Preprint · 2026 · Open access
Institution-Specific LLM Prompting Recovers PHI That De-identification Systems and Their Gold Standards Both Miss
arXiv (Cornell University) · Published 17 August 2026 · DOI 10.48550/arxiv.2608.17051
In brief
“Institution-Specific LLM Prompting Recovers PHI That De-identification Systems and Their Gold Standards Both Miss” is a 2026 preprint by Daniel Palacios and 8 co-authors, published in arXiv (Cornell University). OpenAlex counts 2 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at arXiv (Cornell University).
- Citations
- 2
- Most cited in
- 2026 (2)
- References
- 45
- Authors
- 9
Abstract
Secondary use of electronic health records requires de-identification, yet existing systems miss \emph{institutionally situated} protected health information (PHI) such as hospital abbreviations, building names, and internal codes whose status is locally determined. We ask whether large language models (LLMs) with in-context learning (ICL) can close this gap and control the precision--recall trade-off. On 100 annotated pediatric oncology notes (5,322 PHI spans) from Texas Children's Hospital, we benchmarked eight LLMs against two purpose-built systems (Stanford TiDE, OpenMed PII) and two pattern-based baselines. Each LLM ran under three prompts of increasing specificity: (1) a HIPAA-aligned baseline, (2) baseline plus the institutional PHI categories it missed, and (3) prompt 2 plus instructions against over-redacting clinical content. We then compared 14~multi-agent and ensemble configurations against the best single prompt, with recall the primary safety metric. LLMs outperformed the purpose-built systems (best F1=0.918$\pm$0.001 vs.\ TiDE 0.779), with advantages concentrated in contextual categories. Naming the missed categories recovered 79\% (48/61) of them, and discouraging over-redaction restored precision. No agentic architecture beat calibrated single-pass prompting (F1 0.906--0.907), but LLM outputs surfaced 414~candidate annotation gaps; re-annotation confirmed 227~PHI spans, against which the final prompt reached recall=0.981 (F1=0.907$\pm$0.002). Well-calibrated ICL resolves both the institutional PHI gap and the precision--recall trade-off in one LLM call per note. LLMs cost more to run than traditional methods, but that cost buys a way to audit the reference standard. LLMs are a legitimate, adaptable alternative to purpose-built de-identification systems; institution-specific prompt development should be the primary adaptation strategy.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Palacios, D., Neeley, M. B., Otto, A. A., Dhamodharan, S., Woodhouse, J. P., Lin, C.-F., Zobeck, M. C., Liu, Z., & Jeong, H.-H. (2026). Institution-Specific LLM Prompting Recovers PHI That De-identification Systems and Their Gold Standards Both Miss. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2608.17051
MLA 9
Palacios, Daniel, et al. “Institution-Specific LLM Prompting Recovers PHI That De-identification Systems and Their Gold Standards Both Miss.” arXiv (Cornell University), 2026. https://doi.org/10.48550/arxiv.2608.17051.
Chicago
Palacios, Daniel, Matthew B. Neeley, Angel Adetomike Otto, Shalini Dhamodharan, John P. Woodhouse, Chi-fan Lin, Mark C. Zobeck, Zhandong Liu, and Hyun-Hwan Jeong. 2026. “Institution-Specific LLM Prompting Recovers PHI That De-identification Systems and Their Gold Standards Both Miss.” arXiv (Cornell University). https://doi.org/10.48550/arxiv.2608.17051.
Harvard
Palacios, D. et al. (2026) ‘Institution-Specific LLM Prompting Recovers PHI That De-identification Systems and Their Gold Standards Both Miss’, arXiv (Cornell University). doi:10.48550/arxiv.2608.17051.
IEEE
D. Palacios et al., “Institution-Specific LLM Prompting Recovers PHI That De-identification Systems and Their Gold Standards Both Miss,” arXiv (Cornell University), 2026, doi: 10.48550/arxiv.2608.17051.
BibTeX
@misc{palacios2026institution,
title = {{Institution-Specific LLM Prompting Recovers PHI That De-identification Systems and Their Gold Standards Both Miss}},
author = {Palacios, Daniel and Neeley, Matthew B. and Otto, Angel Adetomike and Dhamodharan, Shalini and Woodhouse, John P. and Lin, Chi-fan and Zobeck, Mark C. and Liu, Zhandong and Jeong, Hyun-Hwan},
howpublished = {arXiv (Cornell University)},
year = {2026},
doi = {10.48550/arxiv.2608.17051},
url = {https://doi.org/10.48550/arxiv.2608.17051}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Topic and keywords
Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- protected health information
- large language models
- in-context learning
- prompt engineering
- multi-agent LLM systems
- precision-recall trade-off
- HIPAA compliance
Work with this paper
Most-cited papers in Computer Science
- ImageNet classification with deep convolutional neural networksAlex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton · Communications of the ACM · 2017 · 114,726 citations
- Long Short-Term MemorySepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 109,097 citations
- A Mathematical Theory of CommunicationClaude E. Shannon · Bell System Technical Journal · 1948 · 91,736 citations
- Deep learningYann LeCun, Yoshua Bengio, Geoffrey E. Hinton · Nature · 2015 · 90,226 citations
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksShaoqing Ren, Kaiming He, Ross Girshick and 1 more · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2016 · 74,143 citations
Questions about this paper
How do I cite this paper in APA style?
Palacios, D., Neeley, M. B., Otto, A. A., Dhamodharan, S., Woodhouse, J. P., Lin, C.-F., Zobeck, M. C., Liu, Z., & Jeong, H.-H. (2026). Institution-Specific LLM Prompting Recovers PHI That De-identification Systems and Their Gold Standards Both Miss. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2608.17051
How many times has this paper been cited?
2 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access at arXiv (Cornell University): sign in and open the PDF from this page.
Who wrote this paper?
Daniel Palacios, Matthew B. Neeley, Angel Adetomike Otto, Shalini Dhamodharan, John P. Woodhouse, Chi-fan Lin, Mark C. Zobeck, Zhandong Liu, Hyun-Hwan Jeong.
What is the DOI of this paper?
10.48550/arxiv.2608.17051, permanently at https://doi.org/10.48550/arxiv.2608.17051. Sign in and open it from this page.
Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.