Good article, however the probability is a bit weird concerning the test false/true positives. something with a false positive rate of 5% with a prevalence of 1

Good article, however the probability is a bit weird concerning the test false/true positives. something with a false positive rate of 5% with a prevalence of 1/1000 only means a 2% true positive rate if you assume the only thing we know about the patient is their test result. If I go in complaining about flu symptoms and my test comes back positive for influenza, there's a much higher likelihood of me having the flu than whatever the prevalence is vs the false positive rate of the test. The choice of which test to administer is also a financial one for doctors, which fact researchers hate, because tests with a high correlation for problems leads to higher insurance payouts for preventative care, etc. Finally, while I do somewhat agree with the main point of the article, that doctors need to be trained better on these tests, right now the movement in testing, especially with imaging-based tests, is to quantify and extract metrics so that interpretation plays a much smaller role. Ultimately the first test administered after a visit needs to be chosen more carefully, however the odds aren't as bleak as the author suggests, in my opinion.