Physician quality pay not paying off

While offering physicians bonuses for hitting quality benchmarks is popular, the incentive programs may not be worth the money.

Linking financial rewards to cost-effective management of patient care or reducing adverse outcomes has not produced the desired results, recent studies show. When it comes to physician pay, some experts are asking if healthcare organizations are moving in the wrong direction.

“The programs are often less effective than the designers hoped for,” said Jessica Greene, associate dean for research at George Washington University. She conducted two studies of an ambitious physician incentive program at Minnesota-based Fairview Health Services. “There is still so much we don't know about how to design effective pay-for-performance programs.”

Behavioral economists and healthcare quality and management experts are urging provider organizations to take a second look at their payment models. Complex compensation designs, poor alignment of goals and lack of clearly defined, actionable measures can lead to failed efforts and unintended consequences, they say. Poorly aligned monetary motivations can even lead to difficulties with staff recruitment or retention and lead to over-focusing on one specific issue at the peril of other, more important ones.

Hospital and physician group leaders long have cited the difficulty of crafting physician payment models that encourage quality processes and outcomes while maintaining the incentive for high productivity. Experts are concerned that any new Medicare value-based system for paying doctors, which must be developed under recent legislation, will run into those same challenges, given the fledgling state of measuring the performance of individual physicians.

“The things that really matter in terms of medical quality are very difficult to measure,” said Dr. Michael Kirsch, a Cleveland-based gastroenterologist and author of a blog called MD Whistleblower. Value-based pay can drive healthy competition, but reliance on metrics that are easy to measure but don't ultimately boost outcomes is “a clumsy response to fee-for-service.”

The science of measuring quality performance in healthcare is still in its infancy. Current measures are limited, and critics say linking them to compensation might be premature. Improvements seen on easily tracked process measures such as checklist use or giving discharge instructions may not lead to improvements in patient outcomes such as lower mortality and lower readmission rates.

About 40% of U.S. healthcare providers had some type of incentive linked to pay in 2013, and within that group an average of more than 4% of total compensation was linked specifically to quality metrics, according to the MGMA. In the coming weeks, the group is expected to release data from its latest physician compensation survey, including more than 70,000 providers and more than 4,100 physician groups.

Advocates of quality incentive pay say the programs have lifted the performance of some physicians and improved collaboration among clinicians. That was the case for Fairview Health Services, which rolled out an ambitious compensation program in 2010 when it tied 40% of clinician pay to performance on a suite of metrics required by state law. Increases in salary were based on how well the medical group performed overall on state benchmarks for diabetes, cardiovascular and asthma care, and for certain evidence-based cancer screenings.

Yet even with the promise of more money, the model “didn't necessarily have an overwhelming impact,” said Valerie Overton, president for quality and innovation at Fairview Medical Group. It's not that quality did not go up at all, she said. It's just that it didn't go up any more than market competitors that had not instituted such a program.

The Fairview payment model also was a source of “significant frustration” among staff, said Greene who, along with Overton, was one of the co-authors of the two studies on Fairview's pay-for-performance effort.

Other studies on financial incentives have come to similar conclusions. In one, some primary-care physicians in New York were eligible to receive up to $200 per patient and up to $100,000 per clinic based on performance on evidence-based heart-care processes and outcome measures. But there were only small improvements despite the financial incentive, according to a 2013 report published in JAMA.

A program in which Houston clinics could receive twice the normal financial incentive given by Medicare for achieving cervical cancer screening, mammography and pediatric immunization targets also had little impact. “Despite considerable initial enthusiasm for the use of financial incentives for quality improvement, this study does not support the efficacy of this approach,” wrote the authors of a 2010 study of that program in the Journal of the American Board of Family Medicine.

Researchers in another study examined 30-day mortality rates among more than 6 million patients who had acute myocardial infarction, congestive heart failure or pneumonia and underwent coronary-artery bypass grafting. The hospital-based, pay-for-performance program linked to Medicare payments did not reduce deaths. The authors of the 2012 article in the New England Journal of Medicine concluded that expectations for similar programs “should remain modest.”

Not surprisingly, when the CMS released its third year of 30-day readmission penalties last fall, quality researchers said that if only 769 of more than 3,370 U.S. hospitals succeeded in avoiding the fines, that program may not be achieving its desired goal of broadly improving quality of care.

“There is essentially no evidence that pay-for-performance works, and certainly no evidence that it works as it is being applied to American healthcare right now,” said Dr. Steffie Woolhandler, a professor at the City University of New York's School of Public Health.

The tendency of pay-for-performance to “dangle money” before doctors has side effects. It turns the intrinsic professional and moral obligation of doing the best thing for the patient into a market transaction governed by price, and also requires excessive amounts of documentation and administrative costs. “If clinicians do have extra time, they should be focusing on real improvement and not just checking boxes to make pay-for-performance goals,” she said.

Still, there are lessons learned from the rollout of the value-based incentive programs. Fairview's project revealed, for example, that the compensation model was most effective with the poorest-performing physicians. Those in the lowest third improved on average three times more than those in the middle group and almost six times more than those in the top-performing group. It also led to greater collaboration between clinicians who were forced to work together to boost clinic-wide performance.

Thoughtful application of the design of pay-for-performance models is one key to their success. Systems with complex mathematical formulas that are difficult to understand don't do as well, said Dave Gans, senior fellow of industry affairs for the MGMA.

Indeed, last October, Fairview revised its model, which originally included complex formulas such as “a sliding scale of down to half the median income for performance between the 20th and 29th percentiles.”

Behavioral economists aren't surprised by the poor initial results from healthcare's pay-for-performance programs. Understanding the actual barriers to the desired behavior—such as lack of knowledge, motivation or resources—is crucial, said Dan Ariely, founder of the Center for Advanced Hindsight. His group studies how people make decisions and has looked at pay-for-performance in other fields such as education.

“Once you understand where the problem is, then you can try to solve it,” he said. “Pay-for-performance is a good idea for things where you can spell out exactly what success is.”