Using AI to Accurately Code and Measure Physician Outcomes, With Solventum’s Dr. Travis Bias

Key Podcast Moments

• Travis draws a sharp line between what physicians think documentation is for and what it actually determines, arguing that the note isn't a billing exercise, it's the only record that exists of how sick your patient was, how complex your judgment was, and whether your organization looks like it's delivering quality care or falling short.
• Travis introduces the concept of the physician-translator as a clinical one: the ability to get to the root of what a customer, a colleague, or a patient actually needs versus what they're asking for, and to set expectations honestly about what the technology can and cannot do.
• Graham makes the case that physicians are being held responsible for quality metrics they don't fully understand, generated by systems they didn't design, from documentation they were too busy to complete
• Travis draws a clear lesson from two years teaching medicine in East Africa: finite resources produce better clinical reasoning, stronger physical exam skills, and a healthcare culture that is genuinely world-class at something U.S. medicine struggles with - knowing when not to order the test.

Dr. Graham Walker sat down with Dr. Travis Bias, Deputy Chief Medical Officer of Health Information Systems at Solventum, to discuss a problem that predates AI and won't be solved by it alone: the gap between what physicians actually do and what the data ever captures. In 80% of U.S. hospitals right now, Solventum's coding and documentation infrastructure is shaping how physicians are paid, how their quality is measured, and whether the care they actually delivered ever makes it into the systems that judge them. Travis is one of the few people in medicine who has seen that infrastructure from both sides, as a practicing family physician who knows what gets missed on a busy shift and as the clinical strategist who has spent years understanding what happens to that missing information downstream.

Travis didn't arrive at this work as a technologist who learned some medicine. He trained as a board-certified family physician, spent years teaching medicine in Kenya and Uganda, earned a diploma in tropical medicine and hygiene from the London School of Hygiene and Tropical Medicine, and maintains 48 active medical licenses across the country. His path into health technology wasn't a pivot away from clinical medicine. It was a decade-long realization that the systems evaluating clinical medicine were broken in ways most physicians couldn't see, and that someone with a foot in both worlds needed to do something about it.

The conversation Graham and Travis have about physician judgment and quality measurement is the one most health IT companies refuse to have in public. Graham names it directly: there's the patient, there's the EHR representation of the patient, and then there's the coded abstraction that flows into US News rankings, Vizient mortality reports, and risk-adjusted payment models. Travis's answer is the through line of the entire episode. Physicians are actually quite dialed into how they're being judged externally, and that is precisely how you get them engaged with documentation. The problem isn't physician apathy. It's that the systems capturing their work were never built to capture all of it, and the burden of filling that gap has historically landed entirely on them. AI ambient documentation, Travis argues, is finally starting to change that equation, not because it replaces physician judgment, but because it captures the clinical complexity that physicians never had time to write down.

Travis is, by his own description, a pragmatist. When he meets with health system clients, his job is to be honest about what the technology is designed to do and what it isn't, because a customer who discovers that gap six months after deployment is a failure, not a success story. That same standard shapes the way he talks about AI regulation, ambient documentation trust, and the role of the physician in a system that is changing faster than the frameworks governing it. The horse, he says, may already be out of the barn.

He also brings a perspective that most health tech voices lack entirely. Two years teaching medicine in Kenya and Uganda gave Travis a different calibration for what clinical excellence actually looks like, one where physical exam skills far exceeded anything he had developed in the United States, where community health workers manage chronic disease at scale through task shifting, and where finite resources produce something U.S. medicine rarely achieves: an honest conversation with the patient about when the test simply isn't worth ordering.

Thank you to our wonderful sponsors for supporting the podcast:

Sevaro is a physician-led telestroke and neurology company that delivers rapid virtual neuro coverage that’s reliable. Learn more at https://sevaro.com/

Evidently - Leading AI-powered clinical data intelligence https://evidently.com/

Top 4 Takeaways

Travis has thought longer and harder about these questions than most people on either side of the medicine-technology divide, and this episode reflects that. Here are the four ideas from this conversation that every physician should sit with.