Monday, December 06, 2004

Yes, but with our product, you'll be less dead...

So, news reports indicate that Celebrex is three times safer than Vioxx! Yay!

Well, Russian roulette is three times safer with one bullet instead of three. That doesn't make it a good idea.

I'm not trying to beat up on Celebrex here. For all I know, it's a miracle drug that'll cure your gout and make your wife love you more. But I am picking on Celebrex because this article is an example of the poor health reporting we see all the time.

Both journalists and researchers share a common bias -- they want to find something new. Both professions thrive on a "publish or perish" mentality, and the best way to get published is to report new things first.

As a result of this pernicious bias, we end up with articles full of misleads like this one. We read along, blissfully thinking there's scientific evidence that Celebrex is safer than Vioxx, and it's not until the 6th paragraph that we get our first indication this finding might not live up to its banner headline. A full third of the way into the article we learn, "But the researchers also found that patients using either drug were not at a significantly greater risk of having a heart attack than those who did not use either drug."

Huh? I thought the FDA pulled Vioxx because it caused heart attacks. But this study didn't find Vioxx caused heart attacks. So what can I learn here? Why should I care how safe this study tells me Celebrex is when it's also telling me Vioxx is safe? I mean, if a guy is telling me grenades are safe for children, should I listen to him when he tells me that land mines are three times safer?

I'm not suggesting that anyone was cooking the data or falsifying their sources. Check out the original journal article -- it's telling the same story as the article, albeit with more exciting math. What I'm saying is that journalists and scientists really wanted this study to have something newsworthy in it. So they found something.

It's bad science and it's bad journalism. Waiting until the end of the article to quote another expert as saying "
these types of studies are not conclusive," isn't enough. I mean, it was important enough to get that headline and a full article! Can you blame a reader for thinking there's some news of importance here? A little "he said, he said" isn't going to meaningfully challenge that assumption. When a headline should actually read, "Jury Still Out on Celebrex's Health Risks," a few last minute hedges aren't really going to undo the false impression created in paragraph one.

So what's the answer? Scientists have to stop inflating the significance of their findings, and reporters need to be less credulous about the integrity of scientists when it comes to the significance of their findings. Since neither of those things seems likely to happen, I suggest we read every health article from the bottom up -- start where the facts are, and end with a good laugh at the overblown headline.


Anonymous Anonymous said...

I am looking for some well written articles on any subject to add to
my article site directory. Please visit my website and post your article
today...if you need some tips on how or why to write an article you can
read some of the articles already posted.....Best Of Success

10:25 AM


Post a Comment

<< Home