We are happy to announce the release of the ICD Helper module: an AI-assisted ICD-10-CM coding tool developed as part of our Master’s thesis at University of Louvain, Belgium.
The module adds an offline coding assistant directly into the OpenMRS clinical workflow. Clinicians can select a visit note from the patient dashboard, analyze it with embedded AI models, and save suggested ICD-10-CM diagnosis codes to the patient’s record, all without any external API or cloud dependency at inference time.A big thanks to the OpenMRS community for the resources and documentation that made this possible!
If you work in a clinical environment and are interested in trying the module, we would also love to hear about your experience. We have prepared a short usability survey; it only takes a few minutes and would be incredibly valuable for our master thesis:
At Partners In Health and our sites around the world (Haiti, Sierra Leone, etc), we use WHO ICD-10 not ICD-10-CM. Someday in the future, we’ll move to using ICD-11. Can your module be customized to work with ICD-10? That would be more valuable to us and others.
Thank you, it is encouraging to hear real-world deployment feedback from Partners In Health directly!
To answer your question honestly: partial adaptation would be feasible in a near-term release, but a full adaptation towards ICD-10-WHO remains a challenge.
For the selection mode (SapBERT), switching to WHO ICD-10 is achievable without retraining the model; it only requires rebuilding the precomputed embedding index against WHO codes and descriptions instead of CM.
For the full note mode (Hierarchical BERT) however, the ICD-10-CM choice was driven by data availability. The model was trained on MIMIC-IV (therefore the ICD-10-CM target labels), and finding a comparable large-scale dataset of clinical notes mapped to ICD-10-WHO codes has proven difficult.
We are still very interested in refining our approach: if you know of any datasets or resources where notes are mapped to WHO ICD-10, we would be eager to explore them!