Related Experiment Videos
Publication bias and clinical trials
K Dickersin1, S Chan, T C Chalmers
1Clinical Trials Unit, Mount Sinai School of Medicine of CUNY, New York.
Controlled Clinical Trials
|December 1, 1987
Summary
Publication bias significantly skews medical literature. Unpublished randomized clinical trials (RCTs) often show negative results, unlike published studies, impacting meta-analysis and trial interpretation.
Area of Science:
- Medical research methodology
- Clinical trial publication ethics
Background:
- Selective publication of randomized clinical trials (RCTs) can distort the medical literature.
- Publication bias, where studies with statistically significant positive results are more likely to be published, is a concern.
Purpose of the Study:
- To assess the extent of publication bias in medical literature.
- To investigate the characteristics of unpublished randomized clinical trials (RCTs).
Main Methods:
- Surveyed authors of published trials regarding participation in unpublished RCTs.
- Collected data on published and unpublished trials from respondents.
- Analyzed the publication trends of completed trials.
Main Results:
- Respondents reported 271 unpublished and 1041 published trials.
- Only 14% of completed unpublished trials favored new therapy versus 55% of published trials (p < 0.001).
- Major reasons for nonpublication included negative results and lack of interest, suggesting failure to submit rather than rejection.
Conclusions:
- A significant publication bias exists in the medical literature.
- This bias impacts the reliability of meta-analyses.
- Careful interpretation of statistically significant positive trial results is crucial.