Evaluating Encoder- and LLM-Based Approaches for Robust Indirect Personal Identifier Detection
Year: 2026
Type: conference-paper
Abstract: Removing explicit protected health information does not fully eliminate re-identification risk in clinical text.Contextual attributes such as socio-economic status, institutional affiliations or detailed life circumstances may still enable linkage attacks.These heterogeneous and often sparsely distributed elements are referred to as Indirect Personal Identifiers, i.e., textual elements that are not always identiyfing in isolation but may enable re-identification when combined with external knowledge.They extend de-identification beyond fixed identifier lists and pose new modeling challenges.Therefore, we present a systematic comparison of encoder-only models, prompt-based LLMs and hybrid pipelines for span-level IPI detection in English discharge summaries.A fine-tuned RoBERTa-large model improves on an existing baseline and substantially outperforms ChatGPT-5.2,achieving 0.906 micro-F1 and 0.724 macro-F1, compared to 0.509 micro-F1 and 0.487 macro-F1.Our findings indicate that IPI detection constitutes a distinct modeling regime characterized by class imbalance and high intra-class variability, where scaling model capacity alone does not guarantee macro-level robustness.We show that supervised encoder models currently provide the most reliable foundation for extending anonymization guarantees and future research.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7166594769