Abstract:

For a longitudinal clinical trial, an efficacy endpoint can be defined on a single time point at the end of trial; or an average of data over all post baseline time points, like in an allergy study; or partial post baseline time points, like in an asthma study. When using multiple time points to define an efficacy endpoint, it is usually assumed that the treatment effect is close to the expected trial value over the interval used. Here, we will define a formula to diagnose how close the expected and actual treatment effect must be in order to realize an equal or better power when comparing to a single time endpoint. Of course, it is well known that an endpoint defined using multiple time points is less sensitive to missing values and we will extend our formula to account for a given fraction of missing values.
