“When your enemy’s making mistakes, don’t interrupt him.” That may be what King Charles wanted to say to Democrats when he addressed Congress this week, but this wisdom is actually from Brad Pitt in Moneyball when he played the real life Oakland A’s famed general manager, Billy Beane. Instead, the King of England congratulated both countries for partnering on “nuclear fusion and quantum computing, and in AI and drug discovery.” Members of Congress were very polite and clapped vigorously, but probably whispered to each other “what’s ‘quantum computing’?” That uncertainty is not limited to Capitol Hill; health care stakeholders are also trying to make sense of what the future looks like. Read more about that in the One Thoughtful Paragraph below.
First, consider this news that demonstrates a focus on the future:
- AI is going to help identify the patients who should participate in clinical trials. This week, the FDA asked for input on a pilot program that uses AI in early-phase clinical trials. Specifically, the FDA issued a request for information (RFI) about how AI may be used to improve aspects of clinical trials like patient recruitment, safety monitoring, and data collection. Comments are due May 29, 2026. The FDA announced that it is already trying an early version of this idea with AstraZeneca and Amgen.
- In the future, if the FDA approves your cool new medical device, Medicare may pay for it right away. This week, CMS and the FDA announced a proposed coverage pathway called the Regulatory Alignment for Predictable and Immediate Device (RAPID) program, which is intended to reduce the lag time between FDA approving medical devices for breakthrough designation and CMS deciding whether Medicare will cover them. [Note: CMS decisions may be faster, but it is not guaranteed that CMS will agree to pay for these devices.]
- In the future, the FDA will totally probably know how to regulate medical devices that are so advanced they start to think for themselves. At an industry conference, FDA Center for Devices and Radiological Health (CDRH) Director Dr. Michelle Tarver said the agency plans to issue final guidance on AI-enabled device and software lifecycle management. She also said the FDA is working on a regulatory oversight approach to generative AI.
“Adapt or die.” That’s how Billy Beane, then General Manager of the Oakland A’s, explained his decision to revolutionize how baseball teams were built, prioritizing players who could get on base instead of relying on traditional scouting methods. The idea, popularized in Moneyball (based on Michael Lewis’ book), helped drive a record 20 consecutive game winning streak and a string of playoff appearances. Just like the institution of baseball, our medical establishment is not quick to embrace major shifts. But if medical schools fail to make AI literacy a core competency for our future doctors, they will be devalued by society, according to an economist-turned-AI consultant (Matt Hasan) in a new Health Affairs Forefront article. Mr. Hasan’s “adapt or die” mantra about modernizing medical education makes sense. First, note the reality of the news listed above. Second, AI is already competing with physician performance on certain tasks, with some models exceeding human doctors and identifying paths to lower costs. But Mr. Hasan makes a point in his post that resonates the most: “The most consequential clinical work involves patients who do not fit established patterns, whose trajectories confound algorithmic prediction precisely because they fall outside the training distribution of any model.” The conclusion is that medical schools should teach how to critically evaluate algorithmic outputs and leave the more mundane tasks to automated tools. It’s like when they were trying to recruit a catcher to be a first baseman in the Moneyball movie. The trainer said the pivot would be “incredibly hard” and Billy Beane shrugged and said to the catcher: “Hey, anything worth doing is. And we’re gonna teach you.”
