Automatic Cardiac Risk Management Classification using large-context Electronic Patients Health Records

TL;DR

This study develops an automated classification system for cardiac risk management using large-context electronic health records, outperforming traditional and generative models with a specialized Transformer architecture.

Contribution

Introduces a novel hierarchical Transformer model for long-context clinical narratives, demonstrating superior performance over classical and LLM-based approaches in cardiac risk classification.

Findings

Transformer outperforms traditional models and LLMs

Hierarchical attention captures long-range dependencies effectively

Automated approach rivals manual workflows in clinical risk stratification

Abstract

To overcome the limitations of manual administrative coding in geriatric Cardiovascular Risk Management, this study introduces an automated classification framework leveraging unstructured Electronic Health Records (EHRs). Using a dataset of 3,482 patients, we benchmarked three distinct modeling paradigms on longitudinal Dutch clinical narratives: classical machine learning baselines, specialized deep learning architectures optimized for large-context sequences, and general-purpose generative Large Language Models (LLMs) in a zero-shot setting. Additionally, we evaluated a late fusion strategy to integrate unstructured text with structured medication embeddings and anthropometric data. Our analysis reveals that the custom Transformer architecture outperforms both traditional methods and generative \acs{llm}s, achieving the highest F1-scores and Matthews Correlation Coefficients. These findings underscore the critical role of specialized hierarchical attention mechanisms in capturing long-range dependencies within medical texts, presenting a robust, automated alternative to manual workflows for clinical risk stratification.