AgenticSum: An Agentic Inference-Time Framework for Faithful Clinical Text Summarization

TL;DR

AgenticSum is a novel inference-time framework that improves clinical text summarization by decomposing the task into stages of context selection, generation, verification, and correction, enhancing factual consistency.

Contribution

It introduces an agentic, multi-stage approach that reduces hallucinations in clinical summarization by targeted correction during inference, a novel strategy in this domain.

Findings

Consistent improvements over baseline models across datasets.

Effective reduction of hallucinated content in clinical summaries.

Enhanced factual accuracy demonstrated through human and automated evaluations.

Abstract

Large language models (LLMs) offer substantial promise for automating clinical text summarization, yet maintaining factual consistency remains challenging due to the length, noise, and heterogeneity of clinical documentation. We present AgenticSum, an inference-time, agentic framework that separates context selection, generation, verification, and targeted correction to reduce hallucinated content. The framework decomposes summarization into coordinated stages that compress task-relevant context, generate an initial draft, identify weakly supported spans using internal attention grounding signals, and selectively revise flagged content under supervisory control. We evaluate AgenticSum on two public datasets, using reference-based metrics, LLM-as-a-judge assessment, and human evaluation. Across various measures, AgenticSum demonstrates consistent improvements compared to vanilla LLMs and other strong baselines. Our results indicate that structured, agentic design with targeted correction offers an effective inference time solution to improve clinical note summarization using LLMs.