Health Care AI Is Exposing an Age-Old Problem: How Work Actually Gets Done

AI technology in healthcare is rapidly evolving, but integrating it into existing workflows and job descriptions remains a significant challenge. While AI can detect critical issues like sepsis before symptoms become apparent, human oversight is still required to determine the appropriate course of action. Similarly, AI can assist primary care doctors in managing conditions that previously required specialists, but it cannot replace the expertise of specialists like nephrologists.

This creates a dilemma as AI can increase capacity, but it cannot create roles that do not already exist. According to a survey by the Scottsdale Institute and Deloitte Center for Health Solutions, a gap exists between healthcare systems' enthusiasm for AI transformation and their ability to operationalize it. Health systems rated themselves only 2.7 out of 5 on their ability to turn transformation efforts into scaled, AI-enabled operations.

Tampa General Hospital has taken a practical approach to AI and predictive analytics, integrating over 110 AI applications without leading to layoffs. Instead, the focus has been on changing how clinicians use the information provided by AI to enhance patient care. For example, the hospital uses AI to monitor patients' physiological changes and alert caregivers when potential problems arise, allowing them to confirm the concern and coordinate appropriate medical teams.

However, implementing AI solutions can be particularly challenging when there is a shortage of healthcare professionals. Dr. David Carmouche, chief medical and commercial officer at Lumeris, highlights the difficulty of meeting the needs of the entire population with a constrained workforce. AI could potentially help primary care physicians manage conditions that require specialist care, but it must be implemented thoughtfully to ensure that the workforce is comfortable with the new tools.

In clinical research, AI may accelerate certain tasks, but the overall process of bringing a trial to fruition can still take months. Dr. Jennifer Litton, chief clinical research officer at MD Anderson Cancer Center, emphasizes the need to streamline processes surrounding AI implementation to expedite patient enrollment. Ultimately, the successful integration of AI in healthcare requires careful consideration of how it fits into existing workflows and addresses workforce limitations to maximize its benefits for patients and healthcare providers.