BYU

Abstract by Luke Sagers

Personal Infomation


Presenter's Name

Luke Sagers

Degree Level

Undergraduate

Co-Authors

Orion Weller

Abstract Infomation


Department

Statistics

Faculty Advisor

Quinn Snell

Title

Validation and Assessment of the Opiod Abuse Risk Screener using Machine Learning Methods

Abstract

The prevalence of opioid abuse in the US presents a dire need for improved assessment and risk stratification tools. The Opioid Abuse Risk Screener (OARS) is a clinician-administered assessment designed to evaluate risk of opioid misuse based on biopsychosocial factors and aberrant behaviors related to opioid abuse. This study aims to validate the ability of OARS and its factors to predict aberrant behaviors including failed urine drug tests (UDTs) and aberrant opiate prescriptions. The data includes OARS responses from 2,051 patients and results of UDTs and controlled substance database checks.  Machine learning methods were applied to build models for predicting aberrant behavior and to identify which items on the OARS showed the strongest association with aberrant opioid related behaviors. The results will help refine the test and better inform clinicians of important risk factors.