Bayesian decision procedures for dose‐escalation based on evidence of undesirable events and therapeutic benefit

Abstract

In this paper, Bayesian decision procedures are developed for dose-escalation studies based on bivariate observations of undesirable events and signs of therapeutic benefit. The methods generalize earlier approaches taking into account only the undesirable outcomes. Logistic regression models are used to model the two responses, which are both assumed to take a binary form. A prior distribution for the unknown model parameters is suggested and an optional safety constraint can be included. Gain functions to be maximized are formulated in terms of accurate estimation of the limits of a 'therapeutic window' or optimal treatment of the next cohort of subjects, although the approach could be applied to achieve any of a wide variety of objectives. The designs introduced are illustrated through simulation and retrospective implementation to a completed dose-escalation study.

A Bayesian Approach for Dose-Escalation in a Phase I Clinical Trial Incorporating Pharmacodynamic Endpoints

“…In other settings, different pharmacodynamic endpoints might be used. Whitehead et al (2006b) of an E max model: conservative in the sense that higher doses will appear more dangerous under the quadratic model and will thus be avoided until sufficient evidence of their safety has been collected. The authors recognise that each clinical study is unique, and that the details of other trials may be different from that studied here.…”

Bayesian statistics in medicine: a 25 year review

“…Concomitant information in bioassay can be incorporated in a semi-parametric model [427]. Dose-finding designs accounted for two endpoints using utilities [428,429] and Bayesian designs for dose-escalation studies in healthy volunteers were compared [430]. Bayesian phase II designs were compared to Simon two-stage designs [431].…”

Bayesian decision procedures for binary and continuous bivariate dose‐escalation studies

“…Therefore, in this paper, we extend the Bayesian approach of Whitehead et al [15,16] so that the bivariate outcomes are binary DLEs and continuous DOs. In doing so, we introduce an allowance for the correlations between repeated observations of DOs on the same subject, although we have not been able to do the same for the binary DLEs.…”

“…Whitehead et al [15] considered dose-escalation studies in which two responses are observed on each subject. One of these was referred to as a dose limiting event (DLE), and the other as a DO.…”

“…Both desirable and undesirable responses were limited to binary outcomes in these papers, and the applications discussed were mainly to bone marrow transplantation. Most of these methods were intensively discussed in [15], where a Bayesian approach based on logistic regression models is presented.…”