer in this case is 7.5% but 95/100 doctors in the study gave answers in the range 70-80%, i.e. their estimates were out by an order of magnitude. [The solution: in each batch of 100,000 tests, 800 (80% of the 1,000 women with breast cancer) will be picked up; but so too will 9,920 (10% FPR) of the 99,200 healthy women. Therefore, the chance of actually being positive (800) if tested positive (800 + 9,920 = 10,720) is only 7.46% (800/10,720).]

Click on the link for the rest.