Crossover Studies with Continuous Variables: Power Analysis
Background: The crossover design is a sensitive means of determining the efficacy of new drugs because it eliminates between subject-variability. However, when the response in the first period carries on into the second (carryover effects) or when time factors can not be kept constant in a lengthy crossover (time effects), the statistical power of testing may be jeopardized. We recently demonstrated that the crossover design with binary variables is a powerful method in spite of such factors as carryover effects. Power analysis of crossover trials with continuous variables have not been explicitly studied.
Objective: Using the Grizzle model for the assessment of treatment effect, carryover effect and time effect, we drew power curves of hypothesized crossover studies with different levels of correlation between drug reponse.
Results: We demonstrate that the sensitivity of testing is largely dependent on the levels of correlation between drug response. Whenever the correlation coefficient is >0, we soon will have better sensitivity to test treatment effect than carryover effect or time effect of similar size. Whenever levels of correlation are not strong positive or negative the statistical power to demonstrate similarly-sized treatment and carryover effect, or treatment and time effect is approximately 80%, which is an acceptable level for reliable testing.
Conclusions: The crossover design is a powerful method for assessing positively correlated treatment comparisons, despite the risk of carryover and time effects.
KeywordsChronic Obstructive Pulmonary Disease Time Effect Crossover Study Carryover Effect Crossover Design
Unable to display preview. Download preview PDF.
- 1.Niemeyer MG, Zwinderman AH, Cleophas TJ, De Vogel EM. Crossover studies are a better format for comparing equivalent treatments than parallel-group studies. In: Kuhlmann J, Mrozikiewicz A, eds, What should a clinical pharmacologist know to start a clinical trial (phase I and II). Munich, Germany, Zuckschwerdt Verlag, 1998, pp 40–48.Google Scholar
- 2.Scheffé H. Mixed models. In: Scheffé H, ed, The analysis of variance. New York, Wiley & Sons, 1959, pp 261–291.Google Scholar
- 9.Cleophas TJ, Van Lier HH. Clinical trials with binary responses: power analyses. J Clin Pharmacol 1996; 36: 198–204.Google Scholar
- 10.Nies AS, Spielberg SP. Individualization of drug therapy. In: Hardman JL et al., eds, Goodman and Gilman’s Pharmacological Basis of Therapeutics. New York: McGraw-Hill, 1996, pp 43–63.Google Scholar
- 11.Grizzle JE. The two-period change-over design and its use in clinical trials. Bbiometrics 1965; 22: 469–480.Google Scholar
- 12.SPSS 8 for Windows 95 and 98, SPSS Benelux, Gorinchem, Netherlands.Google Scholar
- 13.Hays WL. Statistics. Fort Worth, TX, Holt, Rinehart and Winston, Inc, 4th edition, 1991.Google Scholar