Hospital Systems Use Generative AI to Enhance Workflows

• The University of California Irvine is among the latest institutions to use artificial intelligence in a clinical setting.
• Preliminary results from Deloitte’s 2024 Health Care Outlook Survey found that roughly 25% of health care organizations “have a generative AI governance model“ and a team that maintains and updates those models.
• Some industry leaders are receptive to using AI, but understand current limitations and privacy concerns. 

The University of California Irvine is among the latest institutions adapting to the use of artificial intelligence in a clinical setting. 

Abridge, a “generative AI solution for clinical documentation,” became a part of the UCI Health system last month. Founded in 2018, the Abridge technology transfers conversations between patients and clinicians into “a structured, clinical note draft in real-time,” a March news release stated. 

“Abridge is already reducing the ‘pajama time’ that is endemic to clinicians today,” Scott Joslyn, UCI Health’s chief innovation officer, said in the release. “It’s a powerful AI technology that fits seamlessly into our existing workflows and is the beginning of what I believe will be a broad transformation using generative AI to simplify the experience and elevate the quality of health care.” 

UCI Health’s use of generative AI is part of a larger move toward the technology across the nursing industry. Preliminary results from Deloitte’s 2024 Health Care Outlook Survey found that roughly 25% of health care organizations “have a generative AI governance model” and a team that maintains and updates those models. 

Executives who participated in Deloitte’s survey said they believe generative AI could address “many of the sector’s most vexing issues,” such as patient wait times and staff burnout. 

“Dr. Bill Fera, a principal at Deloitte Consulting LLP, recently urged health care leaders to incorporate generative AI in parallel with other digital technologies as they retire legacy hardware and transition to the digital realm,” the health outlook survey stated. “The technology is already being used to predict in-hospital mortality, length of stay, and medical claims denials. At the same time, many consumers are already turning to generative AI to help make decisions about their health and well-being, to locate clinicians, learn about medical conditions, understand treatment options, and decipher technical language.” 

The Bigger Picture with Generative AI

Some health and technology leaders at the World Economic Forum’s annual meeting showed support for AI, but with the caveat that industry professionals take “considered steps” to limit risks while simultaneously finding its benefits, according to a January 2024 report. 

“We truly believe AI will revolutionize health care delivery, generate groundbreaking advances in health research, and contribute to better health for all,” the report stated. “But we must get it right.” 

The World Economic Forum report also noted that AI can perpetuate biases that already exist in health care. 

A study in the February 2024 edition of “Nurse Education in Practice” found that while ChatGPT, for example, resulted in “appropriate nursing diagnoses,” it did not adhere to the requested North American Nursing Diagnosis Association—International classification. 

“It should be noted that some level of human fact-checking will always be needed when using generative artificial intelligence methods,” the authors concluded. “How to implement this step and include a human-in-the-loop approach to optimize the potential time saved using generative artificial intelligence represents one of the most important challenges in the future of modern data-driven nursing care.” 

Navigating Generative AI

Meanwhile, a new consortium of health care leaders, the Trustworthy and Responsible AI Network, will “operationalize responsible AI principles” in the industry. The network was announced in March at the HIMSS 2024 Global Health Conference and features more than a dozen organizations, including Microsoft as the “technology enabling partner.”  

Its goal is to set best practices, provide a way to measure the success of AI implementation, and to help develop a “federated national AI outcomes registry” to help health care professionals nationwide. 

“AI has incredible potential to transform health care; however, it’s crucial that we harness this power ethically, responsibly, and with safety at the forefront,” John Brownstein, Ph.D., chief innovation officer at Boston Children’s Hospital, said in a news release. “As we continue to face complex and multifaceted challenges in health care, it’s imperative that we come together as a community to navigate this new frontier and ensure that the benefits of AI are implemented equitably among all patients and providers.” 

That navigation occurs not only at the health system level but also in education. Jung In Park, an associate professor at UC Irvine’s Sue and Bill Gross School of Nursing, shares her experience working with large data sets and machine learning with students. 

“I believe nursing can benefit from this era of big data science,” she said in a Q&A on the UCI website, “as there is a growing need for new discoveries from large quantities of health data to provide evidence-based care.” 

A UCI nursing student helped create KidoCare, an “AI-driven platform that improves the experience of hospitalized children.” Nursing science graduate student Mahkameh Rasouli, along with five other students, developed the program to engage in conversation with children and tailor entertainment options to the individual. 

The project garnered second place in the university’s ANTrepreneur Center AI Innovation Challenge.  

The Bottom Line

Generative AI is a part of nursing, whether practitioners embrace or discourage its use. Universities are setting the stage for up-and-coming nurses to have at least a base-level understanding of the technology as they move forward in their health care careers. Some industry leaders are receptive to its use, but understand current limitations and privacy concerns, which they hope to address in a code of ethics.

Read More