[sudhirj]

Experience pays in other professions that require competency, though: a surgeon who can proficiently perform a bypass in a short amount of time is far more valuable than a fresh graduate who potters around in there for a while. No one ever complains that a surgeon finished too quickly.

[VLM]

The surgery marketplace is way too ignorant to be a real marketplace. Nobody has any idea how long my mother-in-law was under actually under the knife and nobody involved other than the surgeon, has any idea what the going rate is for gallbladder removal. Two minutes? Two hours? No one involved on the "purchase side" of the market can even guess. Without a functioning marketplace, numerical metrics don't mean much.

Another interesting issue is she shopped for a GP a long time ago, the GP referred her to specialist who referred her to the surgeon, can't have much of a marketplace if the participants aren't making any decisions. The only decision she made was a decade ago she picked the guy who referred her to the guy who referred her to a surgeon.

The craziest part of a crazy "market" or whatever you want to call it, is it mostly works. Something this screwed up should predictably kill everyone involved, but most of the time, it works, other than blowing a lot of money. It is the truly shocking part of the situation, at least from a "free market capitalism is the only system that, in practice, actually works" mantra. Maybe that's not so true, given the observational evidence.

[gwern]

It gets worse, because it's such a probabilistic market. How does any patient tell the difference between a surgeon with a 10% fatality rate and a worse one with an 11% fatality rate? (At the usual estimates of the value of a life in the several millions USD, that 1% is worth a lot.) And if they somehow had that data, how would they adjust for the surgeons' working in different geographies or specializing in slightly different patient types?

This sort of problem is why making hospitals release medical data hasn't been very useful. And if you can't get much reliable signal out of large datasets with rigorous statistical analysis, how are ordinary people and gossip supposed to reach the right answers which could allow market mechanisms work?