Hacker News Reader

Comment by virtuallynathan | Hacker News Reader

There’s quite a few medical practices which guidelines still recommend that do not show benefit in better or newer trials.

https://jamanetwork.com/journals/jama/fullarticle/1104821

View on HN

This paper says we should abandon practices when they are shown to not work. I don't think anyone will disagree with that. But the original post made it seem like there were practices which were adopted willy nilly. But this paper just shows the self correcting nature of science.

They had to write a paper saying that because it doesn’t happen. We are still doing these procedures today.

Do the interventional cardiologists want to stop stenting people? Not really. The orthopedic surgeons still want to mess with your meniscus, etc.

I mean... This is the scientific framework. There was a mistake and the mechanism to fix it is to write more papers. I know it's not fixed yet, but that's the procedure. Do you believe this should never have happened? Then I guess we disagree on how powerful human intellect can be.

I think its fine for this to occur, and we should expect it, but how long should it take to reverse a bad practice? Is a decade or more acceptable?

Of course not. Nobody is endorsing for that. The problem is these are hard problems to solve and forming a consensus is a hard problem in addition to it. If we keep flip flopping on every new data point we will have more misses than hits.

Reply on news.ycombinator.com