Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

What's in a Note? Unpacking Predictive Value in Clinical Note Representations.

Willie Boag1, Dustin Doss1, Tristan Naumann1

  • 1Massachusetts Institute of Technology, Cambridge, MA, USA.

AMIA Joint Summits on Translational Science Proceedings. AMIA Joint Summits on Translational Science
|June 12, 2018
PubMed
Summary

This study explores how to better understand clinical notes within Electronic Health Records (EHRs). We evaluated methods to unlock insights from unstructured text for improved clinical prediction models.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Evaluating the robustness and readiness of large frontier models in health AI applications.

Nature medicine·2026
Same author

CLIN-SUMM: Incremental Longitudinal Summarization of Clinical Notes Enables Scalable Representation and Early Disease Prediction.

medRxiv : the preprint server for health sciences·2026
Same author

Methodological and regulatory considerations for causal AI in drug development.

NPJ digital medicine·2026
Same author

Multimodal AI generates virtual population for tumor microenvironment modeling.

Cell·2025
Same author

Towards responsible artificial intelligence in healthcare-getting real about real-world data and evidence.

Journal of the American Medical Informatics Association : JAMIA·2025
Same author

Toward responsible AI governance: Balancing multi-stakeholder perspectives on AI in healthcare.

International journal of medical informatics·2025

Area of Science:

  • Health Informatics
  • Natural Language Processing
  • Clinical Data Science

Background:

  • Electronic Health Records (EHRs) are increasingly adopted, but the unstructured narrative text in clinical notes remains underutilized.
  • Current methods often use simple information extraction for downstream prediction tasks, evaluating representations extrinsically.
  • Extrinsic evaluations may not capture the full richness of clinical prose.

Purpose of the Study:

  • To investigate intrinsic and extrinsic methods for understanding clinical note representations.
  • To assess the power of expressive clinical prose beyond simple prediction enhancement.
  • To promote transparency and reproducibility in clinical NLP research.

Main Methods:

  • Applied both intrinsic and extrinsic evaluation techniques to common clinical note representations.

Related Experiment Videos

  • Utilized publicly available datasets for all experiments.
  • Developed and shared code to ensure replicability for the clinical modeling community.

    • Main Results:

    • Demonstrated that extrinsic evaluations alone do not fully capture the insights within clinical notes.
    • Identified specific strengths and weaknesses of different note representations using combined evaluation methods.
    • Showcased the potential for deeper understanding of clinical text.

    Conclusions:

    • A combined intrinsic and extrinsic approach offers a more comprehensive understanding of clinical note representations.
    • Better understanding of note representations can lead to more effective clinical NLP applications.
    • Open-source code and public data facilitate community-driven advancements in clinical informatics.