In theory, paying for performance makes logical sense, says Ashish Jha, MD, MPH, director of Harvard's Global Health Institute. But eliminating pay-for performance programs isn't the answer, he says. Instead, an overhaul is necessary. From MedPage Today.
This article first appeared July 7, 2016 on the MedPage Today website.
by Shannon Firth
Should new pay-for-performance programs be improved and expanded on, or are such models simply ill-advised? Health policy experts debated value-based payment models at a recent webinar.
In theory, paying for performance makes logical sense, explained Ashish Jha, MD, MPH, director of Harvard's Global Health Institute in Boston, at the July webinar hosted by the Annenberg Center for Health Journalism at the University of Southern California in Los Angeles, and sponsored by the National Institute of Healthcare Management in Washington.
Take two hospitals, one with a high death rate after heart attacks, and another with a lower death rate after heart attack, but only after the numbers have been risk adjusted. Without incentives or penalties, both are paid the same by Medicare, he said.
"There's been a sense for a very long time, that that's just not right," Jha stated.
One early attempt to enhance care by rewarding performance was the Premier Hospital Quality Incentive Demonstration program of 2003. The program placed 2% of Medicare hospital payments at risk, mainly by measuring clinicians adherence to process measures.
But after 6 years, researchers evaluating the program found "no impact whatsoever on mortality," Jha said.
Despite these weak results, new pay-for-performance models were rolled out as part of the Affordable Care Act.
Under the Hospital Readmissions Reduction Programs (HRRP), which began in 2011, Medicare penalizes hospitals as much as 3% of their Medicare payments for "higher than expected" readmission rates for a handful of conditions, including heart attacks, heart failure, and pneumonia. Chronic obstructive pulmonary disease (COPD) and hip replacement were added later.
About 80% to 90% of hospitals were penalized under the program, but there were some positive results as readmission rates for targeted conditions fell from 21.5% in 2007 to 17.8% in 2015 -- a drop of nearly 4 percentage points -- according to a study in the New England Journal of Medicine.
However, most of the hospitals with the highest readmission rates were safety net hospitals, or facilities with a higher proportion of black patients, Hispanic patients, poor patients, poorly educate patients, and Medicaid patients. This left many policy experts questioning the risk adjustment abilities of such programs.
The results of the Hospital Value-Based Purchasing program, which tied Medicare payment to a broad spectrum of quality measures -- some process measures as well as outcomes, efficiency, and patient experience metrics -- mirrored those of the Premier program.
"In the 3 years or so that [Hospital Value-Based Purchasing] has been up and running there, has been no impact on mortality," Jha said.
In January 2015, the Secretary of the Department of Health and Human Services (HHS) announced a goal of linking 85% of all traditional Medicare payment to quality or value by 2016, and then 90% via incentive programs such as the Hospital Value Based Purchasing and the HRRP.
And now under the umbrella of physician payment reform, one of the two pathways established by the Medicare Access and CHIP Reauthorization Act (MACRA) known as the Merit-Based Incentive Payment Program(MIPS), looks to expand pay for performance (MACRA replaced the Sustainable Growth Rate Formula).
"The fundamental question is: If the old stuff didn't work, is this going to work any better?," Jha said.
He said eliminating pay-for performance programs isn't the answer, but an overhaul is necessary. He suggested that pay-for-performance models might be more effective if incentives were greater.
"Maybe 1, 2, and 3% aren't the right numbers?" he said.
Jha also suggested targeting a smaller number of outcomes, aligning metrics with things doctors and nurses are intrinsically motivated to do, and developing a more "nuanced approach" to safety net hospitals.
In comparing his view of pay-for-performance with Jha, Robert Berenson, MD, a fellow at the left-leaning Urban Institute in Washington, cited the old joke, "the food's terrible and the portions are too small."
Berenson disagreed with Jha's suggestion to scale up incentives.
He said that the science of performance measurement is still very young, and that there should be more investment in its research. While there are things that should be measured -- blood pressure, for example -- there are few uniform quality standards for doing so.
He also argued that one danger of pay-for performance models is "testing to the test," where clinicians acquire a kind of tunnel vision for the targeted metrics. He added that creating such focus on targeted metrics will "crowd out" physicians' professional obligations.
Even if targeted performance metrics do improve, one could see see an overall performance decline, he said. "In other words, we measure much less than what we care about."
Berenson stressed that opposing pay-for-performance does not mean accepting the idea of zero accountability. He said he believes metrics can be useful as part of comprehensive major quality improvement, taking the readmissions program as one example. But "we need to figure out what we're doing here."
This means deciding which measures are valuable, determining how to assess statistical significance, and learning to avoid unintended consequences, such as surgeons "gaming" the system by declining to operate on more challenging patients in order to reduce their mortality rates.
"We can't put the genie of performance measurement back in the bottle, even though some of us would like to ... so we probably should do it better than the way we're doing it now," Berenson pointed out.