The continual reassessment method: comparison of Bayesian stopping rules for dose-ranging studies

The continual reassessment method (CRM) provides a Bayesian estimation of t he maximum tolerated dose (MTD) in phase I clinical trials and is also used to estimate the minimal efficacy dose (MED) in phase II clinical trials. I n this paper we propose Bayesian stopping rules for the CRM, based on eithe r posterior or predictive probability distributions that can be applied seq uentially during the trial. These rules aim at early detection of either th e mis-choice of dose range or a prefixed gain in the point estimate or accu racy of estimated probability of response associated with the MTD (or MED). They were compared through a simulation study under six situations that co uld represent the underlying unknown dose-response (either toxicity or fail ure) relationship, in terms of sample size, probability of correct selectio n and bias of the response probability associated to the MTD (or MED). Our results show that the stopping rules act correctly, with early stopping by using the two first rules based on the posterior distribution when the actu al underlying dose-response relationship is far from that initially suppose d, while the rules based on predictive gain functions provide a discontinua tion of inclusions whatever the actual dose-response curve after 20 patient s on average, that is, depending mostly on the accumulated data. The stoppi ng rules were then applied to a data set from a dose-ranging phase II clini cal trial aiming at estimating the MED dose of midazolam in the sedation of infants during cardiac catheterization. All these findings suggest the ear ly use of the two first rules to detect a mis-choice of dose range, while t hey confirm the requirement of including at least 20 patients at the same d ose to reach an accurate estimate of MTD (MED). A two-stage design is under study. Copyright (C) 2001 John Wiley & Sons, Ltd.