Search research articles

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

Does Domain-Specific Retrieval Augmented Generation Help LLMs Answer Consumer Health Questions?

Proceedings of machine learning research·2026

Same author

Impact of Imaging Protocols on Thermal Detection of Pressure Injuries: Threshold versus Deep Learning Across Skin Tones.

medRxiv : the preprint server for health sciences·2026

Same author

Correction: Impact of skin tone, environmental, and technical factors on thermal imaging.

PloS one·2026

Same author

GDA-AM: ON THE EFFECTIVENESS OF SOLVING MIN-IMAX OPTIMIZATION VIA ANDERSON MIXING.

... International Conference on Learning Representations·2026

Same author

Beyond Composite Indices: Comprehensive Social Determinants Improve Heart Failure Readmission Prediction.

Journal of the American Heart Association·2026

Same author

BMRetriever: Tuning Large Language Models as Better Biomedical Text Retrievers.

Proceedings of the Conference on Empirical Methods in Natural Language Processing. Conference on Empirical Methods in Natural Language Processing·2026

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Oct 5, 2025

Touchscreen Sustained Attention Task SAT for Rats

Touchscreen Sustained Attention Task SAT for Rats

Published on: September 15, 2017

CATAN: Chart-aware temporal attention network for adverse outcome prediction.

Zelalem Gero¹, Joyce C Ho¹

¹Department of Computer Science, Emory University, Atlanta, USA.

Proceedings. IEEE International Conference on Healthcare Informatics

|January 26, 2022

Summary

This study introduces CATAN, a novel network for analyzing clinical notes to predict patient outcomes. CATAN improves accuracy in predicting mortality and hospital readmission by considering the timing of medical records.

More Related Videos

A Dual Task Procedure Combined with Rapid Serial Visual Presentation to Test Attentional Blink for Nontargets

A Dual Task Procedure Combined with Rapid Serial Visual Presentation to Test Attentional Blink for Nontargets

Published on: December 5, 2014

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Published on: January 23, 2017

Related Experiment Videos

Last Updated: Oct 5, 2025

Touchscreen Sustained Attention Task SAT for Rats

Touchscreen Sustained Attention Task SAT for Rats

Published on: September 15, 2017

A Dual Task Procedure Combined with Rapid Serial Visual Presentation to Test Attentional Blink for Nontargets

A Dual Task Procedure Combined with Rapid Serial Visual Presentation to Test Attentional Blink for Nontargets

Published on: December 5, 2014

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Published on: January 23, 2017

Area of Science:

Medical Informatics
Artificial Intelligence in Healthcare
Clinical Data Analysis

Background:

Electronic health record (EHR) systems are increasingly adopted in healthcare settings.
Unstructured clinical text within EHRs represents a valuable, yet underutilized, data source for patient insights.
Automated systems can assist healthcare professionals by extracting information from clinical notes.

Purpose of the Study:

To develop CATAN (chart-aware temporal attention network), a novel deep learning model for learning patient representations from clinical notes.
To leverage temporal information (chart-time) within clinical notes to enhance attention mechanisms for better prediction.
To improve the accuracy of predicting patient outcomes such as mortality and hospital readmission.

Main Methods:

Developed CATAN, a chart-aware temporal attention network that treats clinical notes as units with attention-weighted words.
Introduced a hierarchical attention mechanism: word-level attention within notes and note-level attention for patient representation.
Incorporated chart-time as a constraint in attention calculation to prioritize temporally relevant notes.

Main Results:

CATAN achieved superior performance in one-year mortality prediction and 30-day hospital readmission compared to state-of-the-art baselines on the MIMIC-III dataset.
The model's patient representation, informed by temporal attention, proved effective for outcome prediction.
Attention weights provided transparency into the model's decision-making process.

Conclusions:

The proposed chart-aware temporal attention network (CATAN) effectively learns patient representations from unstructured clinical text.
Incorporating temporal information significantly enhances the predictive power of models for clinical outcomes.
CATAN offers a transparent and high-performing approach for leveraging EHR data in clinical decision support.