Innovative Statistical Approaches for Clinical Trials

Survival Analysis
Survival analysis refers to the statistical methods used to summarize and analyze time to event data. While the term “survival analysis” implies analysis of mortality data, the methods are actually applicable to any situation in which times to a particular event are measured. An important consideration in the analysis of these data is the impact of censoring - when a subject is lost to follow up or the trial is completed prior to the subject experiencing the event of interest. In this module, we will also examine descriptive statistics and graphical displays associated with the analysis of survival data, as well as techniques for comparing treatments and assessing the effect of covariates.

• The nature of time to event data and censoring
• Summary statistics for survival data and estimating survival times
• Non-parametric tests for comparing two or more groups
• Cox proportional hazards regression
• Planning and Conducting a “Survival” Trial

Non-inferiority
There has been a steady increase in the use of active control trials to investigate the efficacy of a new investigational regimen. In this module, we will focus on those active control trials where the primary objective is to demonstrate that a new regimen is not inferior to an active control. We will review fundamental requirements trialists need to consider before they can be assured that such trials can accomplish the ultimate research objectives. In addition, we will discuss in detail issues essential to determination of the non-inferiority margin. Furthermore, we will examine the phenomenon known as the biocreep and the effect of a sloppy trial conduct on the study’s ability to provide correct conclusions.

• Objectives of active control trials
• Differences between non-inferiority trials and superiority trials
• Design considerations of an active control study
• Choice of the non-inferiority margin
• Analysis of an active control study: Confidence interval approach
• What is biocreep? Is it real?
• Testing superiority after concluding non-inferiority in a trial
• Examples

Exercise