Causal Interaction Trees: Tree-Based Subgroup Identification for Observational Data
This method addresses subgroup identification for treatment effects in observational studies, which is important for personalized medicine, but it appears incremental as it builds on existing tree-based algorithms.
The authors tackled the problem of identifying subgroups with enhanced treatment effects from observational data by proposing Causal Interaction Trees, which extend Classification and Regression Trees with splitting criteria to maximize treatment effect heterogeneity. They demonstrated performance through simulations and an application to right heart catheterization in critically ill patients, though no concrete numbers were provided in the abstract.
We propose Causal Interaction Trees for identifying subgroups of participants that have enhanced treatment effects using observational data. We extend the Classification and Regression Tree algorithm by using splitting criteria that focus on maximizing between-group treatment effect heterogeneity based on subgroup-specific treatment effect estimators to dictate decision-making in the algorithm. We derive properties of three subgroup-specific treatment effect estimators that account for the observational nature of the data -- inverse probability weighting, g-formula and doubly robust estimators. We study the performance of the proposed algorithms using simulations and implement the algorithms in an observational study that evaluates the effectiveness of right heart catheterization on critically ill patients.