The New York Times has a pretty gripping series about the quest to develop a targeted drug for certain forms of metastatic melanoma. As of the second installment we're at a cliffhanger (spoiler alert!): the new drug is producing almost magical shrinkage of tumors. But one patient has died from cancer that crossed the blood-brain barrier--which the new drug cannot penetrate. Will the other patients in the early stage trial also succumb? How long will the remissions last?
I don't know how it will end, but I think it illustrates one of the primary problems that will afflict any attempt to control health care costs. When you read the description, it's hard not to be awed at the difference this drug made in the lives of people afflicted by a pretty nasty cancer. But presuming it survives all the trials, this drug will probably be pretty expensive. It serves only a fraction of people with melanoma, those whose cancer has a very specific gene mutation. It probably won't cure them, but only buy them a few weeks or months or years.
That's how cancer treatment has mostly advanced--not with a spectacular cure that can be funded by better targeted NIH money, or identified by comparative effectiveness research. It grinds out small improvements one at a time, experimenting with combinations of drugs and radiation and surgery, dosages, and timing. A lot of the improvement in mortality rates comes from better detection--but that means a lot of money wasted on tests, and biopsies for false positives.
Will the drug be "worth it"? What's the price of giving someone six months instead of one to say good bye to their family, or shrinking their tumors so that they don't die in pain? Technocrats can't answer those questions. We have to.