A Bayesian Credible Subgroups Approach to Identifying Patient Subgroups with Positive Treatment Effects

Abstract

Summary Many new experimental treatments benefit only a subset of the population. Identifying the baseline covariate profiles of patients who benefit from such a treatment, rather than determining whether or not the treatment has a population-level effect, can substantially lessen the risk in undertaking a clinical trial and expose fewer patients to treatments that do not benefit them. The standard analyses for identifying patient subgroups that benefit from an experimental treatment either do not account for multiplicity, or focus on testing for the presence of treatment-covariate interactions rather than the resulting individualized treatment effects. We propose a Bayesian credible subgroups method to identify two bounding subgroups for the benefiting subgroup: one for which it is likely that all members simultaneously have a treatment effect exceeding a specified threshold, and another for which it is likely that no members do. We examine frequentist properties of the credible subgroups method via simulations and illustrate the approach using data from an Alzheimer's disease treatment trial. We conclude with a discussion of the advantages and limitations of this approach to identifying patients for whom the treatment is beneficial.

Subgroup identification via homogeneity pursuit for dense longitudinal/spatial data

“…Lipkovich et al provided a nice tutorial to describe, compare, and summarized the key features of these various methods. Recently, Schnell et al proposed a Bayesian credible subgroups method for simultaneous inference regarding who benefits from treatment in the context of a hierarchical linear model; Schnell et al developed subgroup analysis methods to handle cases in which more than two treatments are being compared with respect to multiple endpoints; Schnell et al studied the details required to follow the credible subgroups approach in more realistic settings by considering nonlinear and semiparametric regression models. Instead of dividing the samples according to cross‐sectional subject characteristics, we will propose a subgroup identification method based on homogeneity pursuit for longitudinal studies.…”

A nonparametric Bayesian basket trial design

“…Some recent frequentist approaches use Bayesian methods to determine the adaptive enrichment to a subpopulation that is most likely to benefit from a treatment (Simon & Simon, 2018). Schnell, Tang, Offen, and Carlin (2016) and Schnell, Tang, Mller, and Carlin (2017) propose a principled Bayesian approach by defining a notion of posterior credible intervals for the estimated subpopulation. One of the challenges of such approaches is the control of (frequentist) operating characteristics.…”

“…Quantifying errors and uncertainties for the more general problem of inference for a benefiting subpopulation without predefined candidates is more challenging. Schnell, Tang, Offen, and Carlin (2016) and Schnell, Tang, Mller, and Carlin (2017) propose a principled Bayesian approach by defining a notion of posterior credible intervals for the estimated subpopulation.…”

A Bayesian subgroup analysis using collections of ANOVA models

“…Later, Hodges et al (2007) and Jones et al (2011) extended the approaches by considering more general random effects and provided more flexible shrinkage estimates for interaction coefficients. Schnell et al (2016) propose an approach that could be characterized as constructing a credible interval for the desired subgroups. The approach defines an inclusive and an exclusive subset of the covariate space, with posterior probability greater than or equal 1 − α that the inclusive set contains all covariate values with differential treatment effect and similarly for the exclusive set to contain only covariates with differential treatment effect.…”