The Two Period Cross Over Trial

A comparison of the effect of paroxetine and amitriptyline on the tyramine pressor response test.

“…Tyramine-dose/pressor response curves were plotted on a semi-log scale and T30s (the doses of tyramine that produced a 30 mm Hg rise in systolic blood pressure) were calculated. The tyramine dose-ratios were derived as follows: T30 after treatment/i30 before treatment Analysis of variance was performed on log tyramine dose by the method of Barker et al (1982) for the repeated dose study, and using the statistical package GLIM to fit volunteer and order effects on the single dose study. Wilcoxon test was used for alertness and salivary volume.…”

“…T30 after treatment/i30 before treatment Analysis of variance was performed on log tyramine dose by the method of Barker et al (1982) for the repeated dose study, and using the statistical package GLIM to fit volunteer and order effects on the single dose study. Wilcoxon test was used for alertness and salivary volume.…”

Bayesian Methods in Practice: Experiences in the Pharmaceutical Industry

“…It is against this background that the Biometrics and Epidemiology Methodology Advisory Committee (BEMAC) ofthe Food and Drug Administration (FDA) has recommended the use of completely randomized designs rather than crossover designs, except when the absence of a carryover effect can be unequivocally demonstrated, either a priori or on the basis of the experimental data themselves. This view has recently been challenged by a group of British statisticians-see, for example, Poloniecki and Daniel (1981), Huitson et al (1982), Barker et al (1982) and Poloniecki and Pearce (1983)-who argue that crossover designs can still be valuable in certain cases where a carryover effect is present but is small relative to the treatment effect; see Grieve (1986) for comments on these arguments.…”

A Three-Step Procedure for Assessing Bioequivalence in the General Mixed Model Framework

“…The 90 per cent confidence interval (from 0.365 to 0.800) for p' indicates that H o l set up for p' is not rejected at the 0.05 level of significance and individual bioequivalence is not achieved ( Table 11). Scheffk's statistic in (12) with an equivalence limit of p1 = 0.50 produces an identical conclusion ( P = 0.187).…”