Dose-response models with covariates

Two approaches are presented to incorporate covariate effects into dose-response models. The first approach postulates a parametric form for the dose-response function with the resulting parameters then assumed to depend on the covariates. The second approach is semi-parametric: it assumes that the logit ratio (the difference between the logit of success rate at the covariate level and that of the baseline level) can be expressed as a regression model in the covariate effects and the dose level. We present results of a simulation study which compares the efficiency of the two approaches.