Achieving the Benefits of Both an Internal Pilot and Interim Analysis in Small Samples

• Activity Number: 187
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 10:30 AM to 12:20 PM
• Sponsor: ENAR
• Author(s): John A. Kairalla*+ and Keith E. Muller and Christopher S. Coffey
• Companies: University of Florida and University of Florida and The University of Iowa
• Keywords: Adaptive designs ; Power ; Sample size re-estimation

Abstract

An internal pilot with interim analysis (IPIA) design combines interim power analysis (an internal pilot) with interim data analysis (a two stage group-sequential design). We describe methods for a practical IPIA design that works well even in small sample studies. The design provides the flexibility and efficiency of both the internal pilot and group sequential by controlling power and sample size simultaneously rather than either alone. New IPIA results for Gaussian linear models are valuable for both small and large studies. Enumerations demonstrate that guaranteeing control of type I error rate requires some form of correction. A simple bounding method approximately controls the type I error rate across all possible population variance values with little cost in power or expected sample size.