Practical Issues in Planning Cross-over Trials in Drug Development

Some practical issues in planning cross-over trials will be considered. A general approach to sample size determination for any arbitrary cross-over will be described. This also permits sample size determination when it is intended to recover inter-block information for incomplete block cross-overs. Consideration will also be given to planning to mimimize the effect of carry-over. Some aspects of period effects will also be covered and implications for use of baselines. Some "optimal" designs will be critically assessed. Finally, a neglected use of cross-overs, for resolving components of variation, will be covered. It will be concluded that cross-over designs could make an important contribution to the field of pharmacogenomics and that current research into the field of cross-over design has a rather distorted emphasis.