Manual Validation
Ninety-four percent of screened potential cross-registrations were confirmed as genuine cross-registrations of the same trial. A non-random sample of 242 potential cross-registrations was manually screened. Manual confirmation used multipl…
1 sources - 5 claims
Ninety-four percent of screened potential cross-registrations were confirmed as genuine cross-registrations of the same trial. A non-random sample of 242 potential cross-registrations was manually screened. Manual confirmation used multiple trial characteristics, including title or sponsor, outcomes, design, intervention, comparator, enrolment, and population. Exact matches across all characteristics were not required for confirmation. Publication full-text TRN mentions had higher false-positive risk than some other identification sources.