From enhancing efficiency to helping organizations strengthen essential face-to-face human interactions, artificial intelligence holds significant promise for healthcare. But strong regulation and clear definitions of the technology's use and capabilities will be required.

WHY IT MATTERS
Those were among the conclusions drawn from a roundtable discussion, featuring two dozen leading experts on healthcare AI, hosted this summer by Epic and UW Health, the integrated health system of the University of Wisconsin.

AI technology's ability to enhance access to care, expedite healthcare delivery and address disparities in underserved communities were key areas of focus.

Participants, which included representatives from the federal government, healthcare providers, insurers and academics, as well as tech leaders, touted AI’s potential to reduce administrative burdens.

However, successful deployment and adoption will require close collaboration among healthcare institutions to train workers and integrate AI into their workflows, with a priority on patient-centered solutions.

Specific use cases discussed include real-time patient monitoring and automated data entry, highlighting AI’s ability to alleviate stressors on the healthcare workforce.

Among the major concerns was the need for transparency as AI integrates into patient care, and the importance healthcare organizations must place on building trust in AI technology.

This can be achieved by informing patients when AI is used and ensuring clarity on data use by external providers.

Participants agreed that AI should support, not replace, healthcare staff, emphasizing the importance of involving employees in testing and trialing new AI tools.

The discussion also highlighted the importance of equitable AI adoption across rural and underserved areas, if AI is to be truly transformative. Infrastructure development and technical assistance were seen as crucial to ensure all patients benefit from AI.

Applications for AI were also discussed in aiding in real-time patient-condition monitoring, generating clinical-deterioration alerts, and providing unified notifications about patient status, thus facilitating prompt care interventions.

Predictive analytics, such as forecasting risks like geriatric falls, could enable preemptive measures and timely referrals for specialty care.

When it comes to enhancing efficiency, participants concluded that deeper integration of AI into clinical workflows promises to enhance efficiency, reduce manual workloads and offer data-driven insights for better treatment plans and patient outcomes.

Concerns about data equity and transparency were also discussed, with a focus on the inclusion of diverse patient populations in AI models.

Effective regulation of AI in healthcare requires appropriate authority and a thorough understanding of AI’s impact on patient care, with regulatory frameworks aimed at encouraging innovation, ensuring equitable access and addressing disparities in quality of care.

"Participants stressed the importance of balancing AI's benefits with ethical considerations, patient engagement, and robust regulatory frameworks," the report noted.

THE LARGER TREND
As AI expands in healthcare, there is an ongoing discussion about balancing its potential for improved health outcomes with concerns over labor, ethics and equity, necessitating careful regulation to prevent harm and build trust.

From AI-enabled computer vision and generative AI (genAI) workflow tools for nurses, to burnout-reduction tools and patient throughput algorithms, the applications for AI in healthcare are broad and expanding at a rapid pace.

The use of AI can also be extended to virtual reality (VR) and augmented reality (AR) applications also gaining traction in the healthcare space.

ON THE RECORD
"For AI efforts to succeed, it is crucial that policymakers and healthcare providers collaborate to ensure patient and workforce dynamics receive appropriate consideration," said Epic and UW Health leaders in the report, noting that the roundtable exploration was "unique in that unlike many other discussions that have focused on the technological aspects of AI, we examined AI through the lens of healthcare delivery and the policy mechanisms needed to drive equity and access.

"By bringing together healthcare providers, payers, policy experts and other key stakeholders," they added, "the roundtable identified best practices for integrating AI into the care process, reviewed current and pending policies, and developed consensus recommendations to establish essential guardrails for AI implementation in healthcare."

Nathan Eddy is a healthcare and technology freelancer based in Berlin.
Email the writer: nathaneddy@gmail.com
Twitter: @dropdeaded209

The HIMSS AI in Healthcare Forum is scheduled to take place September 5-6 in Boston. Learn more and register.