Balancing Model Complexity and Clinical Deployability in Deep Learning for Sociodemographic Information Extraction
Journal of Primary Care & Community Health · 2025-12 · 16
DOI: 10.1177/21501319251404193 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Primary Care & Community Health, online erschienen am 16.12.2025. Verfasst von Rawan Abulibdeh und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Electrical and Computer Engineering, University of Toronto, Toronto, Ontario, Canada. Hauptthemen laut MEDLINE: Deep Learning, Electronic Health Records, Information Storage and Retrieval. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 70 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: Canadian Institutes of Health Research.
Abstract (Originalsprache)
Sociodemographic factors are critical determinants of health outcomes and disparities, yet their documentation in electronic medical records is often sparse and confined to unstructured clinical text. This poses substantial challenges for automated extraction and integration into clinical decision-making. In this study, we systematically evaluate and compare 6 convolutional neural network architectures, including hybrid models that integrate traditional classifiers, for binary classification of multiple sociodemographic characteristics from EMR text using data from 4375 patients across 96 primary care clinics. The goal was to assess how model complexity and lexical diversity influence classification performance. Manual annotation achieved high inter-rater reliability (kappa: 0.98 for documentation status, 0.96 for documented information). We report performance using F1 score, precision, recall, area under the precision-recall curve, and Matthews correlation coefficient. Results showed that simpler architectures, particularly a single-layer CNN, consistently outperform deeper or hybrid models across most characteristics (F1 score: 90.99%), especially under conditions of data imbalance and varied documentation patterns. While hybrid models offered gains for well-documented factors like marital status, they were less effective for sparse or diverse characteristics. These findings provide a practical framework for developing efficient, interpretable clinical NLP pipelines and inform model selection strategies for real-world health equity and EMR research applications.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Electrical and Computer Engineering, University of Toronto, Toronto, Ontario, Canada.
- Toronto Western Hospital Family Health Team, University Health Network, Ontario, Canada.
- Department of Family and Community Medicine, University of Toronto, Toronto, Ontario, Canada.
- North York General Hospital, North York, Ontario, Canada.
ORCID-Profile: Rawan Abulibdeh
Förderung
- Canadian Institutes of Health Research
Zitieren
APA
Abulibdeh, R., Tu, K., & Sejdić, E. (2025). Balancing Model Complexity and Clinical Deployability in Deep Learning for Sociodemographic Information Extraction. Journal of Primary Care & Community Health, 16. https://doi.org/10.1177/21501319251404193
Vancouver
Abulibdeh R, Tu K, Sejdić E. Balancing Model Complexity and Clinical Deployability in Deep Learning for Sociodemographic Information Extraction. Journal of Primary Care & Community Health. 2025;16. doi:10.1177/21501319251404193
BibTeX
@article{Abulibdeh2025,
author = {Abulibdeh, Rawan and Tu, Karen and Sejdić, Ervin},
title = {Balancing Model Complexity and Clinical Deployability in Deep Learning for Sociodemographic Information Extraction},
journal = {Journal of Primary Care & Community Health},
year = {2025},
volume = {16},
doi = {10.1177/21501319251404193}
}
EUVIMED