>In 2006, Potti et al. had published a revolutionary paper in Nature Medicine, proposing using genomic signatures to guide the use of chemotherapeutics.
>Another Potti paper, published in the New England Journal of Medicine, proposed using genomics to assign early-stage lung cancer patients to treatment regimens.
>In both cases, the reported scales of improvements were dramatic.
Academic papers about medical treatment affect life/death of millions. This is frightening.
Academic papers about medical treatment affect life/death of millions. This is frightening.
Absolutely. If you read "The Emperor of All Maladies" you'll discover that a single research lab led by a (probably) fraudulent research lead caused the US healthcare system to spend around a $billion on an aggressive cancer treatment that was no more effective than the existing one back in the 90s. The treatment was extremely invasive and very expensive: the patient was pushed to the edge of organ failure by extremely high doses of chemotherapy which required close management by nursing staff and had to have a bone marrow transplant after the drug regimen was completed because their bone marrow didn't survive. Needless to say, the side effects of this treatment were horrendous & since chemotherapy drugs are usually carcinogenic it's pretty much guaranteed that people who received it that would otherwise have survived died later on as a consequence.
As far as I know the researcher in question has never faced any real sanction for their actions. When the fraud came to light they quit their post & refused to talk to anyone about it ever again.
I wonder if academic journals should be treated similarly to Linux distros (or other software projects really) where less proven/stable features are considered beta and published in journals like Arxiv (say, the Fedora to draw a Linux reference). Once the research has been proven or repeated, it can be published in more "stable" journal like NEJM or Nature. That way the original author still gets the credit, but you have more confidence that what you're reading in those stable journals is indeed truth.
Problem is, a lot of this stuff is expensive to reproduce, and you certainly can't get a grant to do that, so what normally happens is either the research is so unimportant it's ignored (a product of publish or perish, I suppose from both sides), or if it's "high impact" enough, other labs try to build upon it. If the foundation is rotten, they'll figure that out soon enough, and at worst case word will filter through the grapevine.
>Another Potti paper, published in the New England Journal of Medicine, proposed using genomics to assign early-stage lung cancer patients to treatment regimens.
>In both cases, the reported scales of improvements were dramatic.
Academic papers about medical treatment affect life/death of millions. This is frightening.