Key Takeaways
- An AI startup, Ambience Healthcare, has developed a new medical coding AI.
- This AI identifies medical codes 27% more accurately than physicians.
- It aims to cut administrative work and billing mistakes in healthcare.
- Made with OpenAI tech, it’s used by major health providers.
Artificial intelligence company Ambience Healthcare has introduced a new AI model designed for medical coding, reporting a 27% better performance in accuracy compared to doctors.
This AI transcribes and analyzes doctor-patient conversations (with consent) in real time to identify correct medical codes. These codes, known as ICD-10, are a standardized list of about 70,000 classifications for diseases and conditions, vital for billing and tracking health trends.
The goal is to help cut down on billing mistakes and make the work of doctors and professional coders more efficient. This could significantly reduce the paperwork that often bogs down healthcare professionals.
“We’re not replacing doctors or coders,” Brendan Fortuner, Ambience’s head of engineering, told CNBC. “What we’re doing is we’re liberating them from administration, and we’re fixing mistakes that help make health care better, safer, more cost-effective.”
Accurately documenting these medical codes is a detailed and time-consuming job. Dr. Will Morris, Ambience’s chief medical officer, emphasized its importance, stating it’s how health systems compare performance and track quality. “It’s the cornerstone for quality,” Morris said.
Ambience’s technology is already being used in over 40 healthcare organizations, including well-known names like Cleveland Clinic and UCSF Health. The company has successfully raised over $100 million in funding from prominent investors.
The new AI model was developed using advanced training methods from OpenAI, tailoring it for the complexities of healthcare. Ambience rigorously tested the model to validate its accuracy.
This involved comparing the AI’s coding against answers agreed upon by expert clinicians, and then against the performance of 18 board-certified doctors. The AI proved significantly more precise.
Fortuner highlighted the significance, noting, “It shows for the first time that an AI system can actually surpass clinician experts at a very, very important administrative task, especially in coding.” Ambience already offers similar tools for other medical codes.
The company is also exploring AI solutions for other healthcare administrative tasks like prior authorizations. This new ICD-10 coding feature is expected to be available to its customers this summer.
Dr. Morris believes that improving accuracy at the point of care is a “fundamental change” for healthcare, potentially leading to better data and more efficient operations.
