With new artificial intelligence (AI) products emerging (seemingly) every day—generative AI is increasingly becoming part of the healthcare process.
But what does it actually look like to implement an AI program into a health system?
University of Iowa (UI) Health Care in Iowa City has teamed up with Nabla. This Boston-based company develops ambient AI assistant technology, to roll out the company’s new technology that transcribes appointments and reduces clinicians’ time spent on documentation, according to Nabla. UI Health Care rolled out the tech to all its clinicians in September and provided video training for the staff on how to use Nabla. James Blum, an anesthesiologist and chief health information officer at UI Health Care, said.
In 2024, about 62% of clinicians reported that “excessive documentation requirements” is a leading cause of burnout, according to Athenahealth, a health tech and electronic health record (EHR) company. The American Medical Association reported in January that primary care physicians, for example, can spend up to 45.7 minutes on documentation for every 30-minute appointment.
In March and April, UI Health Care ran a five-week pilot program in which a cohort of clinicians used the autonomous scribe to document 2,300 patient visits. Providers reported a 4.3 out of 5 clinical note rating, and saw a 26% reduction in reported burnout rate, according to a press release.
About 78% of office-based physicians and 96% of acute care hospitals had an EHR system in 2021, according to HealthIT.gov, a federal organization that researches health technology and information. By comparison, only 34% of office-based physicians and 28% of hospitals were using EHRs in 2011.
“We, unfortunately, over the last couple of decades, have introduced more and more burden on our providers to document various things,” Blum said. “A technology like ambient [AI] fundamentally changes the game, as it removes that documentation burden, and it removes the cognitive burden when you’re sitting with the patient…If you know that note’s going to be generated, and it’s going to be a high-quality note that you can then edit, that enables you to engage with the patient more readily.”
Nabla also built in a customization feature called Magic Edit, which allows clinicians to make notes more concise or add details about a visit.
“Customization is really our core differentiator regarding the product. We want to customize the product to specialties, but also each physician,” Groll said.
Aside from Nabla, Blum said that the system also uses Evidently, a chart-summarizing cognitive AI tool, alongside other predictive technology to detect conditions and improve early intervention care.
Choosing tech. The built-in privacy measures were a major reason UI Health Care decided to work with Nabla, according to Blum. In addition to not recording the conversations, the company says Nabla automatically deletes UI Health Care transcripts from its system after 14 days, and it does not train its models on information the technology engages with from the health system.
“We take regulation, safety, data privacy, [and] transparency very seriously at Nabla,” Delphine Groll, co-founder and COO of the tech company, told Healthcare Brew. “This is very important for us to have this kind of policy when we partner with health systems, because for them, it’s also a matter of trust.”
Nabla is fully supported in English, Spanish, and French—with 32 more languages available in beta versions—and covers at least 55 specialities, like cardiology, according to the company. In addition to UI Health Care, about 70 health organizations and 30,000 clinicians have deployed Nabla, Groll said, including the Children’s Hospital Los Angeles and Carle Health in Illinois.
About one month since the roll-out, nearly 1,000 providers at UI Health Care now use Nabla, according to Blum.
Zoom out. Nabla is not alone in the ambient listening tech space and competes with AI-assisted healthcare documentation products from Abridge, Insight Health, and Nuance Communications. Ambient AI, according to a 2023 report from the Mayo Clinic, has “great potential to become mainstream over the next 10–15 years.”
No comments:
Post a Comment