NN regulations make such economically-sensible, mutually-beneficial business arrangements illegal. Everyone's got to buy the mandated, "neutral" service tiers – rather than other cheaper mixed-traffic-treatment formulas that could meet their needs.
If you want a high-bandwidth low-latency guaranteed-reliable fully-provisioned circuit then it's going to be expensive. Providing the same guarantees but only for a single application doesn't reduce the cost at all. If the customer is paying you to guarantee the delivery of packets, it doesn't cost you anything more if the customer then decides they're going to be YouTube packets rather than telemedicine packets. None of the things that cause that level of service to be expensive are application-specific so there is no savings to be had by treating applications differently.
As far as I understand it, nothing prevents the hospital to get:
Connection 1: Availability guarantee 99.999%, bandwidth 150Gbps max with 50Gbps dedicated, RTT for last mile: <=80ms
Connection 2: Normal, small business grade connection.
Then the hospital can use connection 1 for critical medical equipment and connection 2 for everything else. But, they could do just the reverse (why would they?) and the ISP would need to abide by connection 1's guarantees regardless of content (assuming the upstream connection from the provider is also good enough).