Report bias in drug research.

Selective reporting of clinical trials to regulatory authorities could result in inappropriate marketing authorization decisions. The present study was conducted to uncover such potential bias, here called 'report bias'. In 1987 the number of clinical drug trials notified to Finnish National Agency for Medicines was 274. Up to the end of 1993, a final report or a statement that the trial had been suspended had only been received in the case of 91 studies. The outcome of 183 trials was not reported. These trials were not significantly different from those reported on, in their design, in blinding, or in the use of a comparator. The distribution in phases I to IV was similar in the reported and non-reported groups. Thus incomplete reporting was not caused by differences in the investigational plan. Other factors, such as differences in results, may be responsible. It is possible that findings that do not support commercial interests are less likely to be reported. This potential bias requires to be investigated and considered.