Friday, February 23, 2024

Science is Dead, Long Live Science!

An AI Scientist
Oliver Wiseman and Vinay Prasad at Freep,  We’re Not Curing Cancer Here, Guys "Are leading scientists just making stuff up? Vinay Prasad breaks down the cancer research scandal."

A top cancer surgeon at Columbia University is under scrutiny after one of his research papers was retracted for containing suspect data. Twenty-six other studies by Dr. Sam S. Yoon, who conducted his research at the Memorial Sloan Kettering Cancer Center, have been flagged as suspicious by a British scientific sleuth called Sholto David. David raised the alarm after spotting the same images across different articles that described wholly different experiments. He has also found duplications and manipulated data in papers published by researchers at the Dana-Farber Cancer Institute in Boston that have since been retracted.

This news shocked me: leading scientists at some of the most respected research centers in the world, working on the very important and well-funded fight against cancer are. . . making stuff up. That seems bad. Really bad. And it poses a lot of unsettling questions, like whether we can really trust medical research at all. But maybe I am missing something. In search of reassurance, I called up an expert: oncologist, UCSF professor, the author of more than 500 academic papers, and Free Press contributor Vinay Prasad.

Here’s an edited version of our conversation. (Spoiler alert: I was not reassured.)

Vinay, how worried should we be about the problem of fraud in cancer research?

Extremely worried. There’s something very unique about all these papers that allows people to find the fraud, and that is they report the raw data, in the form of images. Most papers, though, do not contain images. The data is all hidden. The researchers only provide a summary of the data. You have to worry how much fraud you’d find if everybody provided all the raw data. I suspect you’d find a gargantuan amount of fraud. This is merely the tip of the iceberg.

Most laymen like me assume all the data is transparent in medical research. You’re telling me that’s not how it works?

Scientific papers are like someone’s dating profile on an app. They’re picking what pictures to show you and what stories to tell you. You don’t get to see the whole library of photos on their phone. Researchers are only presenting a sliver of what they’ve actually done. And just like a dating app on your phone, everything is inaccurate. . . .

I've always thought fraud in all science is more common than we have seen documented. Why?  The incentive is high in an era of publish or perish on steroids, and the scrutiny is low and hit and miss. Peer-review? Don't kid me. For the most part, reviewers look for obvious methodological flaws, grammatic mistakes, and rarely for deliberate deception. We all want to believe the manuscript before us is legitimate. 

The Wombat has Rule 5 Sunday: Country Girls up on time at The Other McCain

1 comment:

  1. For a dispiriting read on published research I suggest you read the classic paper: Why Most Published Research Findings are False.