If you’re going to attack a position you don’t like and also attempt to claim the moral and intellectual high ground it’s a good idea to get your basic facts rights. Even more– and it’s so obvious it shouldn’t need to be said– if it’s a scientific paper you are attacking you need to make sure you’re attacking the right paper.
Alas for poor Thomas Sullivan, by day the owner of a medical communications company that helps bring together the pharmaceutical industry and physicians, by night a blogger who– surprise!– defends the pharmaceutical industry and its intimate relationships with physicians.
Earlier this week on his blog Sullivan thought he had found the absolutely perfect example illustrating the way “pharmascolds”– liberal do-gooder academics and crusading journalists– unfairly gang up against industry and the doctors who accept its money. Sullivan set out to attack a new study in JAMA Internal Medicine that showed, as I reported, that physicians who accept free meals from a drug company are more likely to prescribe that company’s brand name drugs.
The problem with Sullivan’s analysis is that although he links to the correct paper at the beginning of his post and summarizes its main findings, he then attempts to demolish the paper based on data from a different study on a similar topic that had appeared in the same journal a few months earlier.
From that point on Sullivan’s blog post becomes a comedy of errors. He reaches spectacularly wrong conclusions because he uses the data from the earlier paper to question the conclusions of the later paper. One of his many wrong points is that the study failed to achieve statistical significance. Of course if he had looked at the correct paper he would have seen that the results were highly significant.
He climbs up on a very high horse and recommends that “the authors and the JAMA internal medicine editors review the Wikipedia definition of statistical significance, to help them understand that drawing inferences on data that does not meet the null hypothesis may not be in the best interest of their journal. But then again, pharmascolds hold to the believe to never let facts get in the way of a good narrative. Unfortunately, their allies in the media hold that same view.”
This is really quite extraordinary. Sullivan has loaded his weapon, aimed, fired, and shot himself in the foot. Does Sullivan, the businessman proprietor of a commercial CME company, really believe he’s in a position to lecture genuine researchers and medical editors at a top journal about basic statistics?
Moving beyond the study authors and the journal editors, Sullivan then goes on to fling self-righteous criticism at media coverage of the paper, including stories published in the New York Times, the Wall Street Journal, and ProPublica.
I don’t want to pick on poor Tom Sullivan. I thoroughly disagree with his strong support for commercial CME and other industry activities designed to influence physician behavior. But I also acknowledge that, unlike almost everyone else on his side of the fence, at least he is willing to take an unpopular public position and defend it on a consistent basis.
But it’s also true that the flaws of this one post are just an extreme example of a much broader range of basic scientific, logical, and writing flaws that bedevil just about everything Sullivan writes. I think it’s unfair to expect a businessman to write intelligently and effectively about medicine, science, and healthcare policy. But Sullivan also claims to be a medical educator, and in the post discussed here he sought to school a whole slew of academic researchers, journal editors, and professional journalists. He’s simply out of his league. The post demonstrates as well as anything I’ve ever seen that commercial CME is about commerce and has little to do with real medical education.