It's football season, so I begin this concluding post on the problem of doctors' conflicts of interest with an admittedly farfetched football-related scenario. Let's say that NFL commissioner Roger Goodell has hired you to make sure that the ceremonial coin used for the Super Bowl coin toss is a fair coin. Your testing procedure consists of flipping the coin some number of times and then tallying the heads and tails. If you flip a fair coin enough times, you're likely to come up with a roughly equal number of heads and tails. But what if you secretly work for a rival ceremonial coin-making company that has an interest in getting the NFL to switch business partners? To make it seem as if the coin isn't fair, you decide to selectively report all of the times it comes up heads, but only 20 percent of the times it comes up tails.
Medical journals are all too familiar with this phenomenon of selective reporting. In the case of research, it occurs when clinical studies funded by commercial interests are more likely to be submitted (and therefore, published) when they show a positive effect of the drug or device, and less likely to do so when the results show no effect. Later on, when a trusted international scientific review group such as the Cochrane Collaboration decides to summarize the results of all studies on this drug or device, publication bias may lead them to draw erroneous conclusions about its true effectiveness.
This is not a small problem in real life care situations. Take depression, for example. The introduction of the drug class of selective serotonin reuptake inhibitors in 1987, beginning with the drug Prozac, in many ways revolutioned the treatment of depression and brought its diagnosis and management into the realm of primary care. Recently, however, serious questions have been raised about the true effectiveness of these drugs, based on the results of a 2008 study published in the online journal PLoS Medicine that analyzed the results of all published and unpublished results submitted to the U.S. Food and Drug Administration and concluded surprisingly that for all but the most severely depressed patients, new antidepressants were no better than placebo pills. This finding doesn't necessarily mean that "the drugs don't work" - since effects were averaged over all patients who took the drugs, some people's depression may have improved, some may have stayed the same, and some may have gotten worse. But it does mean that the benefits of antidepressants have been systematically exaggerated in the medical literature, thanks to selective reporting.
How are medical editors and others fighting back against this now widely recognized problem? In 2004, the International Committee of Medical Journal Editors began requiring that investigators register clinical trials on publicly accessible registries such as Clinicaltrials.gov as a precondition for publication. Earlier this year, this group went a step further by establishing a uniform format for conflict-of-interest disclosures, in the hope of making it more difficult for industry-funded authors to "forget" to inform journals about whom they're working for. Finally, two of the health reform bills pending action in the House and Senate would require all drug and device-makers to file annual reports to the government, to be made freely available online, on payments to physicians and health organizations. Which gives me and all of those who are hoping that health reform passes soon in some form another reason to do so.
Note: if you missed the first two posts in this mini-series, you can access them here:
Part 1 - Drug Reps and Ghostwriting
Part 2 - Conference Exhibitions and Dishonest Disclosures