|
Abstract:
|
The continual reassessment method (CRM) is an adaptive model-based design used to estimate the maximum tolerated dose in dose finding clinical trials. A way to evaluate the sensitivity of a given CRM model including the functional form of the dose-toxicity curve, the prior distribution on the model parameter, and the initial guesses of toxicity probability at each dose is using indifference intervals. However, there are infinitely many ways to specify the initial guesses. In practice, the model is specified by trial and error through extensive simulations. We propose a systematic approach to select the initial guesses with respect to a specified indifference interval. The proposed method is applied to re-calibrate two trials. The models selected by our semi-automatic approach are competitive to the models used in the trials, which were obtained through a time consuming process.
|
