Timing is everything. You have to know just the right time to catch the wave.
As I look at the world of technology, I’m grounded as an engineer and a scientist, so I don’t get too hyped about the latest gadgets and gizmos that promise to make our lives better in every way. In my 26+ year career at Vizient, I have evaluated hundreds of new technologies that have been introduced into health care—some were good; some were bad. But I haven’t used words like “‘transformational” or “disruptive” for pretty much any technology in recent memory, except when it comes to artificial intelligence (AI).
AI will be transformative. It will fundamentally change the way that we practice medicine as well as how we deliver care in our hospitals and provider networks forever. AI is inevitable.
My job is to scan the horizon for technology—mostly medical devices—and help member hospitals identify and evaluate those innovative emerging technologies that can help them with their performance improvement initiatives. Innovation is new thinking not done before, and when you add that to high-impact focus areas, the assessment process and timing to adopt directly correlates to an organization’s true ability to improve patient outcomes.
Benefits of early adoption
I get asked a lot by hospitals about the “why-when-how” of AI and what separates early from late adoption thinking. Timing (when) is important, of course, yet so is the potential impact (why) of new technology and its components. Does AI provide more efficiency, improve quality, better outcomes, or in some cases does the software or algorithm enhance access and equity for certain health populations? The answer is a resounding “yes” to all these questions. AI is currently powering many new devices and is having a significant impact on care delivery.
AI enhancements in the diagnosis of diabetic retinopathy can address both access and equity. For example, what if we make an AI-enabled screening device for diabetic retinopathy conveniently available in local pharmacies or primary care offices. After taking a picture of your retina, the AI might autonomously diagnose more than mild diabetic retinopathy and the screen says, “You have a serious condition; you need to do a follow-up. Here are the phone numbers of caregivers that will schedule that visit for you.” Making this more readily available, especially in underserved urban areas, will greatly improve access and equity for everyone…and could prevent blindness.
Other AI opportunities on the upswing include use for routine colonoscopy where the GI specialist can use real-time AI as a “second set of eyes” to help detect polyps during the procedure. In this use, the AI analyzes each frame in the live video and alerts the doctor by drawing a box around a potential polyp. This technology was cleared for marketing by the U.S. Food and Drug Administration earlier this year and research shows it can improve quality metrics for colonoscopy.
AI also has the potential to make the workflow more efficient and reduce hospital delays in care. For example, AI-assisted stroke detection, triage and notification is a groundbreaking and highly successful application targeting treatment delays. In a new paradigm, AI-based algorithms quickly analyze computed tomography images taken in the emergency department and then automatically initiate the notification process when a stroke is detected. Because of the loss of viable brain tissue with each minute of delay and the narrow time window for effective therapy, this new process could substantially improve stroke outcomes. Nationwide, hundreds of U.S. hospitals are now using this type of AI for stroke detection.
There is also a real competitive advantage for the early adoption of AI by hospitals. Early adopters can work through implementation barriers and be better positioned as AI matures. For many, the barriers are safety and effectiveness, cost, reimbursement and clinical utility. i.e., can it work, does it work and is it worth it? In the case of stroke and diabetic retinopathy, both AI technologies are now reimbursed by the Centers for Medicare & Medicaid Services. This payment mechanism was a critical part of the adoption equation.
Planning for AI adoption
Typical of most newly emerging innovative medical technologies, sometimes working through the evaluation and implementation barriers still requires a necessary leap of faith. But an understanding that AI will eventually become standard care for a given medical scenario can lead some hospitals to early adoption and its advantages. Other hospitals may choose to wait until AI modalities become the “standard of care,” but they risk being left behind in advancing efficiency and improving outcomes.
Establishing strong multi-disciplinary, clinically integrated supply chain and value analysis committees within hospitals and health systems that can evaluate all facets of the use of AI is critical. Leveraging insights and their impact on patient care and outcomes requires input and analysis from across the organization. Important sources include supply chain, IT, medical and operational leadership, finance and strategic leaders, to name a few. Similar committees may already exist for traditional medical devices, but hospitals, in general, may need to ramp up their committees for AI. AI should be evaluated just like it would for any other clinical device that affects patient care, can result in patient harm, improves or reduces efficacy and impacts resource utilization.
And when evaluating the cost, keep in mind it’s more than just the upfront costs. It is about the total episodic cost of care. I would recommend these committees, think about how it fits into a value-based reimbursement model of care delivery and better patient outcomes. It’s a future calculation.
Health care can sometimes be a change-averse industry. But change is coming whether we like it or not. AI is not just a conventional medical technology; it is transformative and different for a hospital’s way of thinking. The future paradigm for hospitals and health systems is “disrupt or be disrupted.” Those in the latter category risk becoming obsolete. Therefore, keep an open mind—and maybe think outside the box a little.
Hospitals need to be ready. You need to follow the medical technology trends and educate yourselves on what’s out there. I have created some resources on AI, available in Vizient’s TechFlash reports, that provide a starting point. In these, I endeavor to cut through the hype, take an objective and transparent look at the clinical evidence of a specific AI opportunity and say, “Yes, this one is ready to implement.”
So, in closing, I’ll start at the beginning. The adoption of AI is all about timing. The timing is now.
About the author: Joe Cummings, PhD., is Vizient’s technology program director within the Performance Improvement Collaboratives group. Dr. Cummings' areas of expertise include technology assessment, evidence-based medicine and clinical supply integration and he has research experience with a broad array of medical devices, including artificial intelligence, remote monitoring, robotics, rapid diagnostics, enhanced biomaterials, metagenomics and telehealth. His educational background is in biomedical engineering.