Innovative Statistical Approaches for Clinical Trials

Elementary topics in statistical inference include methods for comparing the means from two groups. If we have more than two groups, or if we need to account for the possible effects of other factors, more advanced methods are needed. These methods are also used to summarize complex datasets and to reduce large numbers of data points to a smaller set of parameters that can be used to make inference or implement policy. Common techniques such as analysis of variance (ANOVA), analysis of covariance (ANCOVA), and multiple regression allow us to move beyond simple comparison of two means to developing mathematical models for expected responses based on the values of various factors. Methodology allows for continuous, ordinal or categorical outcomes, as well as varying types of input factors. In this module, we will introduce the concept of statistical modeling and describe the more common methods used.

• What is a statistical “model?”
• Data reduction
• Determine the difference in means by examining partitioned variances
• Simple linear regression and the relationship to ANOVA
• Multiple and logistic regression
• Ordinal logistic and non-linear regression
• Examples