Arnold Kling has a series of well-reasoned pieces on health care spending here. In the latest one, he discusses the fiscal dangers of what he calls "activist medicine," which he defines in this way:
Activist medicine means undertaking procedures that have a low probability of affecting the final outcome.
This definition disturbs me somehow. Three problems in my mind. First, there's usually going to be a lot of uncertainty about the estimate of the probability. There may be a low estimate, but with wide confidence bands.
Second, the history of technological change suggests that sometimes only experience and a lot of fiddling and experimentation can convert a weak innovation into a strong one. Labelling a medical innovation as "activist medicine" prematurely could be the kiss of death.
Third, there's a lot of things that people could do for their health which have a high probability of affecting final outcomes, which they don't do. I'm not sure what to make of that, but I think it's important.