Also many providers still avoid the cloud (despite proven PHI compliance) and run their own inefficient, in-house server farms with thousands of tech staff because of this irrational fear. This causes a number of problems:

- It assumes a hospital system is better at infrastructure security and maintenance than Amazon or Microsoft

- Server costs per project are stupidly expensive. A VM that goes for $10/mo in the cloud ends up costing $200/mo to whatever department that foots the bill.

- You can't deploy a VM or app with a click. Instead you need to file a demand, then someone needs to make sure they have the resources available, then someone needs to manually spin up a VM for you...

- You can't build services that dynamically scale: this ends up being a problem for computationally expensive ML work - Tech spending on a hospital is overwhelmingly spend on maintenance and uptime rather than innovation or purchasing solutions

I could go on forever.

* This is also complicated by the fact that the industry isn't really incentivized to improve your health. Why run a predictive model and intervene to reduce your need for a stent when I make money from that operation?

The healthcare industry has two sides: the providers and the insurers. A hospital may not have the incentive to steer you away from unnecessary surgery, but your insurance company definitely does, since they'll be the ones paying for it. And your insurance company may have even more information to base its models on, since it knows about all the procedures you've had at any provider. So I'd expect insurance companies to be pioneering this kind of health-improvement software, not hospitals.

Your health provider has zero financial incentive to improve your health, and your employer or insurer will get more ROI from nitpick things like delaying claims or pushing out prescription fulfillment to reduce the spend this quarter.

Over the long run, it would seem to be in the insurer's best interest to keep everyone healthy for as long as possible.