- Pay for Performance:The Next Best Thing
Here's the galloping history of health care financing. After Medicare began in 1965 and some new technologies were introduced in the '70s, health care costs increased dramatically. In the '80s, there was an attempt to check those costs with HMOs and by instituting "diagnostic related groups," a scheme in which doctors and hospitals were paid according to a fixed schedule of diagnoses. There wasn't much to be done about DRGs except fiddle with definitions to make sure patients were not pitched back out onto the street too quickly, and by and large, DRGs did help control costs. For a while, HMOs and the managed care movement of which they were a part also helped control costs, but in the '90s the model began to give out. Employers pitted HMOs against each other and switched plans often enough that patients in a system one year might not be in it the next, which made it a challenge to meet long-range goals. Also, employers heard complaints from their employees about restricted choices of doctors or hospitals, long waits, and too many gatekeepers. Managed care's control over cost eventually relaxed—and now costs are escalating once again.
"Pay for performance" is the next best idea in the health care financing world, aimed this time not just at controlling cost but improving quality. Experiments in commercial insurance and at Medicare follow the wisdom of Wall Street: tie compensation to performance. Employers, still footing most of the bill for health care, are asking insurers to prove they are worth the money. Insurers are looking for predictable costs and better quality to remain attractive to employers. Medicare, like employers, wants to get what it pays for, and like insurers, wants predictable cost and improved quality.
Pay for performance programs are not one thing, they are many. Some insurers' quality measures are for doctors, others for hospitals. Hospital programs may either reward "good" hospitals (by giving the hospital an opportunity to earn more than what was contracted) or punish "bad" ones (by withholding part of the contracted rate until quality measures are in). Doctor programs reward "high quality" doctors with bonus payments. Each scheme has a different logic and may have a different impact.
One concern about pay for performance plans is that, absent knowledge about whether a particular clinical process results in a better outcome (and absent reliable consensus about what a "better outcome" is), they may tie compensation to something over which a doctor or hospital has no control. Often, though, plenty of evidence supports the intervention that the programs are designed to increase. Diabetes and asthma, for example, affect large numbers of people, with painful, protracted, and expensive consequences. Controlling blood sugar and managing asthma with medication prophylactically are not just good guesses—they really do help patients stay well and they really do save money. Getting that aspirin into a person with chest pain in the ER really has demonstrated effectiveness in reducing mortality from heart attacks. These are good things to reward doctors and hospitals for doing.
It remains to be seen which programs will actually lower costs, improve quality, and keep doctors and hospitals from feeling alienated in the process. Right now, a thousand flowers of pay for performance are blooming. But this is the real problem: such a fragmented approach, resembling U.S. health care in general, is unlikely to yield dramatic improvements. Good care can control the effects of asthma in a child with health insurance, but pay for performance leaves those children with no health insurance—now nine million of them—with no better health care than they had before.
Pay for performance may be a way to play the current health care game more prudently, more cost effectively, and with better outcomes for those inside it. But why does "rearranging the deck chairs on the Titanic" come to mind? Because pay for performance is only the next best thing. The very best thing would be to stop playing the current game and design a system—a costworthy, predictably priced, prevention-oriented, and quality-rewarded system—that works for everybody.
[End Page c3]