Elamur Research⌕
☰
home / papers / 2026 / JointCoder: Exploring Automated ICD Coding on Real-World Chinese EHRs with a Multi-Agent Framework

JointCoder: Exploring Automated ICD Coding on Real-World Chinese EHRs with a Multi-Agent Framework

Paper · 2026 · Kangjun Liu, Zhenyu Li, Jianlei Wang et al.

Summary

Most automated ICD coding research targets English data, focuses on diagnoses rather than procedures, and depends on closed-source or resource-heavy LLMs. These choices limit its use in Chinese hospitals, where an aging population is increasing coding workloads. JointCoder addresses these issues with two contributions.

First, the authors built a de-identified dataset of 6,747 inpatient records, mostly cardiology, from a top-tier Chinese hospital. Professional coders annotated each record with ICD-10 diagnosis codes and ICD-9-CM-3 procedure codes.

Second, they present a multi-agent system that treats diagnosis and procedure coding as one joint task. It splits the work into stages that mirror how human coders operate, with each stage handled by its own agent. All agents run on locally hosted 1.7B-parameter models, which keeps patient data on site and reduces cost.

JointCoder outperforms prompting-based LLMs, fine-tuned Qwen3 models up to 14B, and the prior system AiCoder on every reported metric. The authors conclude that encoding the real clinical workflow matters more than model size.

Key points

  • Dataset: 6,747 de-identified, cardiology-focused hospitalization records from a Chinese hospital, coded by experts with ICD-10 for diseases and ICD-9-CM-3 for procedures. Each record includes fields such as Chief Complaint, Discharge Diagnosis, Operative Report and Clinical Course Notes.
  • Six-stage pipeline: the process begins with Standardization and Candidate Mining, proceeds to a Prioritization step, applies "Code Also" rules, checks results in a Validation step, and finishes by mapping the outputs to ICD codes. Disease coding uses only some of these stages, while procedure coding runs through all six.
  • Deployment: every agent is a locally deployed 1.7B model, so no closed-source API is needed and patient data stays inside the hospital.
  • Results: JointCoder scores 0.9094 on ICD-10 and 0.8615 on ICD-9-CM-3. AiCoder scores 0.5728 on ICD-9-CM-3. JointCoder also beats Vanilla, Few-shot, CoT and CoT-SC prompting, as well as fine-tuned Qwen3 models from 1.7B to 14B. The CoT-style methods gain recall but over-generate procedures, which lowers their precision.
  • Limitations: the work covers only cardiology and only Chinese-language records, and the system can still hallucinate. The authors position it as support for human coders, not a replacement.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

Also on the byline: Shandong First Medical University.

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksBiochemistry, Genetics and Molecular Biology · 234 worksMolecular Biology · 231 worksHealth Professions · 69 worksHealth Information Management · 50 works

Cite

Cite as Kangjun Liu, Zhenyu Li, Jianlei Wang, Hongjiao Guan, Ying Lian, Guoqiang Chen, Tao Xin and Wenpeng Lu. 2026. JointCoder: Exploring Automated ICD Coding on Real-World Chinese EHRs with a Multi-Agent Framework. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-demo.26

@inproceedings{liu2026jointcoder,
  title = {{JointCoder: Exploring Automated ICD Coding on Real-World Chinese EHRs with a Multi-Agent Framework}},
  author = {Kangjun Liu, Zhenyu Li, Jianlei Wang, Hongjiao Guan, Ying Lian, Guoqiang Chen, Tao Xin and Wenpeng Lu},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-demo.26},
  url = {https://doi.org/10.18653/v1/2026.acl-demo.26}
}

