Lies, Damn Lies, and COVID-19 Statistics

The case of the disappearing antigens… A tale of two tests: Vermont town left puzzled by positive, then negative, COVID-19 results. Or – the only place I could actually view the article… The Wayback Machine’s copy.

So a small community in Vermont had 65 people test positive via the “new and improved” antigen test. But in Vermont that isn’t enough…

But in Vermont, positive antigen tests — a relatively new way of detecting an active COVID infection — require confirmation by another test: a polymerase chain reaction, or PCR, test, the gold-standard used by the government, hospitals, and professional sports teams. Of the 65 people who’d received those positive antigen tests, 48 would ultimately test negative via PCR. Liebig was one of them.
A community on edge was now a community confused.

Say that again. Out 65 people who tested positive with the antigen test, 48 tested negative with the better test. That is a 73.8 percent false positive rate. Only 26.2 percent of the people the first test said had the disease, actually seem to have had the disease.

If you read the fine print, the antigen test is still experimental, approved by the FDA under “emergency” measures for the pandemic

And not all states are as fastidious as Vermont. Rhode Island and Massachusetts accept the antigen test and publish the results.

No one is saying that people are lying, though the manufacturer of the test and equipment is in ass-covering mode. And I’m left with the feeling that people just want the data to be as “bad” as possible.

2 thoughts on “Lies, Damn Lies, and COVID-19 Statistics

  1. Pingback: In The Mailbox: 07.30.20 (Evening Edition) : The Other McCain

  2. Vermont is functioning as a large scale clinical trial for this kind of test. Expect there to be hiccups along the way.

    The challenge of this test, and all the COVID tests, is to balance sensitivity against accuracy. Too sensitive and you get too many false positives; not senisitive enough and you get too many false negatives. Hyper accurate tests will return too many false negatives when used to find a virus that mutates. The test has to be broad enough to catch all the mutations. On the other hand, marginally accurate tests will give too many false positives, triggering on this virus and on the other 6 coronaviruses that can infect people.

    I think it is a misnomer to call the PCR test a gold standard, seeing that the test itself is made by hundreds of different companies and the test sample analysis methodology varies considerably from lab to lab. Not to mention that at least one widely used brand of PCR test was recently shown to have a 30% false positive and a 20% false negative rate, when used to test the “gold standard” biological samples that every test on the market is proofed against.

    Perhaps someday, as it appears the ‘rona is not going away anytime soon, getting tested will mean submitting mucus, saliva, and blood samples, which are then run through several kinds of testing.

    Or we could just stop wasting time with testing, seeing that we’re all going to get it sooner or later. Maybe it is better to focus on prophylactic medicine for the whole population, multi approach treatments for those showing strong symptoms, and extreme efforts for those severely ill. At this point the medical community should have enough experience to come up with best practice treatment standards and get everyone on board. Meanwhile, the really high risk among us will have to be better isolated without destroying their quality of life, but the rest of us have to get back to work and school, accepting that this virus, just like every other illness out there, comes with some level of risk. Risk that can not be avoided, or even mitigated beyond a certain minor amount (masks, hand sanitizer, etc) and not allowed to destroy society.


Comments are closed.