ClinStructor: AI-Powered Structuring of Unstructured Clinical Texts

2025/11/14 by K Karthikeyan, K, Karthikeyan, Raghuveer Thirukovalluru +3 · 1 citation
Computer Science · Medicine · #Artificial Intelligence in Healthcare and Education #Computation and Language (cs.CL) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning in Healthcare #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2511.11883

openalex publication_date 2025/11/14 · openalex created_date 2025/11/19 · openalex updated_date 2026/07/28

Abstract

Clinical notes contain valuable, context-rich information, but their unstructured format introduces several challenges, including unintended biases (e.g., gender or racial bias), and poor generalization across clinical settings (e.g., models trained on one EHR system may perform poorly on another due to format differences) and poor interpretability. To address these issues, we present ClinStructor, a pipeline that leverages large language models (LLMs) to convert clinical free-text into structured, task-specific question-answer pairs prior to predictive modeling. Our method substantially enhances transparency and controllability and only leads to a modest reduction in predictive performance (a 2-3% drop in AUC), compared to direct fine-tuning, on the ICU mortality prediction task. ClinStructor lays a strong foundation for building reliable, interpretable, and generalizable machine learning models in clinical environments.

Citations

• Enhanced Electronic Health Records Text Summarization Using Large Language Models
• Dynamic and Adaptive Feature Generation with LLM
• MediQ: Question-Asking LLMs and a Benchmark for Reliable Interactive Clinical Reasoning
• Exploring LLM Multi-Agents for ICD Coding
• Evaluating LLM -- Generated Multimodal Diagnosis from Medical Images and Symptom Analysis
• Large language models in healthcare and medical domain: A review
• Interpretable-by-Design Text Understanding with Iteratively Generated Concept Bottleneck
• Large Language Models in Finance: A Survey
• FineEHR: Refine Clinical Note Representations to Improve Mortality Prediction
• Large Language Models Encode Clinical Knowledge
• LMPriors: Pre-Trained Language Models as Task-Specific Priors
• Large Language Models are Few-Shot Clinical Information Extractors
• Revisiting Methods for Finding Influential Examples
• Clinical Outcome Prediction from Admission Notes using Self-Supervised Knowledge Integration
• The Language Interpretability Tool: Extensible, Interactive Visualizations and Analysis for NLP Models
• Concept Bottleneck Models
• Med-BERT: pre-trained contextualized embeddings on large-scale structured electronic health records for disease prediction
• Neural Additive Models: Interpretable Machine Learning with Neural Nets
• Estimating Training Data Influence by Tracing Gradient Descent
• Mitigating Gender Bias in Natural Language Processing: Literature Review
• A Multiscale Visualization of Attention in the Transformer Model
• ClinicalBERT: Modeling Clinical Notes and Predicting Hospital Readmission
• Techniques for Interpretable Machine Learning
• A Unified Approach to Interpreting Model Predictions
• Prototypical Networks for Few-shot Learning
• Understanding Black-box Predictions via Influence Functions
• Axiomatic Attribution for Deep Networks
• "Why Should I Trust You?": Explaining the Predictions of Any Classifier
• Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps

Cited by

• Additive Large Language Models for Semi-Structured Text

Related

• ClinicRealm: Re-evaluating Large Language Models with Conventional Machine Learning for Non-Generative Clinical Prediction Tasks
• BRIDGE: Benchmarking Large Language Models for Understanding Real-world Clinical Practice Text
• Generalist Foundation Models Are Not Clinical Enough for Hospital Operations
• Scaling Clinician-Grade Feature Generation from Clinical Notes with Multi-Agent Language Models
• MediFact at MEDIQA-CORR 2024: Why AI Needs a Human Touch