Accounting for cross-registration in monitoring transparency in clinical trials: a cross-sectional study of trials at German university medical centres.

Abstract

To identify cross-registrations (ie, registrations of the same trial in different registries) among clinical trials affiliated with German University Medical Centres (UMCs) and to describe how these cross-registrations are linked. To determine the consistency of key information across registrations of the same trial. Observational cross-sectional study. Registry-based study of clinical trials affiliated with German UMCs. 2895 interventional trials affiliated with German UMCs, registered on ClinicalTrials.gov or the German Clinical Trials Register (DRKS) and completed between 2009 and 2017. Number of potential cross-registrations involving the EU Clinical Trials Register (EUCTR) and the type of linkage between these cross-registrations. In a manually validated subset, consistency across registry entries of the same trial was assessed for prospective registration, recruitment status, completion date and availability of summary results. We identified 625 potential EUCTR cross-registrations. The majority of potential cross-registrations (n=470; 75%) were linked through a trial registration number (TRN) in at least one of the registries. Only 109 (17%) were linked through a TRN in both registries. Title matching and TRNs in trial results publications led to an additional 155 (25%) potential cross-registrations. In a sample of 228 manually confirmed EUCTR cross-registrations, 85% (95% CI 80% to 89%; n=189) had the same overall recruitment status while 50% (95% CI 43% to 57%; n=92) had the same completion date across registrations of the same trial. EUCTR was the dominant route for summary results availability and included structured tabular results as well as other formats. Among trials affiliated with German UMCs, cross-registrations are common, yet often inadequately linked and present discrepant information. In Germany, where registration in multiple registries is common, cross-registrations can meaningfully influence assessments of trial registration and results reporting and therefore warrant explicit consideration in transparency monitoring.