Minset mCoder Achieves Top Performance on MDACE Benchmark

Minset announced in a press release that its autonomous medical coding system, mCoder, achieved state-of-the-art results on MDACE, a leading public benchmark for explainable automated medical coding. MDACE was introduced in 2023 by 3M Health Information Systems and Carnegie Mellon University researchers and is used to evaluate coding accuracy and evidence grounding in clinical documentation.

mCoder surpassed all known prior peer-reviewed results in both the 1K-constrained and full-label MDACE settings. The 1K-constrained evaluation tests performance across roughly 1,000 ICD codes, while the full-label setting requires selection from over 70,000 ICD-10 codes, representing a more complex and realistic clinical task.

The system demonstrated the ability to assign medical codes accurately while linking each code to supporting evidence in patient records. This functionality is essential for transparency and auditability in healthcare revenue cycle operations. Minset is currently working with select partners to evaluate mCoder in production environments.

mCoder is part of Minset’s broader revenue cycle platform, which also includes mDenials for automated denials management and m360 for patient engagement. These systems share a unified reasoning framework designed to improve coding accuracy and operational efficiency across healthcare workflows.