Are Healthcare Leaders Getting Their AI Priorities Straight?
By Eric Wicklund
Health system and hospital executives looking to embrace AI will need to think long and hard about how they’ll measure ROI. That may include using the technology to actually replace care providers.
Lee Schwamm, MD, senior vice president, chief digital health officer, and associate dean of digital strategy and transformation for the Yale New Haven Health System and Yale School of Medicine, told a busy audience at last week’s HIMSS AI in Healthcare Forum in Boston that the technology will have a profound impact on healthcare delivery. The challenge, he said, will lie in understanding that impact before it happens.
“We’re going to need better financial models to really understand the ROI,” he said, noting that healthcare organizations have so far found only three or four successful use cases for the technology.
Schwamm says the healthcare industry has become “accustomed” and “complacent” in healthcare IT, and Ai is presenting healthcare leaders with issues they haven’t encountered before. The AI evolution, he pointed out, is similar to the development of the software-as-a-service (SaaS) model, but health systems and hospitals haven’t developed the governance to regulate these tools before they’re used.
“I’ve had e-mails where [doctors] say, ‘I just really enjoy using ChatGPT while in clinic,” she said.
The challenge for hospital leaders like Schwamm and Cunningham is to get ahead of a technology that’s moving faster than anything they’ve seen before, and at a time when the industry is struggling with significant issues that AI could eventually address.
“We have to catch up to that SaaS model,” Schwamm said.
Healthcare leaders across the country are pulling their legal and compliance teams into the conversation, in some cases developing strategies based on hypothetical issues. And they’re trying to educate clinicians who might see just the good in AI and not understand the ramifications of fast adoption without governance.
AI “lowers the bar for non-technologists to use sophisticated technology,” Schwamm said.
Cunningham noted that AI tools are being tested out across the enterprise, often in small programs that show very specific, though limited ROI. Leadership has to find a way to keep track of all these programs and integrate them into a governance structure.
She said health system leadership needs to take a step back and assess the new tools and technology being pitched in healthcare. Many vendors, including those in AI, are aiming at the patient experience and engagement space, with products that promise to improve the clinician-patient relationship. But healthcare organizations are struggling with stress and burnout, to the point that a new tool that offers results “with just one more click” isn’t a good selling point, and products that aim to give clinicians time to take on more patients are just adding to the misery of overloaded workflows.
Several panelists at the HIMSS forum said AI’s potential to synthesize data and take pressure off of clinicians has to be balanced with an understanding of how healthcare should manage the vast amounts of data coming into the enterprise. That may mean creating a change management strategy devoted solely to AI adoption, to give healthcare leaders an understanding of how AI will take that data and make it useful to clinicians.
AI’s potential to address workforce shortages in healthcare may also mean it can be used to replace people, especially for positions that hospitals are having problems filling. And Schwamm noted that as health systems and hospitals focus on operational factors to improve their financial standing, AI could work its way into labor negotiations.
Cunningham said the industry will eventually have to get its act together and pull all the loose AI threads into one organized strategy.
“What does it all look like five to seven years from now?” she asked. “How will all these things that we’re doing come together?”