The Cochrane Collaboration: Fact-Checking Science

From the hysteria that the very mention of comparative-effectiveness research (CER) triggers in some quarters, you'd think that studying which medical treatments work and which do not was one step away from setting up a government-run marketplace in body parts. As I wrote in a column earlier this year:

"The Washington Times said effectiveness research will 'threaten the lives of many Americans' as government decides 'who gets lifesaving treatment and who doesn't.' Sen. Tom Coburn of Oklahoma (a doctor) warned of 'a Soviet-style Federal Health Board that will put bureaucrats and politicians in charge of our nation's health-care system.'"

So I took greater interest than usual in the latest results from the Cochrane Collaboration. The Cochrane is, legally speaking, a charity registered in Britain. But it is also the world's largest group of volunteer scientists and physicians—about 20,000 at last count—assembled in “review groups” to study the effectiveness of different treatments. They don't conduct original research but, instead, review everything they can get their hands on that addresses the safety and efficacy of a drug or other treatment—exactly what President Obama envisions as part of health-care reform: why pay for stuff that doesn't work?

Contrary to claims of how difficult and controversial such research can be—any assessment will be overtaken by new research, according to one objection, or effectiveness cannot be reliably determined because too much research goes unpublished and is therefore hidden, according to another—the Cochrane manages to do it dozens of times each year. They tap scientists and doctors who conduct the assessments for free, out of a sense of public duty, David Tovey, editor in chief of the Cochrane Library, told me. Each review, based on data from a few to hundreds of individual studies, takes 12 to 15 months or so, with only minimal infrastructure costs, says Tovey. And only once since Cochrane was established in 1993 has one of its reviews been overtaken by events. In that case, the reviewers concluded that an anti-clotting drug used during heart surgery was of uncertain effectiveness compared with other such drugs; it was subsequently found to be less effective—a case of giving a drug the benefit of the doubt until more research emerged.

One of the best things about the reviews is that the scientists scrutinize each individual study they include for bias and methodological weaknesses, which means their assessments do much more than count how many studies give some drug a thumbs up and how many a thumbs down. With this rigorous approach, Cochrane has reached conclusions about hundreds of treatments. In 2006, it found that giving women at risk of preterm birth a single course of corticosteroids speeds up fetal lung development and "should be considered routine for preterm delivery" in order to "reduce the number of babies who die or suffer breathing problems at birth." Other reviews, all collected in the Cochrane Library, have found that remaining upright during the first stage of labor can reduce labor by an hour, that weight-loss surgery is more effective than other strategies for shedding pounds but may not be safe, that statins given to the elderly do not reduce the risk of Alzheimer’s or other dementias, that using chest X-rays to detect lung cancer in people with no symptoms doesn’t affect whether they survive and may increase mortality and that Echinacea doesn’t get rid of a cold any faster than doing nothing.

In its latest studies, released Tuesday evening, Cochrane concludes that Chinese herbal medicine may relieve symptoms of endometriosis that have been treated with laparoscopic surgery just as well as, or better than, standard hormonal drugs and with fewer side effects; that progressive resistance strength training helps older adults become stronger; that reducing blood pressure below 140/90—as some public-health authorities have recommended—brings no known clinical benefit, including any reduction in death rates or serious illness; and that psychological interventions intended to prevent post-traumatic stress disorder right after a traumatic experience have not been shown to be effective.

Really, with more than 2,000 such reviews on everything from the common cold to cancer already completed, Cochrane has shown that comparative-effectiveness research works. Of course, there are some questions for which there is not sufficient research to reach a conclusion, but with so many ineffective treatments in wide use—and running up the nation's health-care tab—surely we can at least use CER to eliminate the rotten, low-hanging fruit.