Utilization of clinical trials registries in obstetrics and gynecology systematic reviews
Objectives: We evaluated the use of clinical trials registries in published obstetrics and gynecological systematic reviews and meta-analyses.
Methods: A review of publications between January 1, 2007, and December 31, 2015, from six obstetrical and gynecological journals (Obstetrics & Gynecology, Obstetrical & Gynecological Survey, Human Reproduction Update, Gynecologic Oncology, British Journal of Obstetrics and Gynaecology, and American Journal of Obstetrics & Gynecology) was completed to identify eligible systematic reviews. All systematic reviews included after exclusions were independently reviewed to determine if clinical trials registries had been included as part of the search process. Studies that reported using a trials registry were further examined to determine whether trial data was included in the analysis.
Results: Our initial search resulted in 292 articles, which was narrowed to 256 after exclusions. Of the 256 systematic reviews meeting our selection criteria, 47 utilized a clinical trials registry. Eleven of the 47 systematic reviews found unpublished data, and added the unpublished trial data into their results.
Conclusion: A majority of systematic reviews in clinical obstetrics and gynecology journals do not conduct searches of clinical trials registries or do not make use of data obtained from these searches.
Systematic reviewers are known for using pre-specified, comprehensive searches to locate all relevant studies for inclusion. While thorough, this approach to searching may include an overrepresentation of published, statistically significant outcomes if reviewers search only popular databases such as MEDLINE or Embase. This occurrence is known as publication bias, which Dickersin defines as, “the tendency on the parts of investigators, reviewers, and editors to submit or accept manuscripts for publication based on the direction or strength of the study findings” (Dickersin 1990).(p1385) To mitigate this issue, searches should attempt to use alternative information sources that contain grey literature and unpublished data.
A recent survey of Cochrane systematic review authors examined their experiences when retrieving and using unpublished data for systematic reviews. Non-commercial clinical trials registries accounted for 6.3% of sources reported (Schroll 2013). The infrequent use of trials registries reported by Cochrane reviewers is surprising since they offer promise in locating unpublished trial data. Furthermore, a number of precipitating events have made clinical trials registries a viable information source for systematic reviewers.