STATISTICAL-METHODS FOR 2-SEQUENCE 3-PERIOD CROSS-OVER DESIGNS WITH INCOMPLETE DATA

In clinical trials, and in bioavailability and bioequivalence studies, one often encounters replicate cross-over designs such as a two-seque nce three-period cross-over design to assess treatment and carry-over effects of two formulations of a drug product. Because of the potentia l dropout (or for some administrative reason), however, the observed d ata set from a replicate cross-over design is incomplete or unbalanced so that standard statistical methods for a cross-over design may not apply directly. For inference on the treatment and carry-over effects, we propose a method based on differences of the observations that eli minates the random subject effects and thus does not require any distr ibutional condition on the random subject effects. When no datum is mi ssing, this method provides the same results as the ordinary least squ ares method. When there are missing data, the proposed method still pr ovides exact confidence intervals for the treatment and carry-over eff ects, as long as the dropout is independent of the measurement errors. We provide an example for illustration. (C) 1997 by John Wiley & Sons , Ltd.