DiagnosticTesting – Dr Cowan
Very nice, still the goto stats guy at Jacobi, huh Ethan?
It’s probably much more accessible with you talking folks through it. I’ll settle for getting across a few key points: why sens/spec are generally worthless take home points regarding a dx test, why LR is stable across populations (requires you to supply a pretest prob to interpret), why testing low risk populations always generates more false positives (insert screening discussion of choice:PE, prostate, breast). Looking forward to a post by you on the tyranny of the p value and the bizarre logic of non-inferiority vs superiority.
You must be logged in to post a comment.