Analyzing Longitudinal Clinical Trial Data

Missing data in longitudinal clinical trials has justifiably been the target of considerable research. However, missing data is just one of the many considerations in the analysis of longitudinal data, and focus on the data we don't have should not distract from focus on the data we do have. The statistical theory relevant to analyses of longitu

Background and Setting. Introduction. Objectives and estimands-determining what to estimate. Study design-collecting the intended data. Example data. Mixed effects models review.

Modeling the observed data. Choice of dependent variable and statistical test. modeling covariance (correlation). Modeling means over time. Accounting for covariates. Categorical data. Model checking and verification.

Methods for dealing with missing Data. Overview of missing data. Simple and ad hoc Approaches for dealing with missing data. Direct maximum likelihood. Multiple imputation. Inverse probability. Methods for incomplete categorical data weighted generalized estimated equations. Doubly robust methods. MNAR methods. Methods for incomplete categorical data.

A comprehensive approach to study development and analyses. Developing statistical analysis plans. Example analyses of clinical trial data.