Aspect-Oriented Summarization for Psychiatric Short-Term Readmission Prediction

2025-02-14

AI Summary

Predicting 30-day psychiatric readmission remains challenging due to the clinical complexity and heterogeneity of patient data. Method: This paper proposes a prompt-driven, multi-dimensional (symptoms, medications, social support) LLM summarization framework leveraging complementary long-document summaries from LLaMA-3 and Qwen. It introduces a novel semantic divergence metric based on embedding KL divergence and attention entropy to enable interpretable and controllable fusion of multiple summary signals, bridging zero-shot LLMs and supervised Transformer models. Contribution/Results: Evaluated on real-world data from four hospitals, the method achieves +5.2% AUC and +7.8% F1 over single-summary baselines and outperforms end-to-end fine-tuning, demonstrating both efficacy and generalizability of multi-perspective summary fusion for complex clinical prediction tasks.

Technology Category

• Natural Language Processing: Summarization
• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Data Visualization & Summarization

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Recent progress in large language models (LLMs) has enabled the automated processing of lengthy documents even without supervised training on a task-specific dataset. Yet, their zero-shot performance in complex tasks as opposed to straightforward information extraction tasks remains suboptimal. One feasible approach for tasks with lengthy, complex input is to first summarize the document and then apply supervised fine-tuning to the summary. However, the summarization process inevitably results in some loss of information. In this study we present a method for processing the summaries of long documents aimed to capture different important aspects of the original document. We hypothesize that LLM summaries generated with different aspect-oriented prompts contain different extit{information signals}, and we propose methods to measure these differences. We introduce approaches to effectively integrate signals from these different summaries for supervised training of transformer models. We validate our hypotheses on a high-impact task -- 30-day readmission prediction from a psychiatric discharge -- using real-world data from four hospitals, and show that our proposed method increases the prediction performance for the complex task of predicting patient outcome.

Problem

• Predict 30-day psychiatric readmission
• Integrate aspect-oriented summarization signals
• Enhance transformer model prediction performance

Innovation

• Aspect-Oriented Summarization method
• Integrates multiple summary signals
• Enhances transformer model training