• A Decision Tree Analysis of Opioid and Prescription Drug Interactions Leading to Death Using the FAERS Database

    Author(s):
    Rohit R. Dixit (see profile) , Robert P. Schumaker, Michael A. Veronin
    Date:
    2018
    Subject(s):
    Artificial intelligence
    Item Type:
    Abstract
    Tag(s):
    Adverse Drug Events, decision support, Food and Drug Administration, machine learning
    Permanent URL:
    https://doi.org/10.17613/1q3s-cc46
    Abstract:
    Can unknown and possibly dangerous interactions between opioids and prescription drugs be identified? Is it possible? Our research seeks to answer these questions by applying a supervised machine learning algorithm to the FDA’s Adverse Event Reporting System (FAERS). We trained a decision tree classifier to investigate heroin and prescription drug interactions with an accuracy of 84.9%. We found that heroin and buprenorphine, a commonly prescribed opioid detox drug, led to a 28.0% survival rate among patients. Heroin, buprenorphine, and quinine were even deadlier with a 24.0% survival rate. Our technique can be applied to previously unknown drug combinations to predict mortality and perhaps improve patient safety.
    Metadata:
    xml
    Published as:
    Conference proceeding     Show details
    Publisher:
    INTERNATIONAL INFORMATION MANAGEMENT ASSOCIATION
    Pub. Date:
    2018
    Proceeding:
    IIMA/ICITED Joint Conference 2018
    Page Range:
    67 - 67
    Status:
    Published
    Last Updated:
    3 months ago
    License:
    Attribution

    Downloads

    Item Name: pdf iima-icited2018-abstractbook-final-feb-10-2019.pdf
      Download View in browser
    Activity: Downloads: 18

    Back to Deposits