On the assessment of dose proportionality: a comparison of two slope approaches.

The problem for assessment of dose proportionality (or linearity) is studied. Various methods for assessment of dose proportionality (or linearity) such as ANOVA type F-test have been proposed. Cheng et al. (2006) proposed an alternative approach based on the slopes of adjacent dose levels under a crossover design. They showed that when dose proportionality (or linearity) cannot be established, their proposed slope approach is useful for evaluation of the degree of departure from dose proportionality (or linearity). In this article, we propose the use of slopes between the dose level and the initial dose (baseline), which we refer to as the baseline slope approach. The two slope approaches are compared under a parallel group design by means of an ANOVA type F-test and other tests. Simulation studies show that the proposed method has a satisfactory small sample performance.

Duke Scholars

Author Shein-Chung Chow Biostatistics & Bioinformatics, Division of Biostatistics