Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Toward Relieving Clinician Burden by Automatically Generating Progress Notes using Interim Hospital Data

Sarvesh Soni, Dina Demner-Fushman

TL;DR

This paper proposes a task to automate progress note generation using structured or tabular information present in electronic health records, and presents a novel framework and a large dataset, CHARTPNG, for the task which contains 7089 annotation instances.

Abstract

Regular documentation of progress notes is one of the main contributors to clinician burden. The abundance of structured chart information in medical records further exacerbates the burden, however, it also presents an opportunity to automate the generation of progress notes. In this paper, we propose a task to automate progress note generation using structured or tabular information present in electronic health records. To this end, we present a novel framework and a large dataset, ChartPNG, for the task which contains $7089$ annotation instances (each having a pair of progress notes and interim structured chart data) across $1616$ patients. We establish baselines on the dataset using large language models from general and biomedical domains. We perform both automated (where the best performing Biomistral model achieved a BERTScore F1 of $80.53$ and MEDCON score of $19.61$) and manual (where we found that the model was able to leverage relevant structured data with $76.9\%$ accuracy) analyses to identify the challenges with the proposed task and opportunities for future research.

Toward Relieving Clinician Burden by Automatically Generating Progress Notes using Interim Hospital Data

TL;DR

This paper proposes a task to automate progress note generation using structured or tabular information present in electronic health records, and presents a novel framework and a large dataset, CHARTPNG, for the task which contains 7089 annotation instances.

Abstract

Regular documentation of progress notes is one of the main contributors to clinician burden. The abundance of structured chart information in medical records further exacerbates the burden, however, it also presents an opportunity to automate the generation of progress notes. In this paper, we propose a task to automate progress note generation using structured or tabular information present in electronic health records. To this end, we present a novel framework and a large dataset, ChartPNG, for the task which contains annotation instances (each having a pair of progress notes and interim structured chart data) across patients. We establish baselines on the dataset using large language models from general and biomedical domains. We perform both automated (where the best performing Biomistral model achieved a BERTScore F1 of and MEDCON score of ) and manual (where we found that the model was able to leverage relevant structured data with accuracy) analyses to identify the challenges with the proposed task and opportunities for future research.

Paper Structure

This paper contains 10 sections, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Progress Note Generation
  4. Structured Data for Text Generation
  5. ChartPNG Dataset
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments

Figures (3)

  • Figure 1: Overview of the proposed task to automatically generate the next progress note using the previous note and all interim structured chart data.
  • Figure 2: Overview of the proposed framework for PNG. Prompts used for the generative large language models are shown in teal chat boxes.
  • Figure 3: A detailed example showing the flow of information in the proposed framework as the model makes intermediate and final predictions. The sample run is from the Biomistral model. For brevity and privacy restrictions related to the data source, some parts of the structured data, notes, and the predictions are omitted. Highlighted instances from foohighlight-yellow assessment, foohighlight-green condition, and foohighlight-teal plan illustrate information flow..