| Popis: |
Importance: Clinical trial registries are key for gaining an overview of ongoing research efforts and for deterring and identifying publication bias and selective outcome reporting. It is uncertain how reliable the information in trial registries is. Objective: To assess the reliability of information across registries for trials with multiple registrations. Evidence Review: We had access to 360 protocols of randomized controlled trials (RCTs) approved by research ethics committees in Switzerland, UK, Canada, or Germany in 2012. Clinical trial registries were searched between March 2019 and September 2019 for corresponding registrations of these RCTs. We identified 197 RCTs that were registered in more than one registry. We extracted in duplicate key trial characteristics that should be identical amongst all trial registries (i.e. sponsor, funding source, primary outcome, target sample size, trial status, and date of first patient enrolment, results available, main publication indexed). Agreement between the different trial registries for these key characteristics was analyzed descriptively (analyses conducted May 2020 until November 2020). Representatives from clinical trial registries were interviewed to discuss the study findings between February 2021 and March 2021. Findings: The included 197 RCTs were registered in two (n=151) or three (n=46) trial registries (i.e. 188 in ClinicalTrials.gov; 185 in EudraCT; 20 in ISRCTN; 47 in other registries). The agreement of key information across all registries was as follows: 90% (178/197; 95% confidence interval [CI], 85%-94%) for sponsor, 90% for funding source (18/20; 95% CI, 68-99%; funding not reported on ClinicalTrials.gov), 78% (154/197; 95% CI, 72-84%) for the primary outcome, 46% (90/197; 95% CI, 39-53%) for the trial status, 62% (122/194; 95% CI, 39-53%) for the target sample size, and 75% for the date of first patient enrolment (43/57; 95% CI, 62-86%; date of first patient enrolment not reported on EudraCT) when the comparison time window was broadened to 30 days. The agreement with respect to results availability on trial registry was 62% (122/197; 95% CI, 55-69%) and 46% (91/197; 95% CI, 39-53%) when assessing whether a publication of the main results was indexed. Different legal requirements were stated as main reason for inconsistencies by representatives of clinical trial registries. Conclusions and relevance: For a substantial proportion of RCTs registered, information about key trial characteristics was inconsistent across trial registries, raising concerns about the reliability. Further dialogue and harmonization across clinical trial registries is necessary to increase their usefulness. |
