Methods of Adjusting Sample Size for Noncompliance in Studies with Intent-to-treat Analyses of Survival Endpoints

• Activity Number: 487
• Type: Contributed
• Date/Time: Thursday, August 11, 2005 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Patricia L. Stephenson*+ and Robert J. Gray
• Companies: Rho, Inc. and Dana-Farber Cancer Institute, Harvard School of Public Health
• Keywords: noncompliance ; sample size ; intent-to-treat analyses

Abstract

Sample size determination is a crucial step in the design of clinical trials. Inadequate sample size can lead to serious consequences, including misinterpretation of study results, wasted money and resources, and unethical treatment of study subjects. In addition, sample size calculations often aid investigators in determining the feasibility of the current study. Due to the long-term followup generally needed in clinical trials, many patients will be noncompliant (will not adhere to their assigned treatment regimen). Noncompliance often leads to a loss of power for the standard intent-to-treat analysis, which in turn results in an increase in the required sample size. In this paper, we investigate two simple methods of adjusting sample size for noncompliance when the analysis is by intent-to-treat. The methods are appealing becuase they are used easily with existing approaches to sample size calculation and do not require extensive information about the pattern of noncompliance in the trial at hand. The performance of these methods of adjustment is examined under various noncompliance and clinical trial settings. Guidelines are provided for their appropriate use.