Ambience Healthcare revealed on Tuesday that its new artificial intelligence tool for medical coding delivers results 27 percent beyond physician accuracy. The company developed this advanced model through reinforcement tuning with technology provided by OpenAI.
Ambience is gaining attention in the burgeoning landscape of artificial intelligence that assists in drafting clinical records while doctors speak with their patients. The system is geared to ease some of the pressure on physicians by automatically recording and organizing medical data.
In medical environments, coding is crucial because it ties diagnoses and treatments to billing and reporting. The newest solution from Ambience can analyze patient interactions and identify the correct disease or condition codes, known as ICD 10, from a list that numbers nearly 70000 entries. These codes are fundamental for tracking results, comparing clinical outcomes, and ensuring proper payment.
Ambience Model Tested Against Physicians
To benchmark the model, Ambience evaluated its accuracy against a panel of expert clinicians who had independently agreed on the right coding for several complex medical scenarios. Eighteen physicians participated in a head to head comparison, which found that the AI outmatched the average doctor by 27 percent in coding precision.
While the technology raises questions about the future roles of human professionals, Ambience maintains that their goal is not to replace anyone. Brendan Fortuner from the engineering team at Ambience said their focus is on relieving the administrative strain from health workers, reducing mistakes, and ultimately improving the quality and efficiency of health care.
Traditionally, assigning the proper ICD 10 codes has taken up much of clinicians’ time and effort, but this task is essential for comparing health systems and measuring care outcomes, said Will Morris, Ambience’s chief medical officer. Morris described accurate coding as the foundation of healthcare quality assessment.
Ambience’s system is already deployed at prominent organizations such as UCSF Health and Cleveland Clinic. The company has attracted more than one hundred million dollars in funding from investors including Andreessen Horowitz, Kleiner Perkins and the OpenAI Startup Fund.
Market speculation is swirling that Ambience is preparing a new investment round that could value the business above one billion dollars, though company representatives declined to address rumors of ongoing fundraising.
In addition to ICD 10 codes, AI-assisted CDI and compliance at the point of care offers similar solutions for other coding types like CPT codes and is researching tools for processes like prior authorization and clinical trial enrollment. Its latest medical coding model is expected to become available to customers before fall.
Morris emphasized that capturing codes accurately during each patient encounter could fundamentally change the way clinicians work and how care is delivered. John Muir Health Announces the Full Enterprise Rollout of Ambience Healthcare’s AI Platform.
