This article extends the correlation methodology developed by Chinchilli et al. (2005) for the 22 crossover design to more complex crossover designs for clinical trials. We describe how the methodology can be adapted to a general type of two-treatment crossover design which includes either at least two sequences or at least two treatment periods or both. We then derive the asymptotic theory for the corresponding correlation statistics, investigate the statistical accuracy of the estimators via bootstrap analyses, and demonstrate their use with two real data examples.
All Science Journal Classification (ASJC) codes
- Statistics and Probability