Newest papers

From Structured Records to Patient Narratives: Preserving Clinical Alignment in LLM-Generated Health Summaries

Bayram Ayadi, Florian Geißler, Johanna Schmidhuber · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31467 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

A medida que los sistemas de salud portátiles y digitales generan cada vez más registros fisiológicos longitudinales, comunicar estas observaciones en un lenguaje accesible al paciente sigue siendo un reto significativo. Los modelos de lenguaje grande (LLM) ofrecen un mecanismo prometedor para traducir la información clínica estructurada en narrativas legibles, pero este proceso puede introducir una deriva semántica, donde el significado clínicamente preciso se distorsiona durante la simplificación. Investigamos la síntesis de salud basada en evidencias para el monitoreo de la salud menstrual utilizando un marco de generación restringida que combina el condicionamiento de la evidencia estructurada con la Generación Aumentada de Recuperación (RAG) utilizando un marco de generación restringida basado en el prompt. Evaluamos el alineamiento semántico utilizando la Inferencia

Machine translation (Meta M2M100); the English original follows.

Abstract

As wearable and digital health systems increasingly generate longitudinal physiological records, communicating these observations in patient-accessible language remains a significant challenge. Large Language Models (LLMs) offer a promising mechanism for translating structured clinical information into readable narratives, but this process may introduce semantic drift, where clinically precise meaning is distorted during simplification. We investigate evidence-grounded health summarization for menstrual health monitoring using a constrained generation framework that combines structured evidence conditioning with Retrieval-Augmented Generation (RAG) using a prompt-based constrained generation framework. We evaluate semantic alignment using Natural Language Inference (NLI), atomic factuality metrics, and readability measures. Across three LLMs and 4,500 generated summaries, explicit grounding constraints consistently improve clinical alignment, particularly by improving uncertainty preservation and, once faithfulness is evaluated against retrieved evidence, generally preserving or improving faithfulness, while effects on omission remain model-dependent. We also find that improved clinical alignment comes at the cost of increased reading difficulty and medical terminology. These findings highlight the need to jointly evaluate faithfulness, uncertainty preservation, and readability when developing patient-facing medical LLM systems.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Readability, Automatic summarization, Computer science, Unified Medical Language System, Natural language generation, Natural language processing

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange