Estimating optimal dynamic treatment regime with model selection in STAR*D trial

This NDA Study is linked to the paper "Model selection with uncertainty in estimating optimal dynamic treatment regimes" under preparation for statistical journal submission. We applied our model selection methodology to STAR*D to learn/estimate the optimal dynamic treatment regime for patients who did not achieve remission from initial treatment. We restrict our attention to two treatment options, augmentation and switch, for each decision. The original STAR*D data consists of four levels where all patients were treated with citalopram at level 1 and only switch options were available for patients involved in level 4. We therefore only focus on level 2 (level 2 plus level 2A if level 2A available) and level 3. The decision rules considered at the two stages (i.e. level 2 and level 3) are restricted to interpretable ones. Specifically, we consider linear and tree-based rules.