Model-Based Recursive Partitioning for Stratified Medicine

This project aims at developing methods for inference on patient-specifictreatment effects in the context of stratified medicine.  The term “stratified medicine”refers
to methods “using a biomarker to match a patient to a cohort that has exhibited
a differential response to a treatment”according to the US Food and Drug
Administration report “Paving the Way for Personalized Medicine”published in 2013. 
Established statistical procedures based on stratified or interaction models are routinely used when both biomarkers and cohorts have already been identified by previous research.  This proposal targets statistical procedures tailored for the identification of such cohorts from a larger number of potentially useful biomarkers.  From a conceptual point of view, we distinguish  between three separate steps of
such an analysis: (1) the selection of one (or a few) biomarker(s) which is predictive
for a differential treatment effect, (2) the identification of patient subgroups (or cohorts) from these  selected biomarkers, and (3) the estimation of a subgroup-specific treatment  effect. Established stratified or interaction models for (3) are commonly refered  to as “primary subgroup analyses”in stratified, individualised, or personalised medicine.  With this project, we develop statistical methods for biomarker selection
and subgroup identification and improved estimation procedures for differential treatment effects in steps (1-3).

For more information and to participate please visit the Swiss National Science Foundation.