With the recent boom of artificial intelligence in healthcare, leaders are grappling with the best use cases.
Many systems are piloting AI programs to reduce clinician workloads and improve patient experience. For example, Palo Alto, Calif.-based Stanford Health Care is using an AI tool to help inform patients of test results.
"I personally would not want to have my healthcare, in some specialties, without AI because I firmly believe I will get a better outcome," Gianrico Farrugia, MD, president and CEO of the Rochester-based Mayo Clinic, said at the World Economic Forum's annual meeting in Davos, Switzerland, Jan. 22.
Leaders are grappling with which of the many new tools to implement within their system — and some are still looking ahead to what they wish existed. Here, four leaders share the AI use cases they want to see in healthcare:
Editor's note: Responses have been lightly edited for clarity and length.
Crystal Arthur, MD. Chief Medical Director of Emergency Services for McLaren Health Care (Grand Blanc, Mich.): In a perfect world, I envision an AI overlay for our electronic medical records where critical patient information automatically populates for physicians as soon as they see a patient. Imagine having a patient’s last echo, cardiologist information, and key labs immediately available at the point of care. We’re not quite there yet, but it’s something I’d love to explore.
Eric Smith. Senior Vice President and Chief Digital Officer at Memorial Hermann Health System (Houston): AI has been applied to several different areas of opportunity, both in support of the clinician as well as the patient. However, one of its challenges is that it is typically applied in a very siloed approach, not necessarily taking into account the holistic experience. That’s why a powerful use case for AI would be a comprehensive health companion, a resource that supports the patient throughout their care journey while ensuring the provider acts as the quarterback of the individual’s care. In this use case, AI would understand the specifics of the patient — current health, history, family history, etc. — and account for the care plan as outlined by their provider. AI would then use that information, leveraging learned methods about how the individual prefers to interact (text, voice, etc.), to prompt the patient to take action or congratulate them on key milestones, such as weight loss or blood pressure reduction. Progress, activities and data would be shared continually with the patient’s provider, with adjustments made to the care plan dynamically. This kind of capability would allow for a higher level of engagement on the part of the patient in managing their health, while also scaling the reach of the provider and ultimately resulting in better outcomes.
Varadarajan Subbiah, MD. Chief Clinical Effectiveness Officer at ChristianaCare (Newark, Del.): If I could design an AI solution, it would focus on patient safety—particularly in preventing pressure injuries and line infections. While we have scoring systems to assess risk, AI could continuously monitor patients and flag risks in real time, something humans simply can’t do 24/7. Having AI predict and alert staff about high-risk patients could drastically improve prevention efforts. Staffing someone to monitor these risks around the clock isn’t feasible, but AI could handle this efficiently and cost-effectively.
Gerald Wilmink, PhD. Director of Innovations Business Development and Licensing at Cleveland Clinic: One of the biggest challenges right now is moving AI tools from research into clinical practice. Many predictive algorithms are developed and published but don’t make it to market due to regulatory hurdles like FDA approval. We need to streamline this process so that effective tools can be bundled with others and integrated into clinical workflows. For example, predictive tools could detect life-threatening conditions like aortic dissection earlier. However, rural hospitals without specialists might not have access to these tools. Getting AI into those settings could be life-saving. Our goal is to accelerate this transition from research to real-world application, making these tools accessible to all healthcare providers.