The analysis of dose-response relationships based on PK PD models typically requires the use of nonlinear regression. Although linear models, e.g., quadratic, can be used to describe curvi-linear dose-response relationships, there are usually other nonlinear components of the model, e.g., using exponential functions to describe the time course of onset and loss of effect.
Sometimes doses are coded in sequential categories that lose information about the size of the dose, e.g., doses of 0, 50, 100, 200, 400 might be coded as 1, 2, 3, and 4. This recoding should be avoided when trying to understand dose-response relationships in pharmacological terms.
Was this article helpful?