The Meduit Innovation Lab Blog

Leveraging Artificial Intelligence to Increase Chart Coding Accuracy

Posted by Meduit RCM on Apr 1, 2022 8:00:00 AM

Documenting a patient visit is a critical part of the healthcare provider's revenue cycle, and recording charts accurately is essential to correct reimbursements. According to one study, up to 90% of medical bills contain errors of one kind or another. Those errors are cumulatively estimated to be worth $750 billion a year.1


Meduit has developed MedAutoCode, a solution that utilizes artificial intelligence (AI) and natural language understanding (NLU) to recreate a patient visit and assign accurate, real-time medical codes to the chart with zero human interaction. It’s a groundbreaking technology that eliminates manual processes to lower operational costs, improve coding accuracy and reduce A/R time.

How MedAutoCode Works

Charts are fed from the provider’s EMR to MedAutoCode, which automatically identifies and captures every clinical aspect of a chart. The system understands what is relevant as well as the most accurate code to use and assigns codes only for charts it fully understands.

MedAutoCode requires fewer than five seconds to code a patient chart and requires no human intervention. Unhandled charts are returned for human coding. The program has built-in quality control through auto-validation, as each chart goes through a multi-step linguistic, narrative and clinical validation to ensure that MedAutoCode fully understands the chart. Successfully coded and validated charts are complemented with an audit trail with full visibility into the reasons for each assigned medical code.

MedAutoCode processes up to 80% of ER codes in an automated manner. The solution’s chart accuracy is a 98% human-to-machine code match.

Watch our video to learn more about MedAutoCode:

Watch the Video