mlsanityData to 7 Oct 2026

Synthetic Data for Veterinary EHR De-identification: Benefits, Limits, and Safety Trade-offs Under Fixed Compute

arXiv 2601.09756 ↗First posted 13 Jan 2026cs.CR cs.AI cs.CLcited by 0

On the map: Synthetic data and self-improvement for LLMs → Medical synthetic data and agents

Abstract (arXiv)

Veterinary electronic health records (vEHRs) contain privacy-sensitive identifiers that limit secondary use. While PetEVAL provides a benchmark for veterinary de-identification, the domain remains low-resource. This study evaluates whether large language model (LLM)-generated synthetic narratives improve de-identification safety under distinct training regimes, emphasizing (i) synthetic augmentation and (ii) fixed-budget substitution. We conducted a controlled simulation using a PetEVAL-derived corpus (3,750 holdout/1,249 train). We generated 10,382 synthetic notes using a privacy-preserving "template-only" regime where identifiers were removed prior to LLM prompting. Three transformer backbones (PetBERT, VetBERT, Bio_ClinicalBERT) were trained under varying mixtures. Evaluation prioritized document-level leakage rate (the fraction of documents with at least one missed identifier) as the primary safety outcome. Results show that under fixed-sample substitution, replacing real notes with synthetic ones monotonically increased leakage, indicating synthetic data cannot safely replace real supervision. Under compute-matched training, moderate synthetic mixing matched real-only performance, but high synthetic dominance degraded utility. Conversely, epoch-scaled augmentation improved performance: PetBERT span-overlap F1 increased from 0.831 to 0.850 +/- 0.014, and leakage decreased from 6.32% to 4.02% +/- 0.19%. However, these gains largely reflect increased training exposure rather than intrinsic synthetic data quality. Corpus diagnostics revealed systematic synthetic-real mismatches in note length and label distribution that align with persistent leakage. We conclude that synthetic augmentation is effective for expanding exposure but is complementary, not substitutive, for safety-critical veterinary de-identification.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Medical synthetic data and agents, one of 13 directions in Synthetic data and self-improvement for LLMs. Builds on 17 of its 33 references that are in the map. First posted on arXiv 13 Jan 2026.

Builds on

17 of 33 references.

References · 17

  1. 2020Language Models are Few-Shot Learnerscited by 32,199
  2. 2019BioBERT: a pre-trained biomedical language representation model for biomedical text miningcited by 1,935
  3. 2020Domain-Specific Language Model Pretraining for Biomedical Natural Language Processingcited by 996
  4. 2019Publicly Available Clinical BERT Embeddingscited by 995
  5. 2016De-identification of Patient Notes with Recurrent Neural Networkscited by 88
  6. 2023Two Directions for Clinical Data Generation with Large Language Models: Data-to-Label and Label-to-Datacited by 26
  7. 2024Training and Evaluating Language Models with Template-based Data Generationcited by 6
  8. 2017De-identification of medical records using conditional random fields and long short-term memory networkscited by 5
  9. 2022A Comparative Evaluation Of Transformer Models For De-Identification Of Clinical Text Datacited by 5
  10. 2025How do we measure privacy in text? A survey of text anonymization metricscited by 5
  11. 2023Validating transformers for redaction of text from electronic health records in real-world healthcarecited by 4
  12. 2025Generation of Synthetic Clinical Text: A Systematic Reviewcited by 4
5 more
  1. 2024Disease Entity Recognition and Normalization is Improved with Large Language Model Derived Synthetic Normalized Mentionscited by 3
  2. 2024Fine-tuning foundational models to code diagnoses from veterinary health recordscited by 3
  3. 2024Generating Synthetic Free-text Medical Records with Low Re-identification Risk using Masked Language Modelingcited by 2
  4. 2024Data Augmentation Method Utilizing Template Sentences for Variable Definition Extractioncited by 1
  5. 2024Evaluation of Language Models in the Medical Context Under Resource-Constrained Settingscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024Rephrasing Electronic Health Records for Pretraining Clinical Language Models
  2. 2026Training a Large Language Model for Medical Coding Using Privacy-Preserving Synthetic Clinical Data
  3. 2026Institution-Specific LLM Prompting Recovers PHI That De-identification Systems and Their Gold Standards Both Miss
  4. 2025From Learning to Unlearning: Biomedical Security Protection in Multimodal Large Language Models
  5. 2025Towards Automatic Evaluation and Selection of PHI De-identification Models via Multi-Agent Collaboration
  6. 2024Zero-shot and Few-shot Generation Strategies for Artificial Clinical Records
  7. 2026Fidelity, Diversity, and Privacy: A Multi-Dimensional LLM Evaluation for Clinical Data Augmentation
  8. 2023Reliable Generation of Privacy-preserving Synthetic Electronic Health Record Time Series via Diffusion Models