Learning Optimal Individualized Treatment Rules from Electronic Health Record Data
By Yuanjia Wang, Peng Wu, Ying Liu, Chunhua Weng, Donglin Zeng. 2016
25 citations recorded by the source corpus.
Medical research is experiencing a paradigm shift from “one-size-fits-all” strategy to a precision medicine approach where the right therapy, for the right patient, and at the right time, will be prescribed. We propose a statistical method to estimate the optimal individualized treatment rules (ITRs) that are tailored according to subject-specific features using electronic health records (EHR) data. Our approach merges statistical modeling and medical domain knowledge with machine learning algorithms to assist personalized medical decision making using EHR. We transform the estimation of optimal ITR into a classification problem and account for the non-experimental features of the EHR data and confounding by clinical indication. We create a broad range of feature variables that reflect both patient health status and healthcare data collection process. Using EHR data collected at Columbia University clinical data warehouse, we construct a decision tree for choosing the best second line therapy for treating type 2 diabetes patients.
Subfields: Artificial Intelligence, Endocrinology, Diabetes and Metabolism
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Eric B. Laber
- Michael R. Kosorok
- Ning Liu
- Jian Tang
- Yanzhi Wang
- Zhiyuan Xu
- Ying Liu
- Leonard A. Stefanski
In Endocrinology, Diabetes and Metabolism
Citations in the corpus
- Deep Reinforcement Learning for Dynamic Treatment Regimes on Medical Registry Data (2017, background): “Methods to identify the optimal DTRs have recently received attentions in the statistical community [7], [8], [9], [10], [11], [12], [13], [14], and most of exsiting methods focused on data from randomized clinical trials.”
- Precision Medicine. (2019, background): “…the mathematical underpinnings of directsearch estimation for optimal treatment regimes and led to a series of refinements including residual outcome weighted learning (Laber and Zhao 2015; Wang et al 2016; Zhou et al 2017) and improved bounds based on efficiency theory (Athey and Wager 2017).”