Difficulty

VW and TL – two upcoming colleagues of my department – find my blog on academics reflect (part of) their difficulties.

On a second thought, that's not a problem of new researchers, but the one for medicine as a whole.

Why could we have good single center randomized trials thirty years ago on a few hundred subjects ? Because there were common conditions with high mortality (or morbidity). Once you have a treatment that improves the adverse outcome from, say, 40% to 10%, you don't need many patients to prove the point. (My software says we need 76.)

Very soon, that 10% further comes down to 5%. What can you prove next ? To bring it down to 2% ? You know what: we need 1304 subjects – take aside the possibility of drop out.

Are we getting at a depressing truth that there will soon come to a limit no medical “breakthrough” could practically be proved by clinical trials ?