Skip to main content

Advanced Search

Advanced Search

Current Filters

Filter your query

Publication Types



Newsletter Article


Debating the Power of Performance Measures

By John Reichard, CQ HealthBeat Editor

December 15, 2006 -- Although the pay for performance ship seems to have sailed in the Medicare program, Democrats may not be the most gung-ho of shipmates. Rep. Pete Stark, D-Calif., delivered a blistering attack on "P4P" in a recent speech, and at a Washington, D.C., forum Friday, a respected Democratic health policy analyst voiced pronounced skepticism about instituting P4P for physician payment.

That those two developments signal a potential reversing of the gears on pay for performance in a Democratic-controlled Congress is questionable, but it may sharpen the debate over which measures get added to Medicare payment as time goes by.

A former senior Bush administration official and a private sector executive at the event defended the current thrust toward P4P in Medicare.

Robert Berenson, a senior fellow at the Urban Institute who ran Medicare payment policy and managed care contracting late in the Clinton administration, told the forum that P4P in big doses might not strengthen quality much.

"We need to carefully develop criteria for opportunistically and strategically using P4P, and not overload it with expectations of transforming the health care system," he told the meeting sponsored by the nonpartisan Alliance for Health Reform and The Commonwealth Fund.

Berenson said the incentive of an added bonus payment might be "marginal," particularly with respect to physicians. An added percentage point or two of payment can make a big difference to a health plan or a hospital given their tight margins but may be of little significance to a doctor's office, he said.

The more powerful way to change provider behavior is to address the basic payments they receive, not the small add-ons they get for better performance, he suggested. "The incentives embedded in basic payments applying to all providers are much more powerful than P4P marginal incentives," he said.

"The tail of P4P should not wag the dog of basic payment policy," he said. Thus, if Medicare payments to primary care physicians do not support what they need to be doing for the chronically ill, "change the basic system," he counseled. "Don't expect P4P to solve the problem, although it might be part of the solution."

Performance measures now tilt toward "process" rather than "outcome" measures. Thus they assess whether a provider is giving care in a certain way—for example, giving a hospital patient who has had a heart attack a beta blocker drug—rather than whether the patient was alive a month after leaving the hospital, for example.

Berenson noted that there are problems with both types of measures. Worse outcomes might reflect the fact that a doctor treats a sicker group of patients, for example. And they don't tell the doctor what specifically he or she might do in treating a disease to improve the outcome, he said.

Outcome measures "require case-mix adjustment, create perverse incentives to not treat sicker or more difficult patients and usually do not produce actionable information," Berenson said.

Process measures solve some of these problems, and may not require case-mix adjustments, for example. But they "may not actually be associated with better outcomes," Berenson said. To drive home that point he referred to a study released earlier this week that found only a weak link between hospital performance on process measures used by Medicare to compare hospitals and the actual outcome of treatment at those facilities.

"Hospital performance measures predict small differences in hospital mortality rates," said the study led by Rachel M. Werner of the Philadelphia Veterans Affairs Medical Center. "Efforts should be made to develop performance measures that are tightly linked to patient outcomes." Published Dec. 13 in the Journal of the American Medical Association, the study said many other factors such as hospital staffing levels shape outcomes.

An "ideal" process measure is a "reliable surrogate for outcomes," Berenson said, giving as an example the measurement of hemoglobin A1C levels in diabetics. But many process measures do not fit that description, he suggested.

But Gail Wilensky, who headed the Medicare and Medicaid programs under President George Herbert Walker Bush, responded that changes in basic payment systems are no simple matter and involve many practical difficulties.

Robert Galvin, who oversees health programs at General Electric, suggested that Berenson's critique offers the kind of debate that is needed over performance measures, but said the payoff from P4P in the private sector should not be minimized.

Galvin, who noted that there are some 150 P4P programs now in the private sector, said that a "Bridges to Excellence" P4P program, created by a coalition of business, quality measurement and other organizations, has produced good results in treating diabetic patients. Doctors who score well on measures of diabetic care are getting paid up to $20,000 more per year, patients are getting better treatment and payers are saving money, he said.

"If we know something works, let's do it," he said.

Stark said in a recent speech, "The entire concept of pay for performance is offensive. We shouldn't ever expect anyone to get paid more for doing what they were . . . paid to do." Stark added that "quality should be expected from each and every provider. And my solution would be to the provider who can't provide quality care, to defrock 'em."

Berenson offered a more measured view Friday. Measures are relatively well developed for dialysis centers and managed care plans in Medicare, he said. P4P for hospitals rates a three on a scale of zero to five, he added. But P4P is not well suited to primary care doctors and specialists, he added.

Publication Details