Practical Issues in Sample Size Re-estimation in Clinical Trials - SOLD OUT

• Activity Number: 347
• Type: Luncheons
• Date/Time: Wednesday, August 11, 2004 : 12:30 PM to 1:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Jeff Maca*+
• Companies: Novartis Pharmaceuticals
• Keywords: sample size re-estimation ; clinical trials

Abstract

Planning and designing clinical trials depends in many ways upon various types of information. When many trials are undertaken, much of this information will not be available at the desired level of precision, or possibly there is no directly relevant estimate available. There will almost certainly be some differences between the parameter in the current trial and the source of the estimate. Indeed, it may frequently be the case that the final data from a clinical trial provides the first opportunity to confirm or contradict the sample size assumptions. One may see at that point that these incorrect assumptions led to a trial which had considerably different power than was anticipated. The potential implications are that a successful trial outcome may be jeopardized because power was insufficient, or substantial resources may have been wasted because the trial was needlessly large. The possibility of updating the sample size as the trial progresses if the observed data seems to disagree with the pre-trial assumptions would be useful in this case. The practical issues pertaining to designing and running such a trial will be discussed.