Researcher Urges Fast Cheap Ultrasound Scans Using Bayesian Methods

Of course I'm pro scanning! False positives are not a fundamental problem. Literal textbook example of Bayes' Theorem. Wtf is wrong with people

@JeffLadish The problem is that to actually correctly apply Bayes’ to the problem is difficult. You would need a posterior on future harm with no treatment, which itself depends on a bunch of terms that cannot be solved by more scanning alone.

@JeffLadish Not impossible to solve problems, by any means. But you’d basically need to run experiments to scan people and not treat things that by current standards of care we _would_ treat today

@JeffLadish But i think saying this is a simple application of Bayes’ elides a lot of difficulties with the determination of the probabilities you’d need to know to plug into the formula