Abstract:
|
For many mental disorders, latent mental status from multiple-domain clinical symptoms perform as a better characterization of the underlying disorder than a simple summary score of the symptoms, and they also serve as more reliable and representative features to differentiate treatment responses. We provide a new paradigm for learning optimal individualized treatment rules (ITRs) by modeling patients' latent mental status. We first learn the multi-domain latent states at baseline from the observed symptoms under a machine learning model, through which patients' heterogeneous symptoms are represented using an economical number of latent variables. Following measurement theory, we assume that treatment changes patient's latent mental states but the conditional distribution of the symptoms given the latent states remains the same. We then optimize a value function defined by the latent states after treatment by exploiting a transformation of the observed symptoms without modeling the relation between the latent states before and after treatment. We derive the convergence rate of the proposed estimator for optimal ITR under the latent model.
|