Training a Large Language Model for Medical Coding Using Privacy-Preserving Synthetic Clinical Data

TL;DR

This paper demonstrates that fine-tuning a large language model with synthetic, privacy-preserving clinical data significantly improves its accuracy in automated medical coding, achieving high exact-match F1 scores for ICD-10-CM and CPT codes.

Contribution

The study shows that using synthetic training data enables a general-purpose LLM to excel at medical coding tasks without risking patient privacy, surpassing zero-shot performance.

Findings

Exact-match F1 score exceeded 0.70 after fine-tuning.

Performance remained high on complex, reasoning-intensive categories.

Synthetic data effectively trains models for real-world medical coding applications.

Abstract

Improving the accuracy and reliability of medical coding reduces clinician burnout and supports revenue cycle processes, freeing providers to focus more on patient care. However, automating the assignment of ICD-10-CM and CPT codes from clinical documentation remains a challenge due to heterogeneous records, nuanced coding guidelines, and long-tail distributions. Large language models have been proposed to help or automate specific medical coding tasks. However, foundation models are not explicitly trained for medical coding and zero-shot coding has yielded poor results. We investigate whether a modern open-weight foundation model can be adapted for an expert-level medical coding task using privacy-preserving synthetic training data derived from electronic health records. We fine-tune Llama 3-70B on pairs of clinical notes and gold codes generated from EHR-grounded templates and coding policies, then evaluate exact-code prediction for ICD-10-CM and CPT. A zero-shot baseline with the unadapted model achieved an F1 score of 0.18 for exact code match. After fine-tuning on the synthetic corpus, exact-match F1 exceeded 0.70, representing a large absolute gain across both code systems. Notably, performance remained high on complex categories that often require multi-step clinical reasoning and code composition, including Advanced Illness and Frailty classes, and the model retained its performance on medical comprehension tasks. These results indicate that synthetic, policy-aware data can efficiently teach a general-purpose large language model to support precise medical coding without exposing protected health information. The approach offers a practical path for training coding agents safely and iteratively on specific tasks that represent real-world populations.