3 Tips for Successfully Advancing AI Tools in Clinical Care

In the latest webinar of HealthLeaders’ The Winning Edge series, a two-member panel of experts shared their experience in ensuring successful adoption and implementation of AI tools in clinical care.

Health systems and hospitals are facing a dizzying array of AI tools that impact clinical care, including AI scribes and AI-powered clinical decision support tools. To manage adoption of these AI tools, chief information officers, chief medical information officers, and other senior leaders must make sure their organizations have several mechanisms in place such as AI governance, risk mitigation, and clinician support for new AI solutions.

This week’s webinar, which was titled The Winning Edge for Clinical AI Advancement, featured two experts with extensive knowledge of AI tool adoption in clinical care: James Blum, MD, chief health information officer at University of Iowa Health Care, and Linda Stevenson, MBA, chief digital information officer at Fisher-Titus Medical Center. The topics they tackled included assessing the clinical and financial ROI of AI tools, garnering clinician support for AI tools, and AI governance.

Assessing clinical and financial ROI

Assessing the clinical and financial ROI of an AI tool starts with engaging clinician stakeholders early in the adoption process, one of the panelists said. This engagement effort should include a focus on a particular problem that the new AI tool is designed to solve, the panelists said, adding that focusing on an AI tool’s ability to solve a problem informs its impact on clinical care and whether investing in an AI tool can be justified financially.

The clinical and financial ROI of AI tools in clinical care varies depending on what an AI tool is designed to do, the panelists said.

For some AI tools, the financial ROI is relatively easy to assess, one of the panelists said. For example, AI tools that improve coding lead to better revenue capture, which boosts the bottom line financially.

Other AI tools such as AI scribes have a softer ROI, Blum said.

As adoption of AI scribes has become widespread, it is imperative for health systems to have AI scribes to remain competitive with other health systems from a workforce recruitment and retention perspective, Blum said. University of Iowa Health Care decided to be an early adopter of AI scribes to help recruit and retain medical faculty as well as take better care of its clinicians, he said.

Fisher-Titus Medical Center and its ambulatory clinics serve many rural communities, and the organization has open positions that it struggles to fill. The health system has been able to address this workforce pressure by adopting AI tools that automate functions and ease the workload on clinical staff, Stevenson said.

Garnering clinician support for AI tools

Clinician support is crucial in the adoption of AI tools in clinical care, the panelists said.

Just as in the case of assessing the ROI of an AI tool in clinical care, involving clinicians from the beginning of the adoption process to focus on a particular problem that needs to be solved is pivotal in gaining clinician support for a new AI tool, the panelists said.

This approach to AI tool adoption is in stark contrast to the adoption of other IT tools such as electronic medical records, which were largely forced on clinicians, Blum said. CIOs, CMIOs, and other senior leaders should not tell clinicians that they must use an AI tool; instead, they should identify gaps where AI tools can help clinicians and tell clinicians that AI tools can make their professional life better, he said.

Executive support for AI tools can help drive clinician adoption of AI solutions, Stevenson said. Executives such as IT leaders, CMOs, CNOs, and CMIOs should champion an AI tool and share the “why” behind an AI tool’s adoption, she said.

Senior leaders should also address fear that AI tools are going to replace clinical staff members, the panelists said, noting that there is more than enough work for staff members to do and AI tools have not advanced to the point where they can displace clinical staff.

AI tool governance

Rather than establishing separate processes for AI tool governance, University of Iowa Health Care and Fisher-Titus Medical Center have embraced a streamlined approach to AI tool governance.

Blum and Stevenson said their organizations have largely treated AI solutions like any other IT tools.

For example, University of Iowa Health Care and Fisher-Titus Medical Center use the same acquisition process, cybersecurity reviews, and performance evaluation efforts for AI tools as they use for other IT tools.

The one exception to this streamlined approach at University of Iowa Health Care has been the creation of an AI oversight group, which includes Blum, a geneticist to assess the use of genomic data, and the health system’s director of application services. Members of the AI oversight group have training on AI solutions, so they can assess the performance characteristics of AI tools and ensure that AI tools are appropriate for the health system’s patient population, Blum said.

University of Iowa Health Care and Fisher-Titus Medical Center conduct continuous governance of AI tools, Blum and Stevenson said. These efforts include ensuring that AI tools are generating the ROI that was expected, they said.

In addition, Blum and Stevenson said that whenever there is renewal of an AI tool contract with a vendor, it is a golden governance opportunity to ensure that the tool is performing well and addressing the problem it was intended to solve.