Optimal designs for finding the dose that maximizes a Clinical Utility Index
(English)Manuscript (preprint) (Other academic)
The importance of dose finding studies in the clinical development process cannot be overstated. Phase I dose finding studies usually focus on finding a safe dose while in later phases the focus is on finding an effective dose. The primary objectives in these different phases are often to estimate the maximum tolerable dose and the minimum effective dose respectively. The ultimate goal of any dose finding study is however to estimate the best dose for each patient. For the obvious reason this is not possible but in this paper we build a framework for designing dose finding studies aimed at estimating the single best dose for a population of patients. A dose that is both safe and effective. We use a utility function to model the patient net benefit from different doses of the drug taking both effects and side-effects into account. We then derive some locally c-optimal designs that minimize the asymptotic variance of the estimated target dose, the dose that maximizes the utility function. We use simulation studies to verify that the designs are appropriate and good even for small sample sizes. Efficacy calculations are carried out to study the impact of misspecification of the model parameters. We concluded that the proposed designs are good when the true parameters are equal to or larger than expected.
Emax model, Clinical Utility Index (CUI), c-optimal design, dose-response studies
Probability Theory and Statistics
Research subject Statistics
IdentifiersURN: urn:nbn:se:su:diva-102868OAI: oai:DiVA.org:su-102868DiVA: diva2:713848