Pay for Performance Still Not Showing Efficacy : The findings may show that financial incentives do not work for professionals as other research suggests.


SEATTLE — When the physicians of Rochester, N.Y., first had a pay-for-performance program imposed on them, they ignored it.

“At the beginning of our program, most people would not acknowledge it existed,” said Dr. Howard B. Beckman, the medical director of the Rochester Individual Physician Association (IPA). “As we talked about the profiles, people said 'I never got them,' 'I threw them away,' or 'I don't care.'”

That denial ended when the first performance-based checks were disbursed, and after 3 years, pay-for-performance measures have paid off in reduced health plan costs of almost $5 million, Dr. Beckman said at the annual research meeting of AcademyHealth.

Dr. Beckman was one of three physicians who presented research on whether pay for performance improves quality of care and efficiency in medicine enough to make worthwhile all the effort being put into it. He was the only one of the three to have a positive conclusion.

The other two investigations of pay for performance, in California and Massachusetts, looked more specifically at individual aspects of clinical care. Those investigators found they could not document an impact from the programs.

But those investigators also pointed out that, as in Rochester, it takes time for physicians to get accustomed to the idea of greater accountability, and to develop the capabilities to record and report for the programs, so their findings might reflect the fact that the programs have not been going long enough.

On the other hand, the findings may show that financial incentives do not work for professionals, something research in other fields has suggested, they noted.

The Rochester physicians went through stages of acceptance of pay for performance not unlike the stages of grief defined by Dr. Elisabeth Kübler-Ross, Dr. Beckman said.

After the first performance bonus checks were sent out and denial ended, there was anger. The physicians complained that strict performance measures impinge on their autonomy, and they were even offended by the implication that money could influence their behavior, he said.

Then, after about 2 years, the general resistance abated, and the angry phone calls stopped, Dr. Beckman said. Now when he gets phone calls about the program, it is an individual physician trying to negotiate something.

The Rochester IPA represents all 3,200 physicians in the Rochester area and has insurance contracts that cover about 50% of the community market. Its individual physician profiling program began in 2002.

The program's individual physician payments vary, but overall the program pays out about $15 million a year, and the average internist can earn from $4,000 to $12,000 from the quality reports. The physicians get three reports a year, and payments are made at the end of the year.

Dr. Beckman looked at the provider profile data for patients with diabetes and coronary artery disease. He found that when expected costs were compared with actual costs in the diabetes patients in 2003 and 2004, there was a savings of about $1 million in the first year and $2 million in the second year. Most of that savings, about $1.3 million, came from reduced inpatient hospitalization costs.

The savings for the coronary artery disease patients was about $2 million over the 2 years, for a total savings for just those two groups of patients of about $5 million, Dr. Beckman said. Given what the group had put into the program (about $1.1 million, mostly for computer capability), the return on investment for the program was about four times what was spent.

Dr. Beckman pointed out that many people have expressed concern that pay-for-performance programs could be unfair to physicians with the most difficult, least compliant patients, so he looked at different practices. It appeared that differences were greater between individual doctors than they were between practices and practice locations.

Pay for performance began in California at about the same time as the Rochester program, and it has yet to show any meaningful overall improvement in clinical care, said Cheryl L. Damberg, Ph.D., a researcher for the RAND Corp., who has been analyzing data from the California collaborative managed by the Integrated Healthcare Association, which includes seven HMOs and point-of-service plans contracting with 225 physician groups.

Surveys of patient satisfaction, a part of performance that is rewarded, showed gradual, substantive improvement in the first 2 years of the program. But when Dr. Damberg looked at clinical care measures, such as aspects of diabetes care, Pap smears, and childhood immunization, any improvement seen between years is inconsistent and varied.

She concluded, based on an analysis of the patterns of improvement, that many physicians and groups are getting up to speed with reporting, so it is too early to judge the impact on actual clinical care.


Next Article: