The average patient will generate around 80 MB of data per year. Even a small hospital which serves a population of only 50,000 is going to generate on average 4 terabytes of data in a single year. Try scaling that up to a large hospital with several million patients and the numbers become ridiculous. How is that supposed to be stored on a blockchain?
Well, we can have one blockchain for a section per floor per hospital.
And we're going to have CardioHeidelberg token, GastroPitieSalpetriere and so on
Storing it alone wouldn't be that big of a problem, the current system is doing just this at this point but the problem will be the so-called decentralization, how are you going to keep the nodes up, and most importantly who is going to pay for this. Current databases that are running in hospitals here (although not on that scale, as it would be impossible for a hospital around here to have millions of patients) are optimized for this, there are a few copies of the data in the health ministry storage but even with these tiny things there will be no real decentralization but rather giving custody to another entity.
Why do that? No reason!
Another problem I see in this concept is that as we know and use the blockchain everything is final, there is no chargeback there is no block deletion, what will happen in case of human error while handling the data? It's pretty sure they have a way of removing it, so, at this point, what's the point of an immutable blockchain?
The only potential solution I can see is to store it all on a centralized database as it currently is, and use blockchain to manage access policies to this centralized database. Sure, this is possible, but it is also completely pointless and holds no benefit over the current system.
In short, rely on a traditional database that can act like a blockchain as close as possible. And we already have tons.