In this paper we discuss the Vital role that population (hierarchical) mode
lling can play within the drug development process. Specifically, populatio
n pharmacokinetic/pharmacodynamic models can provide reliable predictions o
f an individualized dose-exposure-response relationship. A predictive model
of this kind can be used to simulate and hence design clinical trials, fin
d initial dosage regimens satisfying an optimality criterion on the populat
ion distribution of responses, and individualized regimens satisfying such
a criterion conditional on individual features, such as sex, age, etc. Thro
ughout we emphasize prediction and advocate mechanistic as opposed to empir
ical modelling, and argue that the Bayesian approach is particularly natura
l in this setting.