Machine Learning Pilot for Electronic Phenotyping of Health Outcomes of Interest

    Basic Details
    Date Posted
    Tuesday, November 13, 2018
    Status
    Complete
    Description

    The aim of this project is to demonstrate the feasibility and efficiency of developing and validating of a claims-based health outcome interest (HOI) algorithm using machine learning classification techniques applied to a linked claims-electronic medical records (EMR) database. This project has the potential to improve the electronic phenotype development and validation process for outcomes that can be detected via standardized information in an EMR to accelerate validation of claims-based signatures.

    Information
    Time Period
    2016 - 2017
    HOI Study Type
    Novel Approaches to More Efficient Outcome Validation
    Population / Cohort
    All relevant records from the IBM Watson Health Claims EMR Database will be extracted for patients meeting the HOI inclusion criteria developed during the study
    Data Source(s)
    IBM Watson Health Claims EMR Database
    Workgroup Leader(s)

    Jenna Wong, PhD, MSc; Darren Toh, ScD; Department of Population Medicine, Harvard Pilgrim Health Care Institute and Harvard Medical School, Boston, MA

    Teresa B. Gibson, PhD, MS, MA; IBM Watson Health, Ann Arbor, MI

    Michael Nguyen, MD; Center for Drug Evaluation and Research, FDA, Silver Spring, MD

    Workgroup Member(s)

    James Williams, MBA; Department of Population Medicine, Harvard Pilgrim Health Care Institute and Harvard Medical School, Boston, MA

    Emma Mollenhauer, BS; David Lewandowski; Timothy Burrell, MD, MBA; Shannon Harrer; IBM Watson Health, Ann Arbor, MI

    Sai Dharmarajan, PhD, MS; Wei Hua, PhD, MHS, MS; Rita Ouellet-Hellstrom, PhD, MPH; Elande Baro, PhD, MS; Robert Ball, MD, MPh; Center for Drug Evaluation and Research, FDA, Silver Spring, MD