Article type
Year
Abstract
Background: The 2011 Institute of Medicine (IOM) standards for systematic reviews recommend handsearching conference abstracts to identify otherwise unpublished research results and address potential reporting biases. We hypothesized that the sparse description of randomized controlled trial (RCT) primary outcomes presented in conference abstracts could be supplemented using trial registry information.
Objectives: To compare primary outcomes described in conference abstracts with those reported in ClinicalTrials.gov and assess the relationship between primary outcome reporting and industry sponsorship.
Methods: Eligible conference abstracts described RCT results, had a valid ClinicalTrials.gov registration number, and were presented at the 2007 to 2009 Association for Research in Vision and Ophthalmology meetings. Two reviewers independently extracted information from the abstract and ClinicalTrials.gov. Discrepancies were resolved by consensus.
Results: We identified 152 eligible abstracts; 40 explicitly reported at least one primary outcome in the abstract and primary outcome field at ClinicalTrials.gov. Of 80 primary outcomes reported in abstracts, 18% (14/80) agreed with one reported in ClinicalTrials.gov. Forty-nine percent (39/80) agreed partially, in that additional information (eg. related to time point) was present in ClinicalTrials.gov (17/39; 44%), the abstract (19/39; 49%), or both (3/39; 8%). Thirty-four percent (27/80) of primary outcomes reported in the abstract were not reported as such in ClinicalTrials.gov, although 6 were reported as secondary outcomes. Industry sponsorship was reported in 6/14 abstracts where a primary outcome was reported in the abstract but not in ClinicalTrials.gov, and in 14/26 abstracts where all primary outcomes reported in the abstract were also reported in ClinicalTrials.gov.
Conclusions: Our data suggest that investigators may not be updating ClinicalTrials.gov trial information. Both conference abstracts and ClinicalTrials.gov provide unique primary outcome information not available in the other source. Systematic reviewers should obtain outcome information from both conference abstracts and registries. Study protocols with primary outcome information should be available from trial registries.
Objectives: To compare primary outcomes described in conference abstracts with those reported in ClinicalTrials.gov and assess the relationship between primary outcome reporting and industry sponsorship.
Methods: Eligible conference abstracts described RCT results, had a valid ClinicalTrials.gov registration number, and were presented at the 2007 to 2009 Association for Research in Vision and Ophthalmology meetings. Two reviewers independently extracted information from the abstract and ClinicalTrials.gov. Discrepancies were resolved by consensus.
Results: We identified 152 eligible abstracts; 40 explicitly reported at least one primary outcome in the abstract and primary outcome field at ClinicalTrials.gov. Of 80 primary outcomes reported in abstracts, 18% (14/80) agreed with one reported in ClinicalTrials.gov. Forty-nine percent (39/80) agreed partially, in that additional information (eg. related to time point) was present in ClinicalTrials.gov (17/39; 44%), the abstract (19/39; 49%), or both (3/39; 8%). Thirty-four percent (27/80) of primary outcomes reported in the abstract were not reported as such in ClinicalTrials.gov, although 6 were reported as secondary outcomes. Industry sponsorship was reported in 6/14 abstracts where a primary outcome was reported in the abstract but not in ClinicalTrials.gov, and in 14/26 abstracts where all primary outcomes reported in the abstract were also reported in ClinicalTrials.gov.
Conclusions: Our data suggest that investigators may not be updating ClinicalTrials.gov trial information. Both conference abstracts and ClinicalTrials.gov provide unique primary outcome information not available in the other source. Systematic reviewers should obtain outcome information from both conference abstracts and registries. Study protocols with primary outcome information should be available from trial registries.