Integrating AI Into Clinical Care: It's All About Efficiency

At the HealthLeaders AI in Clinical Care Mastermind forum this week, execs from 14 health systems discussed how the technology will – eventually – improve patient care.

In the simplest terms, AI is a complex technology that aims to make healthcare simpler.

For healthcare executives, however, the devil's in the details. While AI promises to ease workflows for stressed clinicians and improve patient care, getting to that point takes a lot of time and patience. And a good understanding of what generative technology should and shouldn't be doing.

C-Suite executives from 14 health systems tackled these issues at HealthLeaders' AI in Clinical Care Mastermind forum Wednesday at the Stein Eriksen Lodge in Deer Valley, Utah, the opening event to the HealthLeaders CMO Exchange. The Mastermind program, now in its third year, brings together executives for three virtual round-tables and one in-person forum to discuss and share ideas with their peers on a key healthcare topic.

In this case, it's how AI is being introduced to care management. Early wins in this area have centered on ambient tools and bots, some designed to capture the interaction between doctor and patient (often in an ambulatory setting) and transcribe that data into the EHR, others taking on administrative tasks or even handling queries from patients.

James Blum, MD, CDH-E, Chief Health Information Officer for University of Iowa Health Care, detailed his health system's successful roll-out of an ambient AI tool for doctors, along with the ongoing deployment of chart mining technology designed to pull relevant data from the patient's medical record to help clinicians determine care pathways.

Blum noted the tools not only help doctors, but could also improve prior authorizations and even reduce denials. The key, he said, is developing AI products that address both clinical and financial – particularly revenue cycle management – pain points. That's because a tool that addresses doctor stress and burnout needs to have a more robust ROI to appeal to CFOs and even CEOs and be sustainable.

ROI is indeed a tricky component of AI. In many cases the technology is expensive and uses considerable processing power and data, straining even the largest health systems and out of reach for smaller and rural providers.  AI has to have value across departments to truly succeed.

The Struggle for Clinician Buy-In

At the same time, Ai has to integrate seamlessly into clinical pathways. Many an ambient AI project has failed because the tools require doctors or nurses to take a couple extra steps or adjust their workflows. For clinicians who have weathered EHR implementations, the idea of doing one more thing to add more technology to the process is a deal-breaker; they'll ignore the tool entirely or develop work-arounds.

Corey Cronrath, DO, MPH, MBA, FAAPL, FACOEM, CMO of the Mental Health Cooperative, noted his organization tried an ambient listening tool a few years ago and was forced to discontinue it because their doctors found it to complex and intrusive. He said it took almost three years to bring their doctors back on board.

Candace Robinson, MD, CMO of LCMC Health's Touro Hospital, pointed out that the perception of AI in healthcare is that it has to be perfect, and anything that's less than perfect won't be embraced. But executives noted that AI doesn't have to be perfect to have value, only that it can improve on what clinicians are doing now.

Beyond improving clinical workloads, executives attending the Mastermind program said the value of AI lies in giving them information they need to improve care. Tipu Puri, MD, PhD, CMO of the University of Chicago Medical Center; Roopa Foulger, Vice President of Digital Innovation and Development at OSF HealthCare; and Tom-meka Archinard, MD, MBA, FACEP, SVP and CMO at University of Maryland Capital Region Health, all talked about new tools that analyze data to give clinicians insights on specific medical concerns, such as sepsis or the onset of chronic disease. But those tools need to be carefully managed to ensure that the data they're using is accurate.

While the “human in the loop” strategy is designed to ensure continuous governance of AI before it affects patient care, all the executives pointed out that it's also important to ensure that a human being is still delivering care. Aside from the concerns that AI will replace doctors and nurses, there have been some concerns that clinicians could rely too much on AI to make their decisions for them.

Michael Fiorina, MD, CMO for the Independence Health System, questioned whether AI could have an impact on critical thinking, leading to a discussion on how AI should be included in medical education. There's no doubt that tomorrow's clinicians need to have a good understanding of AI as they enter the healthcare field, but the executives agreed that it's more important that they learn medicine before they learn AI so that they're using AI to augment care delivery rather than replace it.

How Will Patients Use AI?

That same question about relying on AI could be turned around to focus on patients. Some have worried that an AI-emboldened patient could overwhelm doctors, while others feel the technology will help patients become more knowledgeable about their health and be able to contribute more to the doctor-patient relationship.

That, in part, is why Blum said he expects health systems like his will be investing in AI for patient engagement, and why the patient portal could be the next big proving ground for new AI tools and platforms. That's why many health systems are trying out bots and other tools that use AI to communicate with patients.

It's a tricky field. Mark Kandrysawtz, MBA, SVP and Chief Innovation Officer at WellSpan Health, noted his health system's implementation of a bot called Anna, which helps steer patients to the right care provider, has actually evolved from being an efficiency tools to drive growth. But there have been hiccups, from the bot using humor at an inappropriate time to convincing someone that it was an actual human.

At the same time, with some studies showing that patients prefer to talk to AI bots because they think the bots are more empathetic, some are wondering whether AI should prompt the industry to take a closer look at the doctor-patient relationship and teach its doctors and nurses how to be better care providers.

At the end of the day, and at the end of this Mastermind program, the consensus is that AI will improve healthcare, and that the industry has to weather the rough spots, learn how to best use the technology, and make sure there will always be a human being making that final care decision.

As Thomas Balcezak, MD, MPH, EVP and Chief Clinical Officer for Yale New Haven Health, put it, AI will eventually become so accepted and commonplace that we'll forget how much trouble we had putting it in place.