Code Red: Two Economists Examine the U.S. Healthcare System

April 23, 2012

GAO and CMS Spar over Research Methods: Media (as usual) Miss the Boat

Filed under: Uncategorized — David Dranove and Craig Garthwaite (from Oct 11, 2013) @ 1:22 pm

The New York Times and other media outlets are trumpeting a new GAO report that blasts an ongoing $8 billion Medicare demonstration project. CMS has put the money towards a new pay-for-performance scheme for Medicare Advantage plans. The media have focused their attention on part of the GAO study describing how most of the money will go to average performing plans (those receiving as few as three “stars” out of a possible five), with 90 percent of plans receiving some sort of bonus. If that were the gist of the GAO complaint, then the GAO would have been the ones guilty of wasting taxpayer money for writing a useless report. But the key elements of the GAO report pertain to a different matter: will the demonstration allow CMS to determine whether the new pay-for-performance scheme is superior to the existing scheme? Here the GAO report gets bogged down in the details of research methodology. The media understandably got lost in this discussion and have given this part of the report short shrift. As a result, I expect politicians to blast CMS for “giving $8 billion to bad health plans” and other stuff of nonsense.

Let me explain why it is pointless to focus on how the money is distributed among Medicare Advantage plans. The purpose of any pay-for-performance scheme is to provide incentives for improving quality. (A scheme that rewards the best plans but does nothing to alter the status quo really is a waste of money.) Some of the comments by the GAO and aped by the media would have you believe that the best way to improve quality is to reward the top achievers. Tell this to parents of a D student who would be grateful to see their child get C’s. Should they tell their child “we will take you on a nice vacation but only if you get all A’s?” Talk about killing motivation. The point is that there are some situations where a narrowly targeted pay-for-performance scheme would be best and others where it makes sense to offer a ladder of incentives that even the worst performers have some chance of reaching. It was kind of silly for the GAO to mention that the Medicare Advantage scheme is more like the latter than the former; it was a sure bet that the media would pick up on this as if it were some horrible idea. It is neither a horrible nor a splendid idea – given our current state of knowledge there may be no way to know if broad incentives are superior. That is why CMS wanted to do this demonstration! GAO does suggest that the demonstration methodology would not allow CMS to answer this question. This is the important part of the CMS critique but one that is likely to get lost amid the hoopla surrounding the alleged misallocation of $8 billion. Note that CMS has rebutted some of the GAO complaints about its research methods and I will leave this battle to them.

(I suspect that they battled over this for some time and GAO went public only after they had reached an impasse. I should add that others have criticized CMS for spending too much money on the demonstration; this is easily rectified by cutting baseline spending for the very lowest plans, effectively reducing the “intercept” but increasing the slope of the pay for performance curve.)

In closing, note that if CMS wants to save money and still motivate quality improvements, it could simply publish the report card results without offering tangible financial rewards. Lots of research shows that health care consumers, even those in Medicare Advantage, will vote with their feet and enroll in higher scoring plans, provided that the quality measures are easy to understand and widely disseminated. If the New York Times really wants to promote quality in the Medicare program, it should write articles about the results of the demonstrations. Criticizing CMS’s broad incentive scheme for Medicare Advantage serves only to breed further distrust in a program that many consumers look upon with suspicion.

Blog at

%d bloggers like this: