DeepScore Methodologies for AI-Generated Clinical Documentation

This technical report presents DeepScribe's methodologies for measuring the quality of AI-generated clinical documentation, focusing on the composite metric known as 'DeepScore'. The report outlines various metrics used to evaluate documentation quality, including Major Defect-Free Rate (MDFR), Critical Defect-Free Rate (CDFR), Captured Entity Rate (CER), Accurate Entity Rate (AER), Minimally-Edited Note Rate (MNR), and Medical Word Hit Rate (MWHR). Each metric is defined and its significance in assessing the accuracy and completeness of medical notes is explained. The report emphasizes the importance of these metrics in guiding continuous improvement in patient care documentation. Additionally, it details the auditing methodology employed to ensure that AI-generated notes meet high standards of medical documentation. The findings indicate that the DeepScore, derived from these metrics, serves as an overall index of quality and accuracy, crucial for enhancing the effectiveness of AI solutions in clinical settings.