35 – Clinical Trial Design and Randomization
Is Your Study Balanced, as You Had Planned?
Michael Collins
ICON Clinical Research
Eva Ruth Miller
ICON Clinical Research
Randomization coupled with maintaining the blind form the gold standard for minimizing the bias within the clinical trial. The scientific integrity of the trial greatly depends on adherence to the randomization schedule. To be sure that no biases are introduced, randomization procedures must be precise. Clinicians and statisticians who conduct studies and report results depend upon timely, accurate and clean data. While the trial is being conducted, trial statisticians may encounter any number of data discrepancies. These discrepancies can greatly impact the final analyses. Logistical and operational considerations in the actual practice of randomization and study conduct can provide opportunities to correct common errors that could otherwise compromise study validity. Over the course of nine months a collection of misrandomization cases and resolutions were documented. Case studies of “misrandomization events� showed a comprehensive amount of data discrepancies that arose during the actual practice of randomization. These misrandomization events were characterized as: (1) Misclassification of subject stratifying information, (2) Screen failed subjects continuing to the randomization phase, (3) Incorrect treatments administered to randomized subjects, and (4) Multiple randomizations of the same subject. Involving the trial statistician in management of misrandomization events as they unfolded, while maintaining the blind, offered the trial statistician the opportunity to work with the clinical trial study team to minimize the likelihood of recurrence of these same errors within the study.
