Does Evidence-Based PTS Treatment Reduce PTS Symptoms and Suicide in Iraq and Afghanistan Veterans Seeking VA Care

Abstract

We succeeded in developing a Natural Language Processing (NLP) System with excellent performance characteristics for determining the type of psychotherapy described in VA clinical notes. We found that positive predictive value was greater than .91 and sensitivity was greater than .92 for all types of posttraumatic stress (PTS) evidence-based treatment(EBT), including Cognitive Processing Therapy (CPT) individual and group, and Prolonged Exposure (PE) individual therapy. We completed seven major annotation projects resulting in over 11,935 documents, 1,297 of which were annotated by four people (quadruple-annotated) and7,226 of which were double annotated. We also developed an NLP system to extract PT Checklist (PCL) scores from clinical notes with excellent accuracy (98 positive predictive value).

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 01, 2017
Accession Number
AD1053852

Entities

People

  • Erin Madden
  • Shira Maguen

Organizations

  • Northern California Institute for Research and Education

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Accuracy
  • Afghanistan
  • Algorithms
  • Cognition
  • Computer Languages
  • Data Analysis
  • Education
  • Health
  • Health Care
  • Information Science
  • Language
  • Machine Learning
  • Mental Health
  • Natural Language Processing
  • Natural Languages
  • Standards
  • Supervised Machine Learning

Fields of Study

  • Psychology

Readers

  • Computational Linguistics
  • Computational Modeling and Simulation
  • Psychological Intervention/Treatment for Stress, Anxiety, PTSD, and Related Emotional and Cognitive Health Symptoms.

Technology Areas

  • AI & ML
  • AI & ML - Information Retrieval