My Heart Is Breaking. And I Am Angry.
Next month, marks 13 years since my brother, a physician, died following an avoidable misdiagnosis.
I should have mourned and moved on. I couldn’t.
Every 40 seconds someone dies or becomes permanently disabled form a medical error. That’s a lot of fathers, mothers, children, and partners. The Institute of Mediicne predicts that everyone will be victimized by such an error ove rtheir lifetime.
In memory of my incredible brother, who always put his patients first, I became consumed to solve this problem.
Since then, I have spent nearly 40,000 hours, more than $4 million of my own money, and helped bring in another $2 million in support and research funding, including work involving the NIH, NHS, Stanford, and Harvard, trying to solve one enormous problem:
How do we stop people from being harmed or killed because the right diagnosis was missed?
That mission became Medome.
It took years. It took millions. It took two issued patents, with two more pending. It took building technology designed specifically to attack misdiagnosis differently from anything I knew existed.
And now it works.
Recently, Medome faced 28 extremely difficult hematology cases using history alone. It identified the correct diagnosis or the right diagnostic test in 28 out of 28.
That should matter.
Yet while technology like this is becoming available, much of the public conversation remains consumed by what AI might someday do to humanity.
I worry about AI safety too.
But I also worry about the human being sitting in an exam room today whose cancer is missed.
The father whose heart disease is mistaken for something benign.
The mother whose rare disease remains undiagnosed.
The child who loses a parent.
The parent who buries a child.
We can continue to discuss the hypothetical dangers of tomorrow, but can’t afford to ignore the preventable suffering happening today.
I often tell entrepreneurs: Start with something simple. Prove it works. Then expand.
I didn’t follow my own advice.
At HealthDrive, Dr. Alec Jaret and I started with dental care for nursing-home patients. Once that worked, we expanded into a much broader specialty-care platform.
With Medome, I went straight after the monster.
Misdiagnosis.
Some venture capitalists called it “boiling the ocean.”
They weren’t entirely wrong.
Get Steven Charlap, MD, MBA’s stories in your inbox
Join Medium for free to get updates from this writer.
Medicine has tens of thousands of diseases, overlapping symptoms, incomplete histories, medication effects, rare conditions, cognitive biases, and countless ways for important information to be missed.
It was brutally difficult.
But we did it.
What breaks my heart now is not whether the technology can work.
It is how slowly medicine is willing to test, validate, and adopt technologies that could make doctors more accurate and patients safer.
I believe every physician should want a second intelligence looking over their shoulder, asking:
What are we missing?
What doesn’t fit?
What else could this be?
What test would prove us wrong?
People including Zeke Emanuel, Vinod Khosla, RFK Jr., and policymakers at Utah’s Office of Artificial Intelligence Policy are challenging long-held assumptions about who and what can safely deliver medical care.
Good.
Medicine exists for one reason: the patient. At least it should.
Not doctors. Not hospitals. Not insurers. Not technology companies. Not regulators.
The patient.
So test these systems rigorously. Compare them head-to-head with physicians. Measure accuracy, safety, outcomes, cost, and access.
Then let the data speak.
If AI fails, expose it.
If it works, use it.
And if it saves lives, get it into the hands of doctors and patients as fast as responsibly possible.
I have even put my personal phone number on our website because this is not simply another company to me.
It is personal.
I couldn’t save my brother.
Maybe we can save yours.
Please share this.
The next missed diagnosis will not be a statistic to someone.
It will be their mother, father, husband, wife, partner, brother, sister, child, or best friend.
Maybe someday we can make the Institute of Medicine’s prediction that most of us will experience a diagnostic error something that used to be true.
