ABSTRACT
Longitudinal profiles of pharmacodynamic (PD) biomarkers are studied for a variety of objectives. PD biomarkers usually provide further understanding of the mechanistic pathway(s) between the drug and its impact on the disease. The objectives for the study of these biomarkers may yield early proof of concept, provide an early or alternate measurement of clinical response or even help clarify the safety profile. As an example, one could imagine the movement of a set of biomarkers indicating hepatic dysfunction in a subject. Often biomarkers measured in the clinical trial are continuous, and the measurements follow an approximately normal distribution. The log transformation and other boxcox transformation are available to apply to the data as needed. The linear mixed-effects model with repeated measures is used to examine the data, plot the model derived mean and confidence interval, and to make valid statistical inference on treatment effect over time or dose by time interaction.
