From Doctor to Entrepreneur: Why Healthcare Must Move from Diagnosis to Prognosis

I stopped practicing medicine five years ago. Some would say I went to the dark side. In those five years I do not think I ever shared publicly what drove me.

In the hospital I was trained to be very good at one thing: recognising what is already there. A diagnosis. What we almost never had time for was the question that sits in front of it. 

What is likely to happen to this patient, and what should I do about it today?

That gap is not academic as one might think. In our PERISCOPE work we see in real-life data that it takes 7+ days on average to diagnose an infection after surgery. Days in which something was already going wrong and got the chance to grow into a bigger problem. This is mainly because there are so many other things to attend to.

Around ten years ago I started working with what was then ‘new technology’ called AI. The fit with this problem seemed obvious to me. What also became obvious was how often these things get built the wrong way round. Someone has a technique and then goes looking for somewhere in healthcare to put it, instead of starting with a problem that a patient, nurse and/or doctor actually has, and sitting close enough to them to understand it properly.

So you need the clinical view in the middle of it. Someone who speaks the language, who knows which parts of a workflow can realistically be changed and which cannot, and who keeps asking what a prediction is supposed to make someone do differently, and when. Someone who also accepts how much testing medical use demands. I have watched a lot of models get built over the years. Very few of them ever get the evidence and the platform behind them to actually run in a clinic and be governed there.

That is the chasm we set out to cross at Healthplus.ai: from a diagnostic paradigm to an anticipatory one. From a single site AI-model that sits there to a scalable and AI-system that is governed closely.

Today we are serving multiple hospitals and we are active in multiple countries, reaching thousands of patients, and expanding into preoperative risk and other complications we can pre-empt. I never dreamed we would get this far. There is still so much to do and many many challenges to overcome.

And there are many, many colleagues, friends, experts and partners to thank for the part that is already done.