[A new scientific source of bias: SILLY bias. Analysis of citations of BMJ's Christmas articles].

We analysed the scientific impact of systematic reviews and randomised trials published in the BMJ Christmas issues 1997-2006. The articles were mostly interpreted correctly as humorous, but the humorous dimension was overlooked with surprising ease. The result from one ironic-absurd trial on the effect of retroactive remote intercessory prayer for patients already dead or dismissed was taken at face value in 12/36 of the citing articles, and mortality data was unconditionally included in three systematic reviews. Thus, we document a new type of bias in medical research: Serious Idiopathic Loss of Ludic ironY (SILLY) bias, both in citation practices and in metaanalyses.