Cross-over trials

If two treatments (A and B) are to be compared using a cross-over trial design then each patient receives both treatments. Randomisation is used to determine the order in which the patient receives each treatment, i.e. treatment A followed by B, or treatment B followed by A.

One advantage of using a cross-over design is that it reduces the between patient variability, because the comparison of treatment A versus B is made on the same patient. Thus, sample sizes are smaller than for a parallel group trial design. Cross-over trials cannot be used in the situation where patients can only receive one treatment (such as surgery), and are appropriate only for chronic conditions. An analysis of a cross-over trial should investigate if patient outcome is influenced by the order in which treatments are given and if the effects of the first treatment carry over into the period of the second treatment.

Resources:

• Jones B, Kenward G. Design and Analysis of Cross-over Trials. Chapman & Hall, London, 2003.
• Senn S. Cross-over Trials in Clinical Research. 2nd edition. Wiley, Chichester, 2002.

Examples of published studies

• Hong B, Hong JH, Nam KY, Ahn TY. A double-blind crossover study evaluating the efficacy of Korean red ginseng in patients with erectile dysfunction: A preliminary report. Journal of Urology 2002; 168(5): 2070-2073.