Registration of phase 3 crossover trials on ClinicalTrials.gov

Background: In a randomized crossover trial, each participant is randomized to a sequence of treatments and treatment effect is estimated based on within-individual difference because each participant serves as his/her own control. This feature makes the design and reporting of randomized crossover trials different from that of parallel trials. Our objective was to characterize phase 3 crossover trials with results reported on ClinicalTrials.gov and identify issues and best practices for reporting. Methods: We searched ClinicalTrials.gov for phase 3 randomized crossover trials that provided results, registered at least one primary outcome, and included at least one link to a results publication in the record by August 6, 2019. Two reviewers independently assessed the eligibility and extracted information from each record into an electronic form developed and maintained in the Systematic Review Data Repository. Results: Of the 124 crossover trials analyzed, two thirds were a simple "Intervention A then B"or "Intervention B then A"(AB|BA) design. Most trials (78%, 97/124) provided enough information to understand the participant flow throughout the trial. Baseline characteristics were most often reported for all participants as a single group (52%, 65/124). Primary outcomes and adverse events were most commonly reported "per intervention"(85%, 105/124, and 80%, 99/124, respectively). Conclusions: The registration and reporting of randomized crossover trials must account for the paired nature of the design. Our observations and recommendations informed the development of guidelines for good reporting practices in the registration and reporting of randomized crossover trials.