No quality benefit seen in Medicare pay pilot

Washington -- Medicare's largest pay-for-performance pilot project has not produced any better quality gains than a private initiative that used no financial incentives, a recent study said.

Researchers from Duke University in Durham, N.C., analyzed quality scores for the treatment of acute myocardial infarction at 54 hospitals participating in both Medicare's Hospital Quality Incentive Demonstration and a joint American College of Cardiology/American Heart Assn. quality improvement initiative. They then compared these figures with those from 446 hospitals that participated only in the latter project, which unlike Medicare did not offer monetary bonuses to top performers.

The investigation revealed that when it came to treating heart attack patients, the hospitals enrolled in the ACC/AHA quality improvement project did not make any statistically significant gains in quality of care, patient outcomes or mortality by also participating in the Medicare pay-for-performance program. The Medicare project hospitals as a whole made improvements on all of the quality measures they reported to the government, but so did the facilities that participated in only the cardiac associations' initiative. The study appeared in the June 6 Journal of the American Medical Association.

"On the positive side, everybody got better," said Eric D. Peterson, MD, MPH, a Duke researcher and one of the study's authors. "It's just not clear whether pay-for-performance was needed."

Dr. Peterson said the study was well received in the medical community, which has expressed skepticism about the ability of pay-for-performance programs to drive quality of care higher than more traditional clinical improvement efforts by physicians and health systems. He said the conclusions were similar to those made by researchers at Baystate Medical Center in Springfield, Mass., and Tufts University School of Medicine in Boston, who in February published a Medicare-sponsored study. It showed that that pilot program participants' improvement in acute myocardial infarction care and four other clinical areas were not significantly greater than improvement among hospitals that simply reported data without participating.

The Centers for Medicare & Medicaid Services had a different take on the project's outcomes, as did Premier Inc., the organization that runs the more than 250 nonprofit hospitals that participated in the pilot during its first run from 2003 to 2006. Premier published its own analysis earlier this year showing significant improvements among participants but without comparing them to gains made by other hospitals. CMS officials declared the pay-for-performance program a rousing success before the third-year results had even come in. The agency renewed the project for an additional three-year run and said it could serve as a model for the physician side of Medicare.

CMS could try increasing the project's bonus pool above the millions of dollars it has set aside already, but such a move likely would not result in participating hospitals outpacing those that do not take part, Dr. Peterson said. Too much reliance on financial incentives also could encourage some facilities to neglect areas of patient care not covered by the project, though the newest study suggests that this has not yet happened, he said.

The promise of pay-for-performance

Some physicians involved in the program remain convinced that pay-for-performance still has an important role to play in quality improvement.

Because the Duke researchers chose a control group of hospitals that also had committed to a formal quality improvement initiative, it was unlikely that any big improvement gap would emerge between the two groups of facilities, said Louis Teichholz, MD, chief of cardiology at Hackensack (N.J.) University Medical Center. His facility was one of the top performers in the Medicare pay-for-performance program's first run.

If anything, the results show that no matter the form that a quality improvement program takes, physicians and hospitals get into it for the right motivations, he said. "The reason they were doing it was to improve patient care, not because of the money."

Further investigation involving a broader scope of hospitals and clinical measures could uncover areas in which pay-for-performance prompted better outcomes than pay-for-reporting or initiatives that did not incorporate financial incentives, Dr. Teichholz said. Such results could become more pronounced now that CMS plans to pay smaller bonuses to the pilot's top tier of performers and larger ones to those who undergo the most relative improvement over time.

The newest study also does not address the possibility that the Medicare pilot indirectly improved quality among hospitals that didn't participate in it, said Glenn Crotty Jr., MD, chief operating officer at the Charleston (W.Va.) Area Medical Center. His facility was a top performer in the opening years of the program.

Competition between hospitals and the move toward more public reporting of quality scores are encouraging even physicians and hospitals that do not yet strive for pay-for-performance dollars to get into the same league as those who do, Dr. Crotty said.

"People are aware of the demonstration project whether they participate or not," he said. "What you're seeing is the result of a snapshot in time that shows nonparticipants catching up."