Twitter isn’t good for nuance, so here’s what I really think about the SPRINT study news

I had a bit of a fun tweetstorm about this article by Gina Kolata at the NYT talking about the results of the SPRINT study. But Twitter is a terrible way for me to express myself clearly, and people are apt to take my full caps style as true rage. I mean, my avatar is Anger. Take me with a grain of salt.

Let me be clear: I take no issue with the story, the reporting, or the journalism. The NIH had a media briefing. This is news. They’re supposed to cover it.

My issue is that this is basically the release of conclusions without methods or even results. It hasn’t been peer reviewed. I don’t know the details. No one outside the study has checked the findings or statistics.

That’s important. Those who conducted and funded the study can’t be the ones to judge its merit. We have a problem with that when industry is involved. As Austin and I have pointed out, conflicts of interest are more than just financial. And there are even financial ones here.

It may turn out that this study is totally amazing. It may contain life-changing results that are widely applicable. But I can’t tell. What were the exclusion criteria? What were the outcomes? What were the absolute rate reductions? The NNT? What were the harms? What were the limitations?

I have no idea. Neither does anyone else who wasn’t involved with the study.

I wrote a piece for the Upshot last year where I came down hard on researchers who were doing a study on costs to develop drugs because they released findings and held a press conference before they published the manuscript. I think that was problematic. My feelings don’t change because this study was funded with public money.

Peer review and full vetting, along with publication in standard scientific format, are the checks and balances with which we try to make sure science is robust and valid. We’re having a real problem with that recently. This doesn’t help.

@aaronecarroll

Hidden information below

Subscribe

Email Address*