FHexchange: Resources for Family Health History Extraction and Normalization From Consumer Dialog Sources

2026-06-25 · Source: Paper Index on ACL Anthology · Field: Science & Research — Health & Medical Research, Mathematics & Computational Sciences · Depth: Expert, quick

Summary

FHexchange is a new, publicly available resource introduced at BioNLP 2026, designed to benchmark information extraction and entity linking for family health history (FHx) from consumer dialog sources. This resource addresses the critical need to evaluate emerging generative AI tools, such as ambient AI scribes and conversational agents, on dynamic patient-clinician and patient-agent dialogs. It comprises two new datasets of patient FHx dialog documents, distinct from clinician-entered data, which capture unique semantic and content characteristics of patient-reported information. FHexchange includes novel annotations for family members, clinical observations, related entities, and standardized UMLS CUIs, providing the clinical NLP community a robust evaluation platform.

Key takeaway

For NLP engineers and research scientists developing clinical AI tools, FHexchange offers a crucial resource for evaluating family health history extraction. You should utilize these new datasets to benchmark your generative AI models on patient-reported dialogs, ensuring your systems accurately process the unique semantic and content characteristics of consumer-generated health information. This directly supports the development of more patient-centered and robust clinical AI applications.

Key insights

FHexchange provides datasets and annotations for benchmarking FHx extraction from patient dialogs for clinical NLP.

Principles

• Patient-reported data has distinct semantic characteristics.
• FHx in free-text requires processing for utility.
• AI scribes need evaluation on dynamic dialogs.

Method

FHexchange was created by developing two new patient FHx dialog datasets and annotating them for family members, clinical observations, related entities, and standardized UMLS CUIs.

In practice

• Benchmark generative AI tools for FHx extraction.
• Evaluate entity linking in patient dialogs.
• Study patient-centered healthcare semantics.

Topics

• Family Health History
• Information Extraction
• Entity Linking
• Clinical NLP
• Generative AI
• Patient Dialog Data

Best for: AI Scientist, NLP Engineer, Research Scientist

Related on AIssential

• Healthsea: an end-to-end spaCy pipeline for exploring health supplement effects — Healthsea is a spaCy pipeline that uses ML/NLP to extract health effects from supplement user reviews.
• Exploring automatic terminology extraction from historical medical data — Combining multiple automatic methods significantly improves terminology extraction from historical texts with non-standard orthography.
• Team MKC at CLPsych 2026: Capturing and Characterizing Mental Health Changes through Social Media Timeline Dynamics — An LLM-based pipeline analyzes social media timelines for early mental health change detection and continuous monitoring, integrating post and user-level assessments.
• Evaluating FrameNet-Based Semantic Modeling for Gender-Based Violence Detection in Clinical Records — FrameNet-based semantic annotation significantly improves GBV detection in clinical text over categorical data alone.
• ModernBERT is more efficient than conventional BERT for chest CT findings classification in Japanese radiology reports — ModernBERT offers computational efficiency for Japanese medical text classification but requires diverse data for robustness.
• Identifying Signs and Symptoms of Urinary Tract Infection from Emergency Department Clinical Notes Using Large Language Models — Prodigy is an efficient, scriptable annotation tool enabling data scientists to accelerate NLP development.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Paper Index on ACL Anthology.