To Integrate AI Into Clinical Care, Focus on Change Management

“We have the imperative to swim upstream, into undergraduate medical education, in nursing schools and medical schools, and insert some of this foundational education,” he says.

Sheahan says MedStar Health is pursuing an AI strategy that recognizes how the technology will change healthcare delivery, which includes the effect on the provider as well as the patient. As such, the ROI of a new tool or program looks beyond financial costs to include clinical outcomes and provider wellbeing.

That means getting buy-in from all parties, especially clinicians. Sheahan says innovation leads like to participate in town halls with system physicians to introduce new tools like ambient AI, giving them a chance to ask questions and try out the technology before it’s added to their toolkit. The idea, he says, is to include clinicians in the planning so that they’re invested in the process from the start.

“This isn’t forced change,” he points out.

Instead, he says, it’s organizational change. That’s what comes with a holistic technology that is being embraced by consumers as well as clinicians. Organizations that aren’t already properly pursuing this in-demand innovation may otherwise be forced to react rather than act with the times.

With national issues like workforce shortages, cost concerns and increased competition for the patient/consumer plaguing the industry, healthcare leaders are challenged to hit the ground running with AI. Sheahan says he, like any other innovation executive, is excited about the potential for AI to improve clinical care, but they still need to plan out the process and understand the outcomes.

While some health systems have created AI committees or even C-suite roles to manage AI, Sheahan says separating AI from other innovative strategies can hinder progress. He’s more supportive of integrating AI governance into existing management structures. In many cases, he says, a new AI tool can have multiple benefits across different departments—like an ambient listening tool that not only improves clinician note-taking but helps the rev cycle department improve coding.

“AI as a technology shouldn’t be governed entirely differently than other technologies,” he says.

As for the future of AI in the clinical space, Sheahan envisions an AI-enabled operating system serving the enterprise, responding to queries from pretty much anyone within the health system. The platform might help doctors map the best care plan for patients, give nurses direction on inpatient care, help the rev cycle management team deal with a prior authorization or denial, or even map out the best route for a patient to a specialist appointment across town.

“We want to show the many groups that can benefit from AI why and how to use it to make themselves more efficient,” Sheahan said. “Ultimately, that is what is going to deliver ROI over time. We can make our business more efficient if we’re all in this together.”