The Accuracy of Clinical Trial Inferences: The Saw Palmetto Experience

• Activity Number: 360
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Jeannette Y. Lee*+ and Page Moore and Shelly Lensing
• Companies: University of Arkansas for Medical Sciences and University of Arkansas for Medical Sciences and University of Arkansas for Medical Sciences
• Keywords: clinical trials ; statistical inference

Abstract

Planning a clinical trial requires specifying the significance level and power to control type I and type II errors. After a trial is completed, the question turns to whether or not the inference from the trial is correct. Determining the probability that the inference from a clinical trial is correct has been described using "crucial" type I or II error rates. Results of a randomized, placebo-controlled clinical trial of saw palmetto for benign prostatic hyperplasia showed that the product was no more effective than placebo. Since prior studies suggested that the product was effective, it was unclear if the inference from this study was correct. This presentation will demonstrate the false negative rate for this study over a range of underlying probabilities of a negative inference, and its implications for the interpretation of an ongoing clinical trial of saw palmetto.