Computationally Efficient Approach to Operational Prior Specification in Phase I Clinical Trials

The increased interest in the combination of drug agents and/or schedules requires more flexible model-based designs for Phase I clinical trials. The models often involve more design parameters, which have to be specified before starting the trial. Those are often chosen from statistical considerations, that is, need to be optimized in a simulation study. Consequently, an efficient and systematic way for parameter calibration will both improve the model performance and save computational time. The conventional "grid search" calibration approach requires large simulations, which are computationally costly. A novel "cyclic calibration" has been proposed to reduce the computation from multiplicative to additive. Furthermore, calibration processes should consider a wide range of scenarios of true toxicity probabilities to avoid bias. A method to reduce scenarios based on scenario complexity is suggested. This can reduce the computation by more than 500-fold, while maintaining operational characteristics similar to the grid search.