Article type
Year
Abstract
Background: Trial registries are essential sources of information on unpublished and ongoing trials. Searching two trial registries, the ClinicalTrials.gov (CT.gov) and the World Health Organization’s International Clinical Trials Registry Platform (WHO ICTRP), is now mandatory in Cochrane reviews. Minimal information currently exists on how well search functions work on each registry and the potential (or lack of) overlap in trial records retrieved from these registries.
Objectives: To quantify any differences in CT.gov records retrieved using basic and advanced searches in CT.gov and WHO ICTRP.
Methods: We first searched for trial records on ‘breast cancer’ or ‘breast neoplasm’ on the two registries and compared the number and record identification number of CT.gov trials using basic and advanced search functions (Table 1). We then narrowed the search to a specific Cochrane review topic (‘taxanesfor early breast cancer’; Table 2 for search terms) and repeated the analysis.
Results: For the ‘breast cancer’ basic search, CT.gov retrieved 4687 CT.gov records and WHO ICTRP retrieved 3960 CT.gov records (Tables 1, 3). WHO ICTRP detected an extra 50 CT.gov records not retrieved by CT.gov while CT.gov retrieved an extra 777 records. In the advanced ‘breast cancer’ searches, a different total number of CT.gov records were picked-up (Tables 1, 3) with both registries retrieving different CT.gov records as well (i.e. CT.gov = extra 469 records; WHO ICTRP = extra 48 records). For the specific ‘taxanes’ search, basic and advanced searches on CT.gov retrieved 11 and 10 records (respectively) while WHO ICTRP picked-up 5 and 2 CT.gov records (respectively). Again, WHO ICTRP found an extra 2 CT.gov records and CT.gov an extra 6 CT.gov records (post de-duplication).
Conclusions: Multiple basic and advanced searches in both CT.gov and WHO ICTRP registries are necessary to detect all potential CT records; WHO ICTRP detected an additional 6–10% of CT.gov records.
Objectives: To quantify any differences in CT.gov records retrieved using basic and advanced searches in CT.gov and WHO ICTRP.
Methods: We first searched for trial records on ‘breast cancer’ or ‘breast neoplasm’ on the two registries and compared the number and record identification number of CT.gov trials using basic and advanced search functions (Table 1). We then narrowed the search to a specific Cochrane review topic (‘taxanesfor early breast cancer’; Table 2 for search terms) and repeated the analysis.
Results: For the ‘breast cancer’ basic search, CT.gov retrieved 4687 CT.gov records and WHO ICTRP retrieved 3960 CT.gov records (Tables 1, 3). WHO ICTRP detected an extra 50 CT.gov records not retrieved by CT.gov while CT.gov retrieved an extra 777 records. In the advanced ‘breast cancer’ searches, a different total number of CT.gov records were picked-up (Tables 1, 3) with both registries retrieving different CT.gov records as well (i.e. CT.gov = extra 469 records; WHO ICTRP = extra 48 records). For the specific ‘taxanes’ search, basic and advanced searches on CT.gov retrieved 11 and 10 records (respectively) while WHO ICTRP picked-up 5 and 2 CT.gov records (respectively). Again, WHO ICTRP found an extra 2 CT.gov records and CT.gov an extra 6 CT.gov records (post de-duplication).
Conclusions: Multiple basic and advanced searches in both CT.gov and WHO ICTRP registries are necessary to detect all potential CT records; WHO ICTRP detected an additional 6–10% of CT.gov records.
Images
