What 20 years in hospital medicine taught me about starting a health tech company

Twenty years in medicine can take you to strange places. I have practiced as an academic hospitalist and emergency physician at top 10 medical centers, and I have also covered critical access hospitals in the remote corners of Alaska, California, Hawaii and other states. I came to value both: the intensity of a tertiary center, and the slower, more improvised rhythm of community and austere care.

Every one of those places had the same problem waiting for me at the workstation, and I walked past it for two decades before I understood something simple. A problem that follows you across an entire career is not an annoyance. It may be a business.

When I finally started one, I did it with no business background and no coding experience, while continuing to practice. Along the way there were decisions where the physician in me overruled the businessman, and the company is better for every one of them. What follows is some of what I learned.

The idea was lying around the whole time

Over two decades, the buildings changed constantly, but one thing never did. It was not medicine. It was absence paperwork that only a physician could sign, and it arrived every shift, in every waiting room. Our tolerance for it was endless. Nurses joked about it, we grinned and signed, and nobody was ever assigned to fix it. The idea sat in plain sight for 20 years, and none of us picked it up.

It took a forced stop for me to see it. The staffing company I worked for went bankrupt, I had worked too many nights, and I found myself out of the emergency department for a few months with time to think. So I did what we do with any finding that keeps recurring: I quantified it. I estimated, top down and bottom up, what share of emergency physician hours went to encounters that existed only to produce a signature. The number was large enough that I could not un-see it. I kept practicing telehealth to pay the bills, and I decided to act.

What I did first

I did what physicians do when we step outside our specialty. I went looking for experts.

I finished a 10-week online business course, which gave me vocabulary and very little else. I asked peers in private equity and corporate life how to raise money, and every one of them told me the same thing: you do not need other people's money to start, and if the idea is good, the money will find you. I spent months studying the market and the competition. Then I gathered the smartest people I could find into what looked, on paper, like a very good team, and watched the whole thing go downhill.

Realizing step one was wrong

Six months and roughly $80,000 of my own savings later, I was in golden handcuffs. The team was not failing, exactly, but it was working in a dated way: requirements documents (long written specifications agreed on in advance) and a ticket filed for every change. A fix that should have taken an afternoon became a ticket for the following week, and a product in its first months has to be iterated daily.

I also learned that accountability does not travel. In the emergency department, when someone says they will do something, it gets done, because the consequence of not doing it is standing in the room with you. Outside medicine, people would say it and never get around to it unless I chased them by email. The stakes were not life and death, so nothing was urgent.

Days before I made a decision, I ran an experiment. I asked an AI coding assistant whether a physician with no coding experience could build the whole product by describing it in plain language, and the answer was yes. I let the entire team go, keeping one 20-something who cared about the idea as much as I did.

The mistake was never the people. It was my assumption that I needed to be surrounded by experts, when I had been the expert all along. I knew the product better than any of them ever would.

Navigating the hurdles

On day three, the assistant forgot everything we had built, because none of it was written down anywhere it could read. So I treated it like a new resident and asked what it needed in order not to forget. The answer became two documents: a rulebook and a running progress log, read at the start of every session and updated at the end. That discipline still runs the company.

The tools failed me every other day, and physicians are not built for that. We trust the structures around us never to fail. I had to learn to check the work, to ask the same question twice, and to treat a confident answer the way I treat a confident history from an intern. Over the next 180 days, I shipped more than 3,300 changes with no engineering team, while working full time in a virtual urgent care, and most back-end changes passed an automated test before going live.

Meanwhile, colleagues kept asking whether I was still a real doctor. That one stings, because I never left.

What I would tell a physician starting now

None of this required leaving medicine. Anyone who has practiced through an electronic medical record rollout has already spent years translating clinical work into software for engineers who do not understand medicine, and that translation is most of what building a product actually is. The tools let one physician do the rest. But do not believe the hype that one person and one model can run a $10 million company. They are a lever, not a staff.

A few things I would say to a physician starting now:

• Build the worst version that works yourself, and only then decide what you cannot do. Test the idea on strangers, not friends. Everyone in your life will tell you it is wonderful, and none of that is data.
• Before any hire, ask whether it is truly necessary, what it yields, what it costs, and what happens if you do not make it. I have talked myself into needing a chief technology officer more than once and have never hired one.
• When you do hire, vet the person the way you would review a chart, with references as far back as they go, and hire for judgment and steadiness before technical skill. People who care about the work beat people who are in it for the money. That does not mean you pay them any less.
• Write everything down, with people and with machines.
• Try new ideas, and be willing to close them. We are trained to recognize when a treatment is not working and to change course without ego. Give your own ideas the same honest reassessment, on the same timeline, before sunk cost hardens into anchoring bias.
• Let the physician overrule the businessman. Disclose what you do with data, and put help in front of a vulnerable person before a payment button, even when it costs you the sale.

Medicine taught us to be methodical, to check our work, to own our mistakes and to prefer the truth to the comfortable story. Every one of those habits transfers. The line between patient and consumer is blurring, and physicians who build are the people best positioned to hold it. So measure the burden you have tolerated longest. If the hours surprise you, it may be a market. The expertise you need most is the one you already have.

Adam Z. Kawalek, MD, is an emergency physician, hospitalist and the founder of SickSlip, a digital health platform for absence documentation.