Quality in Clinical Research: An Observational Study of Randomisation Techniques in Urological and General Surgical Studies

Abstract

ABSTRACTObjectivesTo quantitatively test the quality of randomised controlled trials reported to international scientific meetings through a critical analysis of randomisation outcomes.Design and Main Outcome MeasuresAll randomised controlled trials presented at international surgical and urological congresses using simple randomisation were identified. Primary analysis of randomisation technique was performed by comparing the observed and expected numbers of trials with equal numbers of participants in each arm. Sensitivity analyses compared study design, type of study and presence of external sponsorship. All abstracts were assessed according to the CONSORT for reporting randomised trials in journal and conference abstracts checklist.Results345 studies met the inclusion criteria. 148 studies reported simple randomisation to allocate 26,510 patients. Randomisation technique could not be identified in 104 studies. Primary and all secondary analyses demonstrated a probability of p<0.0001 that simple randomisation was used for participant allocation in all studies. Mean consort score was 9.4ConclusionsIt is extremely unlikely that simple randomisation was performed as reported in a significant proportion of the 148 RCTs in this study. These results raise concerning questions of the veracity and reliability of current medical research. There needs to be a greater awareness of the potential for methodological inaccuracy and error.