Anshul Kundaje argues Bayes' theorem handles scanning false positives, while an analysis blames differing patient demographics

@BenMazer I’m hearing that bayes theorem means you can just ignore incidental findings you don’t want to follow up

Is it, tho? Prenuvo has been around for quite a while. Incidentally their issue is false negative. Rads got sued for missing obvious things later.

The supposed harm if false positive in screening test is consistently oversold. Mammography isn’t especially effective and there’s a huge false positive rate. But yet very few were “harmed” by subjective report.

@RBMD1982 @BenMazer Can you elaborate here with some earnestness?

... can't you ignore them?

I can mentally model your surgical colleagues as mindless knives who must be kept in the dark about deep-blips.

But surely there's an in-principle-possible way to do sane diagnostics given scans?

@RBMD1982 @BenMazer Lmao

@seanluomdphd @BenMazer Mammography is in high risk populations using a highly predictive measurement. The issues arises when doing low predictive measurements in a low risk population. It’s pure noise and would imply biopsy for 90% of the population.