References

52 references, 6 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Leonor Barreiros, Isabel Coutinho, Gonçalo Correia, and Bruno Martins. 2025. Explainable ICD coding via entity linking. In Proceedings of the 2nd Workshop on Patient-Oriented Language Processing, pages 219–227. find
  2. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel Ziegler, Jeffrey Wu, Clemens Winter, and… find
  3. Pengfei Cao, Chenwei Yan, Xiangling Fu, Yubo Chen, Kang Liu, Jun Zhao, Shengping Liu, and Weifeng Chong. 2020. Clinical-Coder: Assigning interpretable ICD-10 codes to chinese clinical notes. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 294–301. find
  4. Pei Chen, Shuai Zhang, and Boran Han. 2024a. CoMM: Collaborative multi-agent, multi-reasoning-path prompting for complex problem solving. In Proceedings of the 17th North American Chapter of the Association for Computational Linguistics, pages 1720–1738. find
  5. Shuaihang Chen, Yuanxing Liu, Wei Han, Weinan Zhang, and Ting Liu. 2024b. A survey on LLM-based multi-agent system: Recent advances and new frontiers in application. arXiv preprint arXiv:2412.17481. — [arXiv:2412.17481](https://arxiv.org/abs/2412.17481) find
  6. Cheng Cui, Ting Sun, Manhui Lin, Tingquan Gao, Yubo Zhang, Jiaxuan Liu, Xueqing Wang, Zelun Zhang, Changda Zhou, Hongen Liu, and 1 others. 2025. PaddleOCR 3.0 technical report. arXiv preprint arXiv:2507.05595. — [arXiv:2507.05595](https://arxiv.org/abs/2507.05595) find
  7. Hang Dong, Matúš Falis, William Whiteley, Beatrice Alex, Joshua Matterson, Shaoxiong Ji, Jiaoyan Chen, and Honghan Wu. 2022. Automated clinical coding: what, why, and where we are? NPJ digital medicine, 5(1):159. find
  8. Zhihao Fan, Lai Wei, Jialong Tang, Wei Chen, Wang Siyuan, Zhongyu Wei, and Fei Huang. 2025. AI hospital: Benchmarking large language models in a multi-agent medical interaction simulator. In Proceedings of the 31st International Conference on Computational Linguistics, pages 10183–10213.
  9. Zhaopeng Feng, Jiayuan Su, Jiamei Zheng, Jiahan Ren, Yan Zhang, Jian Wu, Hongwei Wang, and Zuozhu Liu. 2025. M-MAD: Multidimensional multi-agent debate for advanced machine translation evaluation. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics, pages 7084–…
  10. Nafiseh Hosseini, Khalil Kimiafar, Sayyed Mostafa Mostafavi, Behzad Kiani, Kazem Zendehdel, Armin Zareiyan, and Saeid Eslami. 2021. Factors affecting the quality of diagnosis coding data with a triangulation view: A qualitative study. The International Journal of Health Planning and Management, 36(5… find
  11. Chao-Wei Huang, Shang-Chi Tsai, and Yun-Nung Chen. 2022. PLM-ICD: Automatic ICD coding with pretrained language models. In Proceedings of the 4th Clinical Natural Language Processing Workshop, pages 10–20. find
  12. Paul Jaccard. 1901. Étude comparative de la distribution florale dans une portion des alpes et des jura. Bulletin de la Société Vaudoise des Sciences Naturelles, 37(1):547–579. find
  13. Guna Sekaran Jaganathan, Indika Kahanda, and Upulee Kanewala. 2025. Metamorphic testing for robustness and fairness evaluation of LLM-based automated ICD coding applications. Smart Health, 36(1):100564. find
  14. Kalervo Järvelin and Jaana Kekäläinen. 2002. Cumulated gain-based evaluation of ir techniques. ACM Transactions on Information Systems, 20(4):422–446. find
  15. Supriya Khadka, Xiaorui Jiang, and Vasile Palade. 2025. Data quality in clinical coding: A critical analysis and preliminary study. medRxiv, 1(1):2025–08. find
  16. Simon A Lee and Timothy Lindsey. 2024. Can large language models abstract medical coded language? arXiv preprint arXiv:2403.10822. — [arXiv:2403.10822](https://arxiv.org/abs/2403.10822) find
  17. Haoran Li, Ziyi Su, Yun Xue, Zhiliang Tian, Yiping Song, and Minlie Huang. 2025a. Advancing collaborative debates with role differentiation through multi-agent reinforcement learning. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics, pages 22655–22666. find
  18. Mingyang Li, Viktor Schlegel, Tingting Mu, Warren Del-Pinto, and Goran Nenadic. 2025b. Structured information matters: Explainable ICD coding with patient-level knowledge graphs. arXiv preprint arXiv:2509.09699, pages 1–16. — [arXiv:2509.09699](https://arxiv.org/abs/2509.09699) find
