Clinical Trial Data Sharing: A Cross-Sectional Study of Outcomes Associated with Two NIH Models

BackgroundThe impact and value of clinical trial data sharing, including the number and quality of publications that result from shared data - "shared data publications" - may differ depending on the data sharing model used. MethodsWe characterized the outcomes associated with two data sharing models previously used by Institutes of the U.S. National Institutes of Health (NIH): NHLBIs centralized model, which uses a repository to manage data sharing requests, and NCIs decentralized model, which entrusted research groups to independently manage data sharing requests. We identified trials completed in 2010 that met NIH data sharing criteria and matched studies sponsored by each Institute based on cost or size, determining whether trial data were shared and the frequency of shared data publications. ResultsWe identified 14 NHLBI-funded trials and 48 NCI-funded trials that met NIH data sharing criteria. We matched 14 NCI-funded trials to the 14 NHLBI-funded trials; among these, 4 NHLBI-sponsored trials (29%) and 2 NCI-sponsored trials (14%) shared data. From the 2 NCI-sponsored trials sharing data, we identified 2 shared data publications, one per trial, both of which were meta-analyses. From the 4 NHLBI-sponsored trials sharing data, we identified 7 shared data publications, all using data from 1 trial, 5 of which were pooled analyses and 2 reported secondary outcomes. ConclusionWhen characterizing the outcomes associated with two NIH data sharing models, both the NHLBI and the NCI models resulted in only 21% of trials sharing data and few shared data publications. There are opportunities to optimize clinical trial data sharing efforts both to enhance clinical trial data sharing and increase the number of shared data publications.