By YiDing Yu, MD, a practicing physician and serial health tech and AI entrepreneur

It was a big day in June. New interns with freshly minted degrees had started on the wards just a week earlier. A two-year, multibillion-dollar installation of a massive electronic health record was going live today. Senior residents braced themselves for chaos. And at around 6:30 a.m. that morning, the first emails began to trickle in. “How do I print the patient rounding list?” More emails asking the same question. All across the wards of our 800-bed hospital, doctors could not print out a list of their patients. Somehow, despite many clinical workflow design sessions and super-user trainings, this detail had been forgotten.

“Who designed this thing?!” is probably a phrase we’ve all uttered at some point. There are probably a multitude of reasons for poor user experience among health care technology offerings, but a common reason among them is the lack of clinical input in the product development process.

While I hope that all technology leaders would agree that clinical input is invaluable during the user research process, some health care companies have gone further, hiring clinical product leaders to shepherd products from the ground up. Over the years, I’ve met a cadre of physician product leaders who have forged incredible careers building products that enhance the work of clinicians and improve health care for patients. These physician product leaders are indeed very rare — rare among physicians and rare among product leaders. However, as health care companies increasingly adopt AI in their solutions, clinical training and input into building responsible AI for health care applications are more important than ever.

So while these positions may be rare today, opportunities for physicians are growing. If you are a physician who has had an interest in building solutions that scale, then product management may be a career jump that is worth exploring.

What Makes a Good Clinical Product Leader?

What makes a good clinical product leader starts with what makes a good product leader: the curiosity and humility to deeply understand customer pain points, a natural intuition for problem-solving within constraints, and an outstanding ability to inspire and communicate across various stakeholders.

Where your clinical background and experience provide immeasurable value is in your ability to distill complex clinical workflows, presentations, or pain points for your colleagues and help design solutions that have broad applicability. Your experience will help your product and engineering teams avoid dead-end misadventures that cost weeks or even months of frustration and wasted effort.

For example, I was once working with a team who were building an AI engine to understand medical necessity guidelines. A particular complex set of published guidelines for spinal imaging was written as decision trees: if you answered “c” to Question 1, jump to Question 3, and so on. Because there were so many indications for the study (concern for new malignancy, staging for known malignancy, rule out cord compression, chronic back pain, post-spinal surgery, concern for infection, multiple sclerosis, etc.), the scenarios were enormous and dizzying for any human to try to map on their own. In addition, some scenarios, like concern for infection, were sometimes secondary to another scenario (post-surgery), causing the decision tree to seemingly have duplicative criteria.

To a non-clinical mind, the order of the decision tree and the nesting of scenarios seem important. To my clinical mind, I realized immediately that it was arbitrary. For the vast majority of clinical scenarios, the order of clinical criteria does not matter. You could just as well determine that a patient requires spine imaging because you are concerned she has an infection in her spine and is status post-surgery, or because you are concerned she has a surgical site infection.

This simple understanding, once explained to a team of brilliant data scientists, made a massive difference. It fundamentally changed the architectural approach of the clinical AI model. Had the team attempted to factor in scenario order, they would have built a more complex architecture, spending unnecessary weeks to months doing so. That complexity becomes exponentially more taxing to maintain over time, especially as guidelines update or variations occur. Instead, clinical product input supercharged the speed of development for AI.

Building Great Clinical Products

Beyond clinical knowledge, I’ve found that my understanding of clinical workflows and first-hand experience of the “pain” of providing health care are the greatest assets I bring to a product team as a clinician. We all hope that product managers spend dozens of hours with their end users to design a product, but the truth is that access to clinicians is incredibly limited. Hospitals and clinics do not have the luxury of providing extra clinical hours to educate vendors on their workflows and practices. The consequences of this are readily evident in lackluster user experience, awkward workflows, and more recently, shiny AI agents that require more work than they are worth.

Health care deserves products that delight end users and deliver meaningful return on investment. The best clinical product leaders I’ve come across have a knack for both. They are hungry problem-solvers, often implementing solutions in their clinical lives — new checklists, workflows, care delivery models, even tech solutions — before jumping into tech. They are no stranger to inspiring stakeholders to support or adopt their changes and know that even the best ideas need a case for funding.

Most clinical product leaders made the jump into tech early in their careers — usually because their interests drove them early to pursue these opportunities. However, I’ve also been privileged to meet, mentor, and even hire several clinical product leaders who made mid-career moves. Wherever you are on your journey, if designing great products is a calling for you, I encourage you to seek out opportunities to network, learn more, and jump in when the role is right.