| Abstrakt: |
In a basic AB/BA crossover trial, patients are randomly assigned to receive either treatment A in the first period followed by treatment B in the second period, or treatment B in the first period followed by treatment A in the second period. Treatment periods are separated by a suitable washout period that is long enough for the treatment effect from the first period to have no residual effect to the second treatment period. The AB/BA crossover trial has many advantages and, when used appropriately, provides an efficient means for comparing 2 treatments. If applied and conducted inappropriately, the AB/BA crossover design has fatal flaws. The objective of this article is to introduce the AB/BA crossover trial and to present some of the issues surrounding crossover trials. The crossover trial, when properly implemented and analyzed, remains a valid and efficient design alternative to the parallel group design and should not be abandoned as has been suggested by the FDA. Crossover studies have an advantage over parallel studies, for example, elimination of between-subject variability, more statistical power because of paired comparisons, and reduced sample sizes to achieve a specified power. Although a test for carryover would be desirable, at present there is no widely recognized method for doing this, and some experts feel that no method can be reliable. [ABSTRACT FROM AUTHOR] |
