Identifying and understanding opioid use disorder in clinical notes

Document Type

Conference Proceeding

Publication Date

1-1-2020

Journal

Proceedings of the 12th IADIS International Conference e-Health 2020, EH 2020 - Part of the 14th Multi Conference on Computer Science and Information Systems, MCCSIS 2020

Keywords

Machine Learning; Opioid Abuse Disorder; U.S. Veterans

Abstract

© Proceedings of the 12th IADIS International Conference e-Health 2020, EH 2020 - Part of the 14th Multi Conference on Computer Science and Information Systems, MCCSIS 2020. All rights reserved. Opioid use, abuse and misuse afflicts many populations, including Veterans. The objective of this ten-year retrospective study was to identify documentation of potential opioid abuse, both treated and untreated, in clinical notes, by developing and applying a natural language processing tool to a corpus of clinical notes documenting the healthcare of U.S. Veterans. To better understand the issue of opioid abuse among Veterans, we also extracted descriptive data on prescription counts, patient demographics, and diagnoses. The natural language processing tool we developed achieved F1 scores of 88% and 91% in identifying opioid abuse with treatment, and without treatment, respectively, among U.S. Veterans receiving healthcare in the Baltimore, Maryland and Washington DC VA service regions. This resulted in identifying 809 additional patients experiencing opioid abuse. The descriptive data give insight by elucidating trends that enhance understanding of opioid abuse among Veterans receiving healthcare in these service regions, and suggest future research.

This document is currently not available here.

Share

COinS