Evaluation of Power and Type I Error on Different Statistical Methods Analyzing Hypoglycemia Data Using Bootstrap Simulation

• Activity Number: 423
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Honghua Jiang*+ and William Huster and Xiao Ni
• Companies: Eli Lilly and Company and Eli Lilly and Company and Eli Lilly and Company
• Keywords: hypoglycemia ; bootstrap

Abstract

Hypoglycemia has long been recognized as a major barrier to achieving normoglycemia with intensive diabetic therapy. It is one common safety concern for the diabetes patients. Therefore, the proper application of statistical methods for analyses of the hypoglycemia data is of importance. The Poisson model is commonly used to analyze count data, like the hypoglycemia event data. However, one characteristic of the Poisson distribution is that its mean and variance are identical. The negative binomial model is an alternative for over-dispersed count data in which the variance exceeds the mean. Non-parametric rank AN(C)OVA model is another way to analyze the hypoglycemia data. Sometimes, simple AN(C)OVA model is also used to analyze the hypoglycemia data. Bootstrap simulation studies are conducted to evaluate the power and type I error of these four statistical methods based on the data from a diabetes clinical trial.