Alex Butler, MD, MS

I practice pediatric hospital medicine and have a master's in clinical informatics from Columbia. Through RAD Medical, I work with early-stage health tech companies as their clinician: the experienced voice in the room while everyone else is moving fast.

I've worked both sides of the divide most health tech falls into. I was head of product at River Records, building ambient documentation and the record underneath it, and I'm head of product and clinical lead at Alcott. My fractional work runs from seed companies with a demo and a hypothesis to Series C companies with live deployments inside health systems. I also still see patients, which is the whole point. The instinct I bring to a product decision comes from the same work your software has to survive.

What I've come to believe after enough of both: the hardest problems in clinical software are rarely about the model. They're judgment calls, dozens of them, made early and quietly by people who have never been responsible for a patient at 3am. Some of those calls were mine, and I got a few of them wrong. My job now is to be there before they're locked in, with the answer pointed at the patient and the person caring for them.

What I care about

Most of my writing comes back to a few questions. When does a human actually need to be in the loop, and when is "human in the loop" just there for the lawyers? Who is accountable when a system does exactly what it was built to do and the patient still ends up worse? What happens to patients as AI works its way into the visit? I try to argue these from the bedside, which is usually the view that's missing from the conversation.

Read the writing →

Who you'll work with

Most of the time, me. Sometimes both of us.

When a question is as much technical as clinical, I bring in Derrick Chu, MD, PhD. He's a primary care pediatrician at the same Boston hospital and an MD-PhD out of Baylor, where his doctorate was in the human microbiome. He's first author in Nature Medicine on how an infant's microbiome develops after birth, which was large-scale data science before the phrase caught on.

He's also an Epic physician builder, so his opinions about the EHR come with build hours behind them. He runs Epic optimization for his department and chairs the workgroup that decides how ambulatory workflows change. We advised on product at River Records together, which is where we first argued about the same problems. Data models, standards, and honest evaluation of AI output are his end of the work.