Voice and language processing solutions company Nuance Communications Inc.  announced the results of a product study today that shows how much healthcare professionals can benefit from artificial intelligence.

The company revealed the study during the Healthcare Information and Management Systems Society 2017 conference as part of a presentation about the company’s use of AI. According to the report, clinicians can save up to 45 percent on documentation time and can improve quality scores up to 36 percent with AI-enabled products.

Nuance, which produces the Dragon Medical One cloud-based clinical speech platform, says its research shows how much AI and deep machine learning can be used to aid in reducing administrative burden for healthcare workers.

To stay on top of their jobs, clinicians and healthcare professionals produce a great deal of documentation, and much of that done by doctors is dictated to save time typing or writing. The software used to turn spoken language into documents requires sophisticated speech recognition. Machine learning systems are used to reduce errors when transcribing speech.

Nuance’s research revealed that the company’s use of AI increased the company’s core speech recognition capabilities with a 30 percent drop in errors, meaning that clinicians didn’t need to spend as much time proofreading the end document.

Using AI for real-time advice

The company also produces technology that can advise doctors based on patterns discovered in patient medical data. Using clinical data from the cloud, Nuance’s computer-assisted physician documentation system can proactively support and allow doctors to ask the right questions and increase understanding of care.

Healthcare solutions that use machine learning are not entirely new. IBM Corp. has been developing its AI system Watson to assist doctors for a few years now.

“I see the technology being our co-pilot,” said David Y. Ting, a doctor at the Massachusetts General Physicians Organization. “In a world now where flying the airplane is way too complex for any one human being to manage on their own, the computer should be my co-pilot to keep me alert, keep me apprised of the situation, warn me when there are dangers coming down the pike and then execute on orders when that needs to happen.”

As a case study of Nuance’s AI use, the company cited impressive results from an integration with a large integrated delivery network. Use of machine learning to provide timely advice led to improved documentation for patients with extreme illnesses by 36 percent and improved details on patients with high risk of death by 24 percent.

Voice-enabled assistants in healthcare

An AI system can be used to better understand doctors dictating documentation and provide timely advice by providing insights, but the hopeful breakthrough of AI is to have a computer that can listen to questions and trigger actions like an expert assistant.

Nuance’s virtual assistant Florence, named after the famous English statistician and founder of modern nursing, Florence Nightingale, can do exactly that.

The Florence virtual assistant can listen to and understand human speech and provide assistance for repetitive activities such as orders for medication, labs or imaging. Ag an AI system, it also learns over time to optimize interactions with doctors by anticipating needs to provide quicker resolutions.

According to Nuance, Florence saved 35 percent of doctors’ time during the initial pilot and now saves over 50 percent compared with before it was integrated. The system also reduced the total number of keystrokes from 87 clicks to zero in just 20 orders.

Nuance says that on a national scale this would amount to 22.6 million hours saved for physicians in a year based on pharmacy, radiology and lab orders in the U.S. alone.

Image: Attribution Engine, source Negative Space. Licensed under CC0