Preprint · 2026 · Open access

Privacy-Preserving Distilled Large Language Models Enhance Multimorbidity Scoring

Raghav Awasthi, Yihe Yang, Mengxuan Li, Xiaofeng Zhu

medRxiv · Published 21 September 2026 · DOI 10.64898/2026.09.19.26363476

Open PDFView at the publisherCite this paper

In brief

“Privacy-Preserving Distilled Large Language Models Enhance Multimorbidity Scoring” is a 2026 preprint by Raghav Awasthi and 3 co-authors, published in medRxiv. OpenAlex counts 0 citations as of 19 November 2026. It is open access: the full text is free to read at medRxiv.

Citations
0
References
57
Authors
4

Abstract

The truthful use of large language models (LLMs) is a growing challenge in safeguarding sensitive patient data from leakage. We introduce and evaluate a privacy-preserving knowledge distillation framework for LLM-based clinical modeling, using multimorbidity scoring as a healthcare task. Although LLMs can encode rich clinical knowledge and improve upon traditional rule-based comorbidity scoring, their direct evaluation on large-scale biobank data remains constrained by patient privacy. In our framework, multimorbidity reasoning is distilled from state-of-the-art LLM teacher models into compact student models (CoLLMs) using synthetic cohorts that preserve UK Biobank distributions, without exposing real patient data. This approach achieves high-fidelity knowledge transfer (Spearman ρ = 0.75–0.89). Independent LLM-as-Judge evaluation confirms the clinical significance of the distilled knowledge and reveals substantial variability among teacher models. When applied to real UK Biobank data, CoLLM-derived multimorbidity scores improve survival prediction (C-index up to 0.91) and exhibit higher SNP heritability (h 2 ≈ 0.05). Our work establishes a trustworthy, privacy-compliant pathway for large-scale healthcare applications of LLMs.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Awasthi, R., Yang, Y., Li, M., & Zhu, X. (2026). Privacy-Preserving Distilled Large Language Models Enhance Multimorbidity Scoring. medRxiv. https://doi.org/10.64898/2026.09.19.26363476

MLA 9

Awasthi, Raghav, et al. “Privacy-Preserving Distilled Large Language Models Enhance Multimorbidity Scoring.” medRxiv, 2026. https://doi.org/10.64898/2026.09.19.26363476.

Chicago

Awasthi, Raghav, Yihe Yang, Mengxuan Li, and Xiaofeng Zhu. 2026. “Privacy-Preserving Distilled Large Language Models Enhance Multimorbidity Scoring.” medRxiv. https://doi.org/10.64898/2026.09.19.26363476.

Harvard

Awasthi, R. et al. (2026) ‘Privacy-Preserving Distilled Large Language Models Enhance Multimorbidity Scoring’, medRxiv. doi:10.64898/2026.09.19.26363476.

IEEE

R. Awasthi, Y. Yang, M. Li, and X. Zhu, “Privacy-Preserving Distilled Large Language Models Enhance Multimorbidity Scoring,” medRxiv, 2026, doi: 10.64898/2026.09.19.26363476.

BibTeX

@misc{awasthi2026privacy,
  title     = {{Privacy-Preserving Distilled Large Language Models Enhance Multimorbidity Scoring}},
  author    = {Awasthi, Raghav and Yang, Yihe and Li, Mengxuan and Zhu, Xiaofeng},
  howpublished = {medRxiv},
  year      = {2026},
  doi       = {10.64898/2026.09.19.26363476},
  url       = {https://doi.org/10.64898/2026.09.19.26363476}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Topic and keywords

Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare

  • knowledge distillation
  • large language models
  • privacy-preserving machine learning
  • synthetic cohort
  • UK Biobank
  • LLM-as-judge evaluation
  • survival prediction
  • SNP heritability

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Awasthi, R., Yang, Y., Li, M., & Zhu, X. (2026). Privacy-Preserving Distilled Large Language Models Enhance Multimorbidity Scoring. medRxiv. https://doi.org/10.64898/2026.09.19.26363476

How many times has this paper been cited?

0 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 medRxiv: sign in and open the PDF from this page.

Who wrote this paper?

Raghav Awasthi, Yihe Yang, Mengxuan Li, Xiaofeng Zhu.

What is the DOI of this paper?

10.64898/2026.09.19.26363476, permanently at https://doi.org/10.64898/2026.09.19.26363476. 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.