My regular readers (both of them) have noticed that I spend almost as much time writing about new studies you should ignore as about new studies you should pay attention to. That’s because the media is driven by hype, not by sober science, and there’s no incentive for an editor to get rid of a story just because the study is misleading or meaningless. (I’m not complaining. That’s a consequence of having a free press, and it’s much better than the alternative.) So it’s up to us, the consumers of news, to figure out how to separate fact from fluff.
The upcoming LA Times health section on Monday has an important article that helps us with that task and explains why there are so many misleading studies out there. You may have noticed in my previous posts that what separates reliable from unreliable studies is whether the study was observational (also called epidemiologic) or randomized. Don’t get intimidated by the jargon; the difference is simple. Pretend I want to test whether drinking four cups of coffee per day increases the risk of heart attacks. An observational study would consist of me getting lots of people, inquiring about their coffee intake (or actually having assistants observe them and measuring their coffee intake) and then following them to see how many heart attacks they have. A randomized study would involve enrolling lots of volunteers and randomly splitting them into two groups. One group would be forced to drink four cups of coffee daily, and the second would not be allowed to drink coffee. These groups would then be followed to see how many heart attacks happen in each group.
The Times article does a great job of detailing the reasons that observational studies are unreliable. Even if coffee has no effect on heart attacks, my imaginary observational study above may show an effect simply because some other factor (sleep deprivation? smoking?) causes people both to drink more coffee and have more heart attacks. The only way to know for sure is to do a randomized trial. Observational studies have caused enormous confusion and much harm to patients. The article relates how physicians were fooled for a generation into believing the health benefits of estrogen replacement on the basis of observational studies. Randomized trials shocked us by proving the opposite. Epidemiology has fooled us again and again, about estrogen, vitamins, the effects of diet on cancer, and many other subjects. Hopefully the Times will help us all ignore such studies in the future.
Tangential Miscellany:
Did you know that anyone certified by the American Board of Internal Medicine after 1990 has to retake an examination and complete other requirements every ten years to maintain certification? It’s true, and in a field that changes as quickly as internal medicine, I think it’s a very good idea. If you’ll excuse a moment of bragging, I took the examination a few months ago and I’m happy to announce that I’m certified for another ten years. You can check if an internist is board certified here.