Bayesian optimal designs for pharmacokinetic models: sensitivity to uncertainty.

We studied the sensitivity of the number of unique design points and their placement, in Bayesian optimal designs for pharmacokinetic models, with respect to the magnitude of prior uncertainty. We used two and three-parameter pharmacokinetic models with fixed and mixed effects and two Bayesian optimal design criteria, namely ED and API, using different error weighting schemes. We found that by increasing the magnitude of the uncertainty, in most cases, additional design points appear, compared to the corresponding local design, and this happens gradually, forming bifurcation patterns. These bifurcation patterns were interpreted as high sensitivity of the design from the magnitude of the uncertainty.