It is a very inefficient system with a lot of unnecessary couplings and vested interests and it does not allow for a major improvement via a sequence of local o

It is a very inefficient system with a lot of unnecessary couplings and vested interests and it does not allow for a major improvement via a sequence of local optimizations. Any change is likely to hurt someone and vested interests scream: see what you do to some_choosen_groop. IMO a rebuild is needed to fix it. For example, if $X gets billed it is seldom the price that gets paid. Usually insurance has negotiated rates of 0.1-0.2X and pays that. Different insurances pay different amounts for the same procedure. Hospitals know this and internally amortize expectations (to, say, 0.15X) pretty well. Now for the stinky part: someone without an insurance gets a strong push and occasionally a threat to pay the full bill. The patient might push back, but success varies and poverty is the only high probability way to successfully lower the bill. It is a BIG pain for both sides, as it usually takes years to collect even a part of this and any money will likely be shared between hospital and any organizations that help collect. Many hospitals would gladly take say, 0.2X upfront (their amortized expectation is $.15X). But they must bill same amount to everyone; thus they bill an outrageous amount and the few without insurance lose BIG.