Dancing around a problem

WonkBlog notes a CBO writeup on the failure of some Medicare demonstration projects to reduce spending.  The summation (emphasis mine):

Why didn’t the demonstrations reduce costs? Largely because they didn’t reduce the quantity of care delivered. Some programs actually correlated with increased hospital admissions. A few saw reductions. On balance, it was pretty much a wash — and a troublesome sign for the health reform law’s soon-to-launch attempts to curb Medicare spending.

That rang a bell, which led me to dig up this passage from Money-Driven Medicine (emphasis mine):

Critics [of pay for performance] point out that very few of the performance targets address the problems of overtreatment.  While CMS is rewarding health care providers to do "more" in the form of tests and procedures that they might overlook, there are few obvious incentives to do less.  "Counting how many patients survived bypass surgery is one thing," says one New York City hospital executive.  "But how many survived a surgery that they didn't need?  That's the important number that you'll never see."

Dr. Stephen Jencks, Medicare's director of quality coordination, concedes that the critics have a point:  "I would say we are moving much more slowly on trying to prevent overuse than in trying to fix underuse," he acknowledged at the end of 2004.  "If I tell a physician he shouldn't do a surgery he wants to do, I personally would anticipate a lot more resistance than if I told him he should give a medicine he wasn't thinking of giving."  Yet if Medicare and other payers don't find ways to locate and discourage unnecessary treatment, pay for performance will only add another layer to health care inflation.

Telling doctors what to do is hard, but it's even harder telling them what not to do.  We're going to have to grapple with that in a serious way if we want affordable health care.