Cross posted from Scientific American’s Observations blog on behalf of Katherine Harmon.
Big clinical trials—to test new drugs or procedures—generate reams of important data about safety and efficacy. Only a fraction of that information sees the light of day, a publishing practice that could put patients at risk, according to a special report published this week in the British Medical Journal (BMJ).
Even though scientific and medical journals are loaded with what might seem like endless reports—and lengthy methodology descriptions–from clinical trails each year, about half of clinical trial results go unpublished, An-Wen Chan, of the University of Toronto’s Women’s College Research Institute, noted in one of the seven new papers in the BMJ special section. Published results typically lack details about how studies were conducted and outcomes for individual participants. Although these particulars might seem negligible or dull to a causal reader, “the overall result is that the published literature tends to overestimate the efficacy and underestimate the harms of a given intervention,” she noted.
Even major trials for drugs submitted to the U.S. Food and Drug Administration (FDA) contained substantial data holes in their published reports, according to one new analysis. A San Francisco- and Denmark-based team performed meta-analyses that included previously unpublished data for nine drugs that were submitted for FDA approval. With this newly included data, they found that 38 of 41 meta-analyses about these drugs were off: 19 of them overestimated the efficacy of the drug, and 19 of them underestimated it. It can be difficult to get a journal to publish—or a drug company to support publication of—harmful—or null, i.e. “negative”–results from a trial. But, as that paper’s authors noted, “when unfavorable results of drug trails are not published, meta-analyses and systematic reviews that are based only on published data may overestimate the efficacy of the drugs.”
Read the rest of this story on Scientific American.