Nikki Patton · Advisory, clinical AI and the end of life
I spent two decades in post-acute care, from the field through vice president. That meant two jobs running at once. One was commercial: build the referral base, grow the census, make the numbers. The other was human: be in the room when a physician said the word hospice out loud for the first time, and stay there afterward.
Most people in healthcare do one of those jobs. Doing both changes what you notice. You stop seeing prognosis as clinical information and start seeing it as a load somebody has to carry across a room and hand to a person who is not ready for it.
Underneath every title I have held, the actual work has been the same: normalizing death and grief — at work, at home, and in healthcare settings where even experienced providers do not feel prepared to deliver a poor prognosis with transparency and grace. Saying the hard thing out loud. Holding space for a decision nobody wants to make. Making room for the family conversation that has been buried for a lifetime.
A prognosis is not data. It is something one person hands to another, and the handing matters as much as the thing.
Then I moved inside health technology, where I still work today, and the two halves came back together in a way I did not expect. I watch roadmaps get built. I have learned what actually ships, what "the model flags it" means operationally, and how far a product decision sits from the bedside where it lands.
And I saw what nobody was building. Predictive models are getting good at identifying who is likely to die. Governance programs are getting good at validating and monitoring them. Between those two competencies sits a room, a family, and a person who has to say something — and almost nothing has been written about it.
That gap is the work. It is the only thing I do here.
