# Can large language models be privacy preserving and fair medical coders?

arXiv: https://arxiv.org/abs/2412.05533 · HTML: https://mlsanity.com/paper/2412.05533/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 7 of its 28 references that are in the map.
- First posted on arXiv 7 Dec 2024.

## Builds on (most cited first, 7 of 7)

- [HuggingFace's Transformers: State-of-the-art Natural Language Processing](https://mlsanity.com/paper/1910.03771/) (2019)
- [MEDITRON-70B: Scaling Medical Pretraining for Large Language Models](https://mlsanity.com/paper/2311.16079/) (2023)
- [R'enyi Differential Privacy of the Sampled Gaussian Mechanism](https://mlsanity.com/paper/1908.10530/) (2019)
- [ICD Coding from Clinical Text Using Multi-Filter Residual Convolutional Neural Network](https://mlsanity.com/paper/1912.00862/) (2019)
- [A Label Attention Model for ICD Coding from Clinical Text](https://mlsanity.com/paper/2007.06351/) (2020)
- [Learning Differentially Private Recurrent Language Models](https://mlsanity.com/paper/1710.06963/) (2017)
- [On the accuracy and efficiency of group-wise clipping in differentially private optimization](https://mlsanity.com/paper/2310.19215/) (2023)

## Built on by (most cited first, 1 of 1)

- [How to Train Private Clinical Language Models: A Comparative Study of Privacy-Preserving Pipelines for ICD-9 Coding](https://mlsanity.com/paper/2511.14936/) (2025)

## Abstract (arXiv)

Protecting patient data privacy is a critical concern when deploying machine learning algorithms in healthcare. Differential privacy (DP) is a common method for preserving privacy in such settings and, in this work, we examine two key trade-offs in applying DP to the NLP task of medical coding (ICD classification). Regarding the privacy-utility trade-off, we observe a significant performance drop in the privacy preserving models, with more than a 40% reduction in micro F1 scores on the top 50 labels in the MIMIC-III dataset. From the perspective of the privacy-fairness trade-off, we also observe an increase of over 3% in the recall gap between male and female patients in the DP models. Further understanding these trade-offs will help towards the challenges of real-world deployment.