The healthcare industry is navigating the integration of AI with unique challenges and opportunities. Unlike other sectors, the stakes are exceptionally high, with successful AI implementation having the potential to improve patient outcomes and alleviate administrative burdens.
Becker's spoke to IT leaders at multiple hospitals and health systems to discuss how their organizations are preparing for AI integration and what AI initiatives they are most excited to explore.
CommonSpirit
At Chicago-based CommonSpirit, the organization's AI approach focuses on enhancement.
"We are not trying to replace care, diagnoses, or treatment by humans, rather we are implementing AI to improve the care we deliver," CIO Daniel Barchi told Becker's. "Long before the advent of AI, safety and quality have guided every decision at every step. As we evaluate every new tool, procedure, drug and protocol, we must balance risks."
CommonSpirit is laser focused on AI initiatives that identify clinical needs and offer the greatest clinical change.
The healthcare organization currently employs AI tools to identify issues as significant as aortic stenosis and as routine as an unscheduled annual exam.
"This is where AI has great potential both to advance care as well as offload burden from our clinicians and support teams," Mr. Barchi said.
Kaiser Permanente
At Oakland, Calif.-based Kaiser Permanente, the health system is focused on the responsible use of AI, and placing people at the heart of every decision.
"We only use AI tools that solve the problems and outcomes that matter to our patients, members, clinicians and care teams," Daniel Yang, MD, vice president of AI and emerging technologies for Kaiser Permanente told Becker's. "We thoughtfully evaluate the return on health of an AI tool — or, how it drives improved outcomes while simplifying healthcare complexities."
Integrating new technologies into existing workflows is a crucial yet challenging aspect of AI deployment. To meet this challenge, Kaiser uses advanced implementation of science practices to ensure that patient and clinician preferences are considered before AI tools are integrated.
"It allows us to consider patient and clinician preference before any AI tool or technology is integrated into the workflows of our care providers and across our health systems," Dr. Yang said.
When it comes to what most excites Kaiser about exploring the AI space, Dr. Yang says its ambient scribe technology, an AI tool that reduces the time physicians spend on administrative tasks and enables them to be more present with their patients.
"This is exciting because it's a perfect example of how we approach responsibly using AI to improve care, enhance the relationships between our patients and clinicians and alleviate the root causes of clinician burnout," he said.
UPMC
At Pittsburgh-based UPMC, the organization began its AI journey by establishing an AI policy and standard to ensure that employees were appropriately utilizing the technology.
"We are focused on getting our workforce engaged in how we can think and work differently using AI tools, both in the administrative and clinical spaces. We have established a team that is dedicated to empowering UPMC employees to leverage technology, enhancing how they work today, as well as moving toward new levels of productivity, creativity and collaboration," Ed McCallister, senior vice president and CIO of UPMC told Becker's. "Through our AI policy and employee education, we believe that we are well positioned to harness the tremendous potential that AI brings to healthcare."
Mr. McCallister said the UPMC team is fascinated by the potential for AI to transform healthcare as we know it. Particularly, what role AI will play in managing burnout, reducing administrative burdens and increasing productivity for clinicians.
"We want to give clinicians more bandwidth so they can focus their attention and energy on taking care of patients," he said.
For example, UPMC was an early adoption of Abridge's technology to transcribe patient interactions.
"We are finding immense value in Abridge helping to make physician documentation more efficient," Mr. McCallister said. "During my 30-plus years as an IT professional, I have witnessed many impactful transformations driven by technology — AI is a game changer."
Stanford Health Care
At Palo Alto, Calif.-based Stanford Health Care, a multi-pronged approach was taken to be AI-ready.
For example, in 2022, Stanford established a new data science team to develop technical expertise and a platform for data-intensive tasks, especially AI applications. The health system also implemented a responsible AI life-cycle governance strategy, integrating AI assessment and decision-making into its existing organizational processes.
"The most important piece of this is to have high level executive decision-making (via a Data Science Executive Committee, co-chaired by the CEO) that balances innovation with sustainability," Nigam Shah, PhD, chief data scientist at Stanford Health Care, told Becker's.
What Stanford is most excited about is the responsible use of AI across biomedical research, education and patient care.
"High on the list are meaningful strategies for evaluating AI — what works and what breaks — are crucial components for building trust and positive impact," Dr. Shah said. "We think about projects in terms of a few application areas, such as administrative support, patient engagement, clinical decision support and workforce efficiency. The goal for any project is to have a meaningful impact, that is sustainable long term."
Providence
Over the past five years, Renton, Wash.-based Providence focused on three strategic pillars: simplify, modernize and innovate. According to B.J. Moore, CIO of the organization, this groundwork, which included moving to a single Epic platform, adopting Oracle Cloud ERP and standardizing on Microsoft 365, has positioned the organization well for AI adoption.
"By simplifying and modernizing our technology stack, we created a robust foundation that allows us to effectively leverage AI capabilities," he told Becker's.
Providence's AI strategy operates in two main buckets — "easy button" solutions and "bespoke AI" solutions.
Mr. Moore said the "easy button" approach comes with Providence already standardizing its systems.
"We're now partnering with Epic, leveraging their latest software version and Hyperdrive platform. This allows us to seamlessly integrate their AI capabilities," he said. "We're currently evaluating around 50 of Epic's AI solutions, and if they prove effective, we can quickly deploy them within our environment. This streamlined process is why we call it the 'easy button.'"
The bespoke AI solutions are the AI tools Providence has developed that are tailored to its needs.
Looking ahead, Mr. Moore said he is particularly enthusiastic about future AI initiatives that will further enhance patient care and reduce administrative burdens.
"We are just at the tip of the iceberg," Mr. Moore said. "Our goal is to use AI to take away the administrative burdens from our clinicians and allow them to focus on what they do best — providing care."
AdventHealth
At Altamonte Springs, Fla.-based AdventHealth, its AI advisory board provides the framework for how the health system vets, tests and implements AI into its operations.
"This enables our rigorous and principled approach to AI adoption and development within the organization as it gathers a cross-functional team of experts including doctors, IT experts, data scientists and vendor partners," Rob Purinton, vice president of analytics and performance at AdventHealth told Becker's.
Additionally, AdventHealth has invested significant time and effort into implementing a single electronic health record across all sites and care settings as clean, consistent data is key to effective use of AI technologies.
When it comes to AI excitement, Mr. Purinton said the health system is most excited about initiatives that transform the experience of its patients and caregivers.
For example, Mr. Purinton referred to ambient documentation assistants, which allow physicians to focus on patients rather than the computer, while the "documentation mostly takes care of itself."
"This can improve the attention patients receive and their experience in settings like their primary care physician's office," he said.
Mr. Purinton also said that AI works on top of rich new data such as digital pathology studies. This, in turn, can allow earlier and more accurate identification of early-stage cancers, which can be treated more easily than tumors identified later.
Mount Sinai
At New York City-based Mount Sinai Health System the organization has established a comprehensive governance structure to ensure its AI initiatives align with its organizational priorities.
"Part of AI governance is measuring how our AI products are performing in order to assure safety, equity and financial sustainability," Bruce Darrow, MD, PhD, interim chief digital and information officer of Mount Sinai told Becker's.
Additionally, Mount Sinai has invested in its cloud capabilities and internal talent, in order to provide its teams with the tools they need to drive AI advancements.
When it comes to the most exciting part of AI Mount Sinai is looking to explore, Dr. Darrow mentioned the potential of generative AI to unlock a new set of use-cases.
"Some promising opportunities include ambient listening to relieve documentation burden for the clinical team and our AI assistant for patients which can handle an increasing number of use-cases," he said.