Measuring Use of Evidence Based Psychotherapy for Posttraumatic Stress Disorder in a Large National Healthcare System

Adm Policy Ment Health. 2018 Jul;45(4):519-529. doi: 10.1007/s10488-018-0850-5.

Abstract

To derive a method of identifying use of evidence-based psychotherapy (EBP) for post-traumatic stress disorder (PTSD), we used clinical note text from national Veterans Health Administration (VHA) medical records. Using natural language processing, we developed machine-learning algorithms to classify note text on a large scale in an observational study of Iraq and Afghanistan veterans with PTSD and one post-deployment psychotherapy visit by 8/5/15 (N = 255,968). PTSD visits were linked to 8.1 million psychotherapy notes. Annotators labeled 3467 randomly-selected psychotherapy notes (kappa = 0.88) to indicate receipt of EBP. We met our performance targets of overall classification accuracy (0.92); 20.2% of veterans received ≥ one session of EBP over the study period. Our method can assist with identifying EBP use and studying EBP-associated outcomes in routine clinical practice.

Keywords: Evidence-based medicine; Health services utilization; Natural language processing; Posttraumatic stress disorder; Psychotherapy.

Publication types

  • Observational Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms*
  • Cognitive Behavioral Therapy / statistics & numerical data*
  • Evidence-Based Medicine / statistics & numerical data*
  • Family Therapy / statistics & numerical data
  • Humans
  • Implosive Therapy / statistics & numerical data*
  • Machine Learning*
  • Natural Language Processing*
  • Psychotherapy / statistics & numerical data
  • Psychotherapy, Group / statistics & numerical data
  • Retrospective Studies
  • Stress Disorders, Post-Traumatic / therapy*
  • United States
  • United States Department of Veterans Affairs
  • Veterans / psychology