Peter Fitzgerald, during his keynote address at ISET 2024, affirmed that algorithms cannot replace physicians.
In Miami Beach, FL, discussions surrounding the future of artificial intelligence (AI) in the medical field go beyond mere speculation. The tools not only have the potential to refine clinical decision-making but also to address healthcare delivery challenges, as highlighted in a session at ISET 2024.
Emphasizing that physicians play an active role in embracing this transformation, Peter Fitzgerald, MD, PhD (Stanford University, CA), stated that they are well-equipped to explore the full capabilities of AI. His keynote presentation at the ISET Town Hall, themed “the good, the bad, and the scary” of AI this year, captivated the audience.
Fitzgerald pointed out a recent development where Google researchers shared data on an open-source platform, arXiv, indicating that their AI model “exhibits superior bedside manner compared to human doctors and provides more accurate diagnoses.” However, he cautioned that understanding the nuances of healthcare goes beyond algorithmic capabilities.
Highlighting the essence of human judgment in healthcare, Fitzgerald stressed the importance of knowing not just how to act, but also when to act and, crucially, when not to act—an aspect that algorithms may not encapsulate.
Acknowledging the existing wealth of healthcare data, Fitzgerald noted the challenges posed by its disorganization, distribution, deviation from protocols, siloed nature, and lack of structure. Deciphering and sharing this data effectively within healthcare settings require a unique set of skills not commonly found in traditional healthcare environments.
In discussions with TCTMD, Fitzgerald recognized the apprehension among clinicians towards integrating AI tools, fearing the imposition of unnecessary interventions on patients or a loss of control over decision-making. However, he highlighted the potential of AI to enhance communication of health metrics, provided it is implemented within a supportive infrastructure without overwhelming clinicians with excessive data. He noted that successful adoption would necessitate the involvement of the entire healthcare team, with nurses showing greater openness to this shift than physicians.
AI’s Role in Patient Engagement
Fitzgerald underscored the vast potential of AI in healthcare, distinguishing between predictive AI for anticipating future events and generative AI for creating new content. Beyond conventional applications such as image analysis, he envisioned AI’s role in promoting preventive care and addressing healthcare disparities.
He emphasized the need to proactively engage individuals before they present with acute conditions, advocating for a shift towards preventive strategies. Companies like Walgreens, Walmart, Best Buy, and UnitedHealth are leveraging their customer engagement expertise to develop innovative ways to interact with patients using AI.
While acknowledging the complexities surrounding AI integration in medicine, Fitzgerald highlighted persistent challenges such as adoption barriers, regulatory hurdles, data privacy concerns, and interoperability issues. Amidst these obstacles, he underscored the indispensable role of clinicians within the healthcare ecosystem.
Mark Coticchia, corporate vice president of Baptist Health Innovations (Miami, FL), echoed the sentiment of significant advancements in health system efficiencies and precision care facilitated by AI. However, he cautioned that the transition won’t be seamless, citing the importance of training algorithms on diverse datasets to avoid biased outcomes and the necessity of integrating AI insights into clinical workflows effectively.
In navigating these challenges, Coticchia emphasized the gradual nature of transformative solutions and the time required for their impact to be realized.
