April 24, 2007

Survey: Doctors support linking pay to performance, wary of implementation

Doctors like the idea of using financial incentives to reward physicians for improving the quality of their care, but only a minority support the public reporting of the quality assessments used to determine the rewards, according to a new study from University researchers.

The survey, which was published in the March/April issue of Health Affairs, found that doctors are concerned that current measures do not accurately reflect quality of care. Additionally, the survey found that physicians do not believe the government or insurance companies would make an effort to correct existing measures. The researchers polled 1,168 general internists, of whom 556 responded.

The study asked doctors about their opinions on “pay-for-performance” programs, in which health insurance companies would pay physicians for improving patient care in an organized way. For example, providing mammograms on a more consistent basis requires more equipment and personnel, but in a “pay-for-performance” program health plans would help offset the costs of these improvements.

“There was a lot more support for financial incentives than public reporting,” said Lawrence Casalino, study author and an assistant professor of health studies, in an e-mail interview.

While the survey provided no direct evidence for why physicians felt so differently about incentives and reporting, the researchers have a hypothesis. “[A] relatively small fraction of their income would be based in financial incentives for quality.… But the idea of being publicly shamed if they didn’t score well, especially if the measures are perceived as unfair, is something that any physician would be quite upset about,” Casalino said.

Doctors were also concerned that “pay-for-performance” programs could actually have a negative effect on patients. Eighty-two percent of physicians worried that inaccurate measurements that do not adjust for the seriousness of a patient’s illness could cause doctors to shy away from high-risk cases.

“We have very strong evidence that doctors are afraid that physicians who take care of sicker, poorer, and less compliant patients would be punished. Some physicians said they would get rid of patients who lowered their scores,” Casalino said. One physician responded that he would have to drop 10–15 patients under such a program.

Overall, 73 percent of doctors support financial incentives if accurate quality measures are used, but only 30 percent feel that the current criteria are accurate. Thirty-eight percent believe “health plans will try hard to make quality measures as accurate as possible,” while 35 percent believe the government would improve those measures.

But even if measures correctly reflect physicians’ work, only 32 percent of the internists would support the public reporting of individual doctors’ quality assessments, while 45 percent favor reporting the scores for medical groups.

Casalino says while “pay-for-performance” programs are not widespread in the U.S., the situation might change in the next several years. Medicare is currently considering instituting a “pay-for-performance” program for its physicians; and since Medicare is the largest healthcare payer in the country, commercial health plans tend to follow its lead. “If [commercial plans] adopt it, it’ll become much more of a big deal” in the next three years, Casalino said.

He cautioned, however, that if companies try to institute “pay-for-performance”schemes without physician input, there may be a backlash from the medical community that could destroy the programs. “I think pay-for-performance and public reporting are probably good ideas, but they need to be done in a way to minimize unintended consequences. They’re not being designed that way now,” Casalino said.

The study was co-authored by Caleb Alexander, an assistant professor of general internal

medicine; Lei Jin, a programmer analyst for general internal medicine; and Tamara Konetzka, an assistant professor in health studies at the U of C.