Augmedix Derives More Value from Automated Generation of Medical Notes

According to Manny Krakaris, CEO of Augmedix, this company has led the field of medical notes over the past 11 years from human scribes to full automation. To heighten accuracy, they use a combination of the most modern technologies. The transcript of the conversation is fed into a typical large language model (LLM) used by generative AI. But Augmedix has also analyzed more than 500 medical complaints brought by patients to create other data models. They apply both the LLM and the structured data model from their analysis to each transcript in their Augmedix Go service. The use of both modles produces more accuracy and eliminates most oddities (known as “hallucinations” in the AI field). If the two models disagree on an interpretation. the user interface flags the conflict so the doctor can check them. Augmedix Go also compares the note it generated to the edited version saved by the doctor, thus enabling continuous learning. Augmedix Go goes beyond the note to extract the various topics and create a separate organized display of metadata, such as medications and family history. Through this interface, a doctor can ask questions with a high chance of getting accurate answers. As Krakaris says, “Automation is table stakes. You need it just to get in the game.” They are constantly working on new ways to “get the most” out of the doctor’s note. They not only take in information, but feed ideas to doctors in real time, such as suggesting ...
Source: EMR and HIPAA - Category: Information Technology Authors: Tags: AI/Machine Learning Ambulatory C-Suite Leadership Clinical EMR-EHR Health IT Company Healthcare IT Hospital - Health System Ambient Clinical Voice Augmedix Augmedix Go Healthcare AI Healthcare IT Video Interviews Healthcare LLMs Source Type: blogs