ABSTRACT
Precision medicine, in the sense of tailoring the choice of medical treatment to patients’ pretreatment characteristics, is nowadays gaining a lot of attention. Preferably, this tailoring should be realized in an evidence-based way, with key evidence in this regard pertaining to subgroups of patients that respond differentially to treatment (i.e., to subgroups involved in treatment–subgroup interactions). Often a-priori hypotheses on subgroups involved in treatment–subgroup interactions are lacking or are incomplete at best. Therefore, methods are needed that can induce such subgroups from empirical data on treatment effectiveness in a post hoc manner. Recently, quite a few such methods have been developed. So far, however, there is little empirical experience in their usage. This may be problematic for medical statisticians and statistically minded medical researchers, as many (nontrivial) choices have to be made during the data-analytic process. The main purpose of this paper is to discuss the major concepts and considerations when using these methods. This discussion will be based on a systematic, conceptual, and technical analysis of the type of research questions at play, and of the type of data that the methods can handle along with the available software, and a review of available empirical evidence. We will illustrate all this with the analysis of a dataset comparing several anti-depressant treatments.
Acknowledgments
Data used in the preparation of this manuscript were obtained and analyzed from the controlled access datasets distributed from the NIMH-supported National Database for Clinical Trials (NDCT). NDCT is a collaborative informatics system created by the National Institute of Mental Health to provide a national resource to support and accelerate discovery related to clinical trial research in mental health. Dataset identifiers: Study ID: N01 MH090003-02, clinical trial ID: NCT00590863. This manuscript reflects the views of the authors and may not reflect the opinions or views of the NIMH or of the Submitters submitting original data to NDCT.
Supplementary material
Supplemental data for this article can be accessed here.
Notes
1. Generally, methods that can handle observational data are based on inversely weighting each observation by the probability of receiving the treatment that it actually received using a propensity model (which models the probability of receiving treatment , given a covariate pattern ). Often, whether or not these methods are consistent depends on whether or not the propensity model is correctly specified.