The entry of AI into healthcare is starkly different from the accepted adage of “move fast and break things” in consumerism.
In episode 1 of AI Meets Life Sci, Kayleen Brown, managing editor at DeviceTalks, and Brian Buntz, pharma and biotech editor, discuss the opportunities, limitations, and direct impact of Ai in healthcare fields including cardiovascular diagnostics and clinical trials.
Dr. Paul Friedman, chair of the Department of Cardiovascular Medicine at Mayo Clinic, and Dave McMullin, chief business officer of Anumana, share their successful experience detecting 33% more potential heart failure in 20,000 patients globally through an AI software. They explore the applications of predictive and generative AI, followed by the importance of algorithm training.
In one case, they ended up with an entirely fake yet believable study reference. The moment proved that in healthcare, we must be careful to move fast and cure — not break — things.
In the second segment, Jeff Elton, PhD, CEO of ConcertAI, dives into how leveraging AI in clinical trials results in clinicians and staff having more time for human interaction, essentially reducing the administrative pains hospital technologies created in the first place. The medical arena may never be ready for true disruption with patients hanging in the balance, but AI can and will responsibly improve workflows, screen, diagnose, and eventually treat diseases earlier.
Technology in medicine today has made real strides in access to care, and AI will materially impact the quality of care. Is AI the key to getting us all back to health?
Tune in and subscribe to AI Meets Life Sci on all major podcast channels and follow youtube.com/@DeviceTalks or AI Meets Life Sci YouTube Podcast to ensure you never miss an episode.
More on how AI could enable healthcare innovation
Friedman and McMullin represent a collaboration involving a world-renowned institution, Mayo Clinic, with years of real-world experience developing and commercializing life science products. This same relationship is well-reflected across new hospital system venture arms such as Kaiser and Providence working with industry to redefine how healthcare will be delivered over the next 10 years. The FDA has recently announced the Digital Health Center of Excellence, recognizing that the pace of AI means that regulation no longer has years to follow innovation — it has minutes. This model brings together clinicians, engineers, regulatory experts and business minds who have the know-how to develop and launch healthcare software with proper oversight. Most recently, the FDA has revised guidance to regulate health apps that have snuck by as “wellness devices” to be “Software as a Medical Device,” thus helping reduce unintended consequences.
Anumana created an AI algorithm that enabled more potential heart failure to be detected with incredible success using traditional echocardiograms (ECGs). Friedman makes an important point that ECGs are an extremely common tool, allowing clinicians to trust the data inputs and objectively determine whether or not they agreed with the AI’s output. This approach has naturally reduced providers’ inherent distrust in new technology due to the chance of a negative patient outcome. In the study, the AI algorithm analyzed ECGs from 20,000 patients and detected potential heart failure 33% more than traditional methods. In a smaller related study of 100 pregnant women, detecting 6% more potential heart failure patients. This is remarkable since their pregnancy symptoms masked the heart failure, thus paving the way for silent atrial fibrillation and other asymptomatic diseases to be AI’s next targets.
While this is a remarkable success, Friedman and McMullin also had examples of why AI innovators need to be careful with AI.
Jeff Elton, CEO of ConcertAI takes another approach to defining AI as a tool that generates insights and allows us to understand processes. Through this lens, AI can review existing literature to determine where causation may be at play, not just correlation, and vice versa. AI can review claims and scan thousands of patient records to determine whether or not they meet inclusion and exclusion criteria for clinical trials. AI can help match patients with the right clinical interventions and, almost more importantly, minimize the introduction of interventions that would not work. Though still in its infancy and far too costly, we’ve seen this idea used with great success in the pharmacogenetic space.ConcertAI’s software technology has enabled research teams to become more effective in their processes and take on three to four times more studies with the same staff. As the team has built confidence, they have been able to expand beyond Centers of Excellence into academic and community centers, further supporting their vision to expand access to care for patients. Elton mentions embracing AI is key to balancing the conservatism in healthcare and that success will solidify the next wave of advancements in medicine over time.
Haley Schwartz is a passionate commercialization expert with over 10 years of sales and marketing experience in medtech. She started Catalyze Healthcare to work with medical device companies struggling to commercialize in the U.S. by focusing on developing marketing plans, launch plans, go-to-market strategies, competitive analysis, pricing models and more. Contact her at inquiries@catalyzehealthcare.com or via LinkedIn for advice on your next product launch.
The opinions expressed in this contribution are the author’s only and do not necessarily reflect those of Drug Discovery & Development or its employees.
Filed Under: AI Meets Life Sci, Cardiovascular, clinical trials, Drug Discovery, Podcast, Special Feature