Saved in:
Bibliographic Details
Main Authors: Posada, Jose D., Love, David, Datta, Somalee, Desai, Priya
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.03301
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911646749818880
author Posada, Jose D.
Love, David
Datta, Somalee
Desai, Priya
author_facet Posada, Jose D.
Love, David
Datta, Somalee
Desai, Priya
contents De-identification of clinical text remains essential for secondary use of electronic health records (EHRs), yet public benchmarks such as i2b2 2006/2014 are over a decade old and lack the semantic and demographic diversity of modern narratives. While Large Language Models (LLMs) achieve state-of-the-art zero-shot extraction, enterprise deployment is hindered by compute costs and governance restricting Protected Health Information (PHI) from cloud APIs. We introduce SHIELD (Synthetic Human-annotated Identifier-replaced Entries for Learning and De-identification), a diverse dataset of 1,394 notes with 10,505 gold-standard PHI spans across 9 categories, built via set-cover diversity sampling with human-in-the-loop adjudication. We evaluate four LLMs (two proprietary, two open-weight) to establish a performance ceiling, then distill these capabilities into locally deployable Small Language Models (SLMs). Distributional analysis using Frechet Text Distance and Jensen-Shannon Divergence confirms SHIELD occupies a distinct region of biomedical embedding and vocabulary space versus legacy benchmarks. Our best distilled model matches its teacher on structured PHI categories (DATE, DOCTOR, ID, PATIENT, PHONE) and achieves micro-averaged span-level precision of 0.88 and recall of 0.86 on standard workstation hardware. Cross-dataset evaluation shows diversity-trained models generalize well on universal structured PHI, while institution-specific entities remain hard to transfer, suggesting optimal deployment combines broad-coverage models with specialized models for high-volume notes. We publicly release the SHIELD dataset and the distilled DeBERTa v3 model.
format Preprint
id arxiv_https___arxiv_org_abs_2605_03301
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SHIELD: A Diverse Clinical Note Dataset and Distilled Small Language Models for Enterprise-Scale De-identification
Posada, Jose D.
Love, David
Datta, Somalee
Desai, Priya
Computation and Language
Artificial Intelligence
De-identification of clinical text remains essential for secondary use of electronic health records (EHRs), yet public benchmarks such as i2b2 2006/2014 are over a decade old and lack the semantic and demographic diversity of modern narratives. While Large Language Models (LLMs) achieve state-of-the-art zero-shot extraction, enterprise deployment is hindered by compute costs and governance restricting Protected Health Information (PHI) from cloud APIs. We introduce SHIELD (Synthetic Human-annotated Identifier-replaced Entries for Learning and De-identification), a diverse dataset of 1,394 notes with 10,505 gold-standard PHI spans across 9 categories, built via set-cover diversity sampling with human-in-the-loop adjudication. We evaluate four LLMs (two proprietary, two open-weight) to establish a performance ceiling, then distill these capabilities into locally deployable Small Language Models (SLMs). Distributional analysis using Frechet Text Distance and Jensen-Shannon Divergence confirms SHIELD occupies a distinct region of biomedical embedding and vocabulary space versus legacy benchmarks. Our best distilled model matches its teacher on structured PHI categories (DATE, DOCTOR, ID, PATIENT, PHONE) and achieves micro-averaged span-level precision of 0.88 and recall of 0.86 on standard workstation hardware. Cross-dataset evaluation shows diversity-trained models generalize well on universal structured PHI, while institution-specific entities remain hard to transfer, suggesting optimal deployment combines broad-coverage models with specialized models for high-volume notes. We publicly release the SHIELD dataset and the distilled DeBERTa v3 model.
title SHIELD: A Diverse Clinical Note Dataset and Distilled Small Language Models for Enterprise-Scale De-identification
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2605.03301