  19. Rumeng Li, Xun Wang, and Hong Yu. 2025c. Improving rare and common ICD coding via a multi-agent LLM-based approach. In Proceedings of the 34th ACM International Conference on Information and Knowledge Management, pages 4945–4949. find
  20. Xiaobo Li, Yijia Zhang, Xiaodi Hou, Shilong Wang, and Hongfei Lin. 2025d. Deep learning for automatic ICD coding: Review, opportunities and challenges. Artificial Intelligence in Medicine, 168(1):103187. find
  21. Xinyi Li, Sai Wang, Siqi Zeng, Yu Wu, and Yi Yang. 2024. A survey on LLM-based multi-agent systems: workflow, infrastructure, and challenges. Vicinagearth, 1(1):9. find
  22. Zhenpeng Liang, Hongjiao Guan, Wenpeng Lu, Xueping Peng, Bing Xu, and Muyun Yang. 2025a. HiRes: Hierarchical feature optimization and rescorer for automatic ICD coding. In Proceedings of the 50th IEEE International Conference on Acoustics, Speech, and Signal Processing, pages 1–5. find
  23. Zhenpeng Liang, Hongjiao Guan, Weiyu Zhang, Ying Lian, Bing Xu, and Wenpeng Lu. 2025b. AiCoder: Exploring automated ICD coding on chinese EMRs with a multi-agent framework. In Proceedings of the 19th IEEE International Conference on Bioinformatics and Biomedicine, pages 3823–3827. find
  24. Chang Lu, Chandan Reddy, Ping Wang, and Yue Ning. 2023. Towards semi-structured automatic ICD coding via tree-based contrastive learning. In Proceedings of the 37th Conference on Neural Information Processing Systems, pages 68300– 68315. find
  25. Wenpeng Lu, Kangjun Liu, Jianlei Wang, Xueping Peng, Tao Shen, Fa Zhu, Weiyu Zhang, Jiabing Zhu, Tao Xin, and Athanasios V Vasilakos. 2025. Advancing Chinese conversation-based patient guidance with a benchmark and knowledge-evolvable assistant. IEEE Journal of Biomedical and Health Informatics, 1(1… find
  26. Julia Medori and Cédrick Fairon. 2010. Machine learning and features selection for semi-automatic ICD-9-CM encoding. In Proceedings of the 11th Annual Meeting of the North American chapter of the Association for Computational Linguistics, pages 84–89. find
  27. Andreas Motzfeldt, Joakim Edin, Casper L Christensen, Christian Hardmeier, Lars Maaløe, and Anna Rogers. 2025. Code like humans: A multi-agent solution for medical coding. In Proceedings of the 30th Conference on Empirical Methods in Natural Language Processing, pages 22612–22627. find
  28. James Mullenbach, Sarah Wiegreffe, Jon Duke, Jimeng Sun, and Jacob Eisenstein. 2018. Explainable prediction of medical codes from clinical text. In Proceedings of the 16th North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1101–1111.
  29. Akram Mustafa, Usman Naseem, and Mostafa Rahimi Azghadi. 2025. Large language models vs human for classifying clinical documents. International Journal of Medical Informatics, 195(1):105800. find
  30. National Center for Health Statistics. 1980. International Classification of Diseases, Ninth Revision, Clinical Modification: Procedures: Tabular List and Alphabetic Index. U.S. Department of Health and Human Services. find
  31. Adler Perotte, Rimma Pivovarov, Karthik Natarajan, Nicole Weiskopf, Frank Wood, and Noémie Elhadad. 2014. Diagnosis code assignment: models and evaluation metrics. Journal of the American Medical Informatics Association, 21(2):231–237. find
  32. Haoran Shi, Pengtao Xie, Zhiting Hu, Ming Zhang, and Eric P Xing. 2017. Towards automated ICD coding using deep learning. arXiv preprint arXiv:1711.04075. find
  33. Fei Teng, Yiming Liu, Tianrui Li, Yi Zhang, Shuangqing Li, and Yue Zhao. 2022. A review on deep neural networks for ICD coding. IEEE Transactions on Knowledge and Data Engineering, 35(5):4357–4375. find
  34. Khanh-Tung Tran, Dung Dao, Minh-Duong Nguyen, Quoc-Viet Pham, Barry O’Sullivan, and Hoang D Nguyen. 2025. Multi-agent collaboration mechanisms: A survey of LLMs. arXiv preprint arXiv:2501.06322. find
  35. Thanh Vu, Dat Quoc Nguyen, and Anthony Nguyen. 2021. A label attention model for ICD coding from clinical text. In Proceedings of the 29th International Conference on International Joint Conferences on Artificial Intelligence, pages 3335–3341. find
