Megan McArdle writes,

I think the central difference between me, and the people who think that IPAB’s reimbursement-rate powers will be a big help in controlling health care costs, is that the latter group tends to think that a lot of expensive health care problems are like back surgery–something that doesn’t do any good, but gets done anyway, because of desperate patients and arrogant/ignorant/greedy surgeons. I tend to think that more of the questions are like this one. Is spending $50,000 to give a pancreatic cancer patient an extra 5-9 months of life a wasted expenditure, or a medical advance?

I think that the tougher questions are ones where probability is involved. A screening protocol has a probability of x of providing a benefit, but it has a probability of y of showingn nothing and a probability of z of being a costly false positive. Overall, the average cost per life saved may be in the hundreds of thousands of dollars. But try arguing that to the person whose cancer was caught early and treated because of the protocol.