Drug treatment likely to be based on biased evidence

Drug treatment is likely to be founded on biased evidence because drug companies tend to publish studies with more favourable results, suggest researchers in the BMJ.

They identified 42 studies submitted to the Swedish drug regulatory authority to secure marketing approval for five antidepressant drugs. These studies were then compared with studies actually published between 1983 and 1999.

They found evidence of three sources of bias: duplicate publication, selective publication, and selective reporting. For instance, 21 studies contributed to at least two publications each, and three studies contributed to five publications. Studies showing significant effects of a drug were published as stand alone publications more often than studies with non-significant results. The tendency to report the more favourable results only, in studies actually published, was a major cause for bias.

These results should not be used to dispute the value of analysing the medical literature, say the authors.

However, they are likely to be valid for other classes of drugs, so for anyone who relies on published studies alone to choose a specific drug, they should be a cause for concern.

Without access to all studies (positive as well as negative, published as well as unpublished) any attempt to recommend a specific drug is likely to be based on biased evidence, they conclude.