Please login to view this media

  • Talk
  • 30/09/2024
  • USA

CODEX AI - Clinical Orthopaedic Data Extraction

Description

Dr Cachaito presents a project called Codex AI, which aims to utilize deep learning language models to enhance the efficiency of billing and coding within medical settings, particularly in orthopedic surgeries. The discussion begins with the acknowledgment of the team involved in the initiative and presents the challenges faced by healthcare providers regarding Electronic Health Record (EHR) documentation, which is often complicated and time-consuming.



Cachaito explains that Codex AI seeks to automate and standardize documentation practices to ensure compliance with billing requirements and reduce errors in data capture for research. The model will be particularly focused on solving issues related to accurate documentation of surgical procedures, like total hip arthroplasty, using a dataset of verified operative notes.



The presentation outlines a roadmap for developing the proof of concept over the next year, with an ultimate goal of expanding its application across various medical specialties. Additionally, the transcript touches upon the competitive landscape, raising concerns about larger organizations potentially overshadowing their efforts, but highlighting Codex AI's unique approach and data collection methods as a key differentiator. The anticipated benefits of using this AI-driven solution include reduction in billing errors, improved compliance for insurance claims, and enhanced data integrity for clinical research.



Towards the end, the speaker answers questions regarding data integrity and the model's ability to accurately reflect surgical outcomes, emphasizing that the AI will be trained on high-fidelity, validated data to ensure both accuracy and compliance in coding.