Publication bias in clinical trials and economic analyses.

Publication bias is an established notion in the clinical literature; essentially, large studies or positive results are more likely to find their way into the public domain than small studies or negative results. Meta-analysis presents a focus for the problems of publication bias, seeking to summarise the evidence in a particular therapeutic area by retrieving and analysing all available clinical studies. Economic analyses are also vulnerable to publication bias, and at 3 levels: first, in the health-outcomes data available for modelling (resulting from publication bias in the clinical literature); second, in the motivations for conducting an economic analysis; and third, in repeating the process of seeking publication.