Comparing Four Methods for Estimating Tree-Based Treatment Regimes

Abstract

When multiple treatment alternatives are available for a certain psychological or medical problem, an important challenge is to find an optimal treatment regime, which specifies for each patient the most effective treatment alternative given his or her pattern of pretreatment characteristics. The focus of this paper is on tree-based treatment regimes, which link an optimal treatment alternative to each leaf of a tree; as such they provide an insightful representation of the decision structure underlying the regime. This paper compares the absolute and relative performance of four methods for estimating regimes of that sort (viz., Interaction Trees, Model-based Recursive Partitioning, an approach developed by Zhang et al. and Qualitative Interaction Trees) in an extensive simulation study. The evaluation criteria were, on the one hand, the expected outcome if the entire population would be subjected to the treatment regime resulting from each method under study and the proportion of clients assigned to the truly best treatment alternative, and, on the other hand, the Type I and Type II error probabilities of each method. The method of Zhang et al. was superior regarding the first two outcome measures and the Type II error probabilities, but performed worst in some conditions of the simulation study regarding Type I error probabilities.

Detecting treatment-subgroup interactions in clustered data with generalized linear mixed-effects model trees

“…These methods have different objectives and there is not yet an agreed-upon single best method. In a simulation study, Sies and Van Mechelen (2016) found the method of Zhang et al (2012a) to perform best, followed by MOB. However, the method of Zhang et al performed worst under some conditions of the simulation study in terms of the type I error rate.…”

A comparison of subgroup identification methods in clinical drug development: Simulation study and regulatory considerations

“…Therefore, our comparative study focuses on a selection criterion appropriate from a regulatory point of view. Other comparisons are based on a single dataset, evaluate the estimation of tree‐based treatment regimes or focus on identifying covariates affecting treatment effects rather than identifying subgroups.…”

“…Although there are numerous methods for this purpose available, comparisons of methods applicable to similar settings are lacking, with the notable exceptions of articles by Doove et al, Alemayehu et al, and Sies and Mechelen . Boulesteix et al point out that more neutral comparison studies, which evaluate the behavior of existing methods, are needed.…”

Qualitative treatment-subgroup interactions in the antidepressant treatment of major depression: Application of QUINT to individual participant data from seven placebo-controlled randomized controlled trials

“…' It is mainly the latter type of interaction that matters in personalized medicine [38]. Only recently, however, important methodological developments have been made for a statistical method that focuses on https://doi.org/10.1016/j.pmip.2019.100054 qualitative interaction [25,33]. Qualitative interaction trees (QUINT) is one such method recently developed by Dusseldorp et al [9,10].…”