Discrepancies between ClinicalTrials.gov recruitment status and actual trial status: a cross-sectional analysis

OBJECTIVES: To determine the accuracy of the recruitment status listed on ClinicalTrials.gov as compared with the actual trial status. DESIGN: Cross-sectional analysis. SETTING: Random sample of interventional phase 2–4 clinical trials registered between 2010 and 2012 on ClinicalTrials.gov. PRIMARY OUTCOME MEASURE: For each trial which was listed within ClinicalTrials.gov as ongoing, two investigators performed a comprehensive literature search for evidence that the trial had actually been completed. For each trial listed as completed or terminated early by ClinicalTrials.gov, we compared the date that the trial was actually concluded with the date the registry was updated to reflect the study’s conclusion status. RESULTS: Among the 405 included trials, 92 had a registry status indicating that study activity was either ongoing or the recruitment status was unknown. Of these, published results were available for 34 (37%). Among the 313 concluded trials, the median delay between study completion and a registry update reflecting that the study had ended was 141 days (IQR 48–419), with delays of over 1 year present for 29%. In total, 125 trials (31%) either had a listed recruitment status which was incorrect or had a delay of more than 1 year between the time the study was concluded and the time the registry recruitment status was updated. CONCLUSIONS: At present, registry recruitment status information in ClinicalTrials.gov is often outdated or wrong. This inaccuracy has implications for the ability of researchers to identify completed trials and accurately characterise all available medical knowledge on a given subject.