Prizes, Prose Messages, and Prior-Auth Battles: AI in Oncology is Here to Stay
Expert Perspectives · Rahul Banerjee, MD, FACP · 2024-07-24
There were so many updates with artificial intelligence (AI) in the news this month! Here are just a few:
- Prizes: Our very own Dr. Waqas Haque, MD, MPH (Twitter/X: @waqashaquemd), a member of Oncology Data Advisor’s Fellows Forum, has won an APEX Award for his series entitled Exploring AI in Oncology: The Convergence of Technology and Cancer Care. Congrats, Waqas! The award, bestowed to him and the Oncology Data Advisor team, awards excellence in communication. Read our press release about it or check out the podcast and blog series here
- Prose Messages: Researchers at New York University asked primary care providers to rate the quality of 344 messages written in response to in-basket messages from patients via electronic patient portals: half were written by healthcare providers, while the other half were written by AI. The study, published on July 16 in JAMA Network Open, showed that AI-written messages were rated higher in terms of empathy and positivity; however, AI-written messages were more likely to use flowery prose language (mean Flesch-Kincaid grade levels 8.1 vs 6.5, P<0.001)
- Prior-Auth Battles: On July 10, the New York Times published an intriguing article about physicians using AI chatbots to help compose prior-authorization requests and written appeals. The New York Times described these AI tools as a “cudgel” against reckless behavior by insurance companies—their words, not mine, but I can’t say I disagree with the sentiment
So what’s the future of AI in oncology? The cliche thing to do would be to ask an AI chatbot and analyze its response, but I will hold off. Instead, I’ll highlight a nice June 2024 position paper by the American College of Physicians. One of their key pillars is that “AI-enabled technologies should complement and not supplant the logic and decision making of physicians and other clinicians.” Another key pillar is that “in all stages of development and use, AI tools should be designed to reduce physician and other clinician burden in support of patient care.”
I love these two pillars – putting them together, AI should help us accomplish what we want to do and streamline what we don’t have time to do. Many of Dr. Haque’s podcasts for Oncology Data Advisor focus on the former goal: for example, using AI to help physician-scientists identify new drug targets in precision oncology. I am encouraged to see AI innovators working on the latter goal as well, focusing not just on things we don’t have time to do but on things we don’t want to do. I’m hopeful for more AI-fueled innovations in years to come that help prevent physician burnout and that allow us to focus on what matters—good patient care.