  36. Rui Wang, Yonghe Chen, Weiyu Zhang, Jiasheng Si, Hongjiao Guan, Xueping Peng, and Wenpeng Lu. 2025. MedConMA: A confidence-driven multi-agent framework for medical Q&A. In Proceedings of the 29th Pacific-Asia Conference on Knowledge Discovery and Data Mining, pages 421–433. find
  37. Xindi Wang, Robert Mercer, and Frank Rudzicz. 2024. Multi-stage retrieve and re-rank model for automatic medical coding recommendation. In Proceedings of the 17th Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4881–4891. find
  38. Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc V Le, Ed H Chi, Sharan Narang, Aakanksha Chowdhery, and Denny Zhou. 2023. Self-consistency improves chain of thought reasoning in language models. In Proceedings of the 11th International Conference on Learning Representations, pages 1–24.
  39. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, and 1 others. 2022. Chain-of-thought prompting elicits reasoning in large language models. In Proceedings of the 36th Conference on Neural Information Processing Systems, pages 24824–24837.
  40. World Health Organization. 2004. ICD-10: International statistical classification of diseases and related health problems: 10th revision. World Health Organization. find
  41. John Wu, David Wu, and Jimeng Sun. 2024a. Beyond label attention: Transparency in language models for automated medical coding via dictionary learning. In Proceedings of the 29th Conference on Empirical Methods in Natural Language Processing, pages 8848–8871. find
  42. Yuzhou Wu, Xuechen Chen, Xin Yao, Yongang Yu, and Zhigang Chen. 2024b. Hyperbolic graph convolutional neural network with contrastive learning for automated ICD coding. Computers in Biology and Medicine, 168(1):107797. find
  43. Suyang Xi, Jiesen Shi, Jiachen Yan, MingJing Lin, Xinyi Zhou, Yuan Cheng, Hong Ding, and Chia Chao Kang. 2025. Breaking barriers in ICD classification with a robust graph neural network for hierarchical coding. Scientific Reports, 15(1):25676. find
  44. Pengtao Xie and Eric Xing. 2018. A neural architecture for automated ICD coding. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics, pages 1066–1076. find
  45. Chenwei Yan, Xiangling Fu, Xien Liu, Yuanqiu Zhang, Yue Gao, Ji Wu, and Qiang Li. 2022. A survey of automated international classification of diseases coding: Development, challenges, and applications. Intelligent Medicine, 2(3):161– 173. 1 others. 2025. Qwen3 technical report. arXiv preprint arXiv:…
  46. Danchen Zhang, Daqing He, Sanqiang Zhao, and Lei Li. 2017. Enhancing automatic ICD-9-CM code assignment for medical texts with pubmed. In Proceedings of the 16th Biomedical Natural Language Processing Workshop, pages 263–271. find
  47. Kaiyuan Zhang, Qian Liu, Luyang Zhang, Chaoqun Zheng, Shuaimin Li, Bing Xu, Muyun Yang, Xinxiao Qiao, and Wenpeng Lu. 2025a. MADAWSD: Multi-agent debate framework for adversarial word sense disambiguation. In Proceedings of the 30th Conference on Empirical Methods in Natural Language Processing, pag… find
  48. Xu Zhang, Kun Zhang, Wenxin Ma, Rongsheng Wang, Chenxu Wu, Yingtai Li, and S Kevin Zhou. 2025b. A general knowledge injection framework for ICD coding. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics, pages 7180–7189. find
  49. Xueguan Zhao, Wenpeng Lu, Chaoqun Zheng, Weiyu Zhang, Jiasheng Si, and Deyu Zhou. 2025. Plan dynamically, express rhetorically: A debate-driven rhetorical framework for argumentative writing. In Proceedings of the 30th Conference on Empirical Methods in Natural Language Processing, pages 9562–9584. find
  50. Zhizhuo Zhao, Wenpeng Lu, Xueping Peng, Lumin Xing, Weiyu Zhang, and Chaoqun Zheng. 2024. Automated ICD coding via contrastive learning with back-reference and synonym knowledge for smart self-diagnosis applications. IEEE Transactions on Consumer Electronics, 70(3):6042– 6053. find
  51. Chang Zong, Yuchen Yan, Weiming Lu, Jian Shao, Yongfeng Huang, Heng Chang, and Yueting Zhuang. 2024. Triad: A framework leveraging a multi-role LLM-based agent to solve knowledge base question answering. In Proceedings of the 29th Conference on Empirical Methods in Natural Language Processing, pages… find
  52. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, and find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/liu-2026-jointcoder-exploring-automated-icd-coding/