I have been pointing this out for months:
In places where the virus is relatively scarce, false positives may even outnumber actual positives — eroding trust in tests and, under some circumstances, prompting outbreaks of their own.
Some time after the election, this will become a bigger story in the media. May be not on November 4th, but likely during November or December.
BinaxNOW Covid test claims to have a 1.5% false positive rate and is now widely deployed to screen asymptomatic sample populations.
If less than 1.5% of the sample population are active cases of Covid-19, then this test will produce more false positives than true positives.
At a 1.5% false positive rate, this means that if 1 in 67 people currently have the disease (in my county it is currently estimated that about 1 in 500 people has an active case) then about half of the positives are false positives. That’s just how the statistics works out.
In my county, this test will find 7.5 (let’s round to 7) false positives for every true positive (1.5% of 500). Thus, false positives are 7x greater than true positives. False positives can be greatly reduced by immediately doing a second test and the rapid test manufacturers ask that positive tests be followed up with an RT-PCR test; however, we know for a fact this has not occurred in real life (example – OSU screening of all students).
The NY Times article fails to make the statistical problem clear by providing examples with numbers. Bottom line: if the incidence of currently active Covid-19 cases in your population is less than 1.5%, then the majority of positive test results are false positives. This is simple statistical math.
(Note – this has to do with conditional probability and is a problem when tests are used to screen essentially randomly selected asymptomatic people. When used to test symptomatic individuals, however, the test is more effective. Since 5% to 10% of that group seem to have the disease, this suggest the incidence of Covid-19 among symptomatic individuals is 1 in 20 to 1 in 10, in which case the false positive problem is negligible. The problem occurs when screening large asymptomatic population samples where the incidence of the disease is low – which is might be just about everywhere.)
Covid-19 is an awful disease for many, and perhaps 10x more likely to be fatal than influenza. The increase in case counts is due to many factors and not just false positives. False positives are part of the reason but we do not know how much because public health is not transparent on this issue.
But what do I know? I am an idiot with no qualifications in health care and my comments are for Entertainment Only.