Defective rewards: How design flaws have hobbled Medicare's incentive programs

Hebrew Home at Riverdale in New York City eagerly joined Medicare's 2009 test of cash incentives for nursing home quality and quickly spotted a way to avoid unnecessary hospital visits for frail residents. Five years later, the number of Hebrew Home residents who land in the hospital for anemia has plunged 74%.

Under the three-year demonstration that ended in 2012, Hebrew Home began to routinely identify anemic residents who could skip the hospital and instead get outpatient blood transfusions. Fragile residents benefited by avoiding risky hospital stays. And Medicare saved, paying an average of $567 for care that would have cost about $10,000 if all the costs of a hospital stay were factored in.

Yet Hebrew Home did not receive a dime in Medicare incentive bonuses, despite success reducing unnecessary medical costs and strong performance on quality. That's because the demonstration required all 78 participating New York state nursing homes to collectively save at least $9.6 million or none received rewards. The New York nursing homes did not achieve that combined savings target.

This illustrates the problem of poorly designed financial incentive programs. The basic problem with the design was that Hebrew Home could not control whether its effort would produce a reward.

The demonstration also found that nursing home quality largely did not improve when compared with nursing homes that did not get incentives. The number of avoidable hospital visits and the number of safety violations were no better. Nursing homes showed no improvement in removing catheters on time. Patients were just as likely to get bed sores, which are almost always preventable.

The results, though disappointing to those who hope to see gains in nursing home quality, did not surprise researchers who study financial incentive programs. The demonstration did not incorporate many of the lessons that have been learned about how to structure incentives, said David Grabowski, a Harvard University professor of healthcare policy who evaluated the program for federal officials. “They broke a lot of the rules,” he said.

The CMS, private payers, policymakers and provider systems are placing heavy reliance on using financial incentives to change provider behavior to improve quality of care and reduce costs. In January, Obama administration officials said the CMS would significantly expand use of financial incentives in the next three years. Half of what the traditional Medicare program spends will be tied to rewards for quality and cost control by 2018. That will mean continued growth of accountable care and bundled-payment contracts. Other incentives that penalize or reward hospitals for quality and safety, known as value-based purchasing, will be linked to 90% of Medicare spending by 2018.

But researchers say much depends on properly structuring the incentive programs to achieve the desired results. Public policy inside and outside healthcare is strewn with incentive programs that have fallen short of producing the desired results. The consequences of poorly designed incentives can be higher spending and lower quality care. For evidence, look no further than Medicare's fee-for-service payment model, which has been widely blamed for encouraging greater volume of services, causing higher spending and exposing patients to the risks of unnecessary and disjointed care.

The CMS did not provide a comment for this article.

A financial incentive program works best when there is a strong relationship between effort, performance and reward, said Andrew Ryan, an associate professor of health management and policy at the University of Michigan. “When you look at the incentive programs out there, so many of them fail this basic idea,” Ryan said. Incentives also must be large enough to make people take notice, but not so big that incentives overshadow other priorities. In addition, programs must measure the quality and outcome factors that matter most.

In the case of the nursing home demonstration, the relationship between effort, results and reward was too tenuous, Grabowski said.

The design flaws in the Medicare incentive programs aren't entirely the CMS' fault. Ryan said the flaws can be attributed in part to the legal requirement that CMS demonstrations either reduce federal costs or at least curb spending growth. The link between effort and reward is weakened because rewards and penalties must be budget-neutral. Thus, participants don't know until all results are in what their performance will be worth financially.

Now, the CMS is preparing to implement new Medicare financial incentives for skilled-nursing facilities nationally. That program will take effect in 2019. The proposal for value-based purchasing at nursing homes is different from Medicare's prior nursing home initiative in key ways. Medicare will withhold a percentage of payments from SNFs that can be earned back based on each facility's performance relative to other facilities' scores. The program will put 2% of payment on the line.

But some experts say that initiative has a potentially fatal design flaw—the financial incentive may be too small. How much of an incentive to offer has proven tricky for Medicare's recent efforts. Too little and no one pays attention. Too much and the potential for unintended consequences increases. “Nobody knows where that sweet spot is,” said Dr. Rachel Werner, an investigator for the Veterans Health Administration's Center for Health Equity Research and Promotion who studies incentives and provider behavior.

The amount Medicare has relied on for various incentives probably isn't right, Werner said. “Two percent is probably too low,” she said. “We have lots of experience with 2% and it doesn't work.”

Dr. David Gifford, senior vice president of quality and regulatory affairs for the American Health Care Association, a trade group for post-acute providers, agreed that 2% might not be enough to get nursing homes' attention.

The size of the financial incentive was an issue in Medicare's Premier Hospital Quality Incentive Demonstration, which ran from 2003 to 2009. Hospitals in the Premier test were eligible for incentives of 1% to 2%. Quality improved at first when compared with hospitals not eligible for incentives. But after five years, the two groups' scores were virtually identical, according to Werner and her colleagues who studied the program.

Medicare's subsequent mandatory effort, the hospital value-based purchasing program that's been underway since 2012, gradually increases the incentive from 1% to 2% by 2017. One study found no change in quality or patient experience in the early months of the incentives compared with the five years earlier. Treasure Valley Hospital in Boise, Idaho, received the largest award measured by percentage in 2013—a total of $6,424, amounting to 0.8% of its Medicare revenue.

The designers of the hospital value-based purchasing program have faced another tough challenge—how to select the best measures for performance. Too often, incentives for quality are tied to performance on irrelevant measures. The measures selected often are those that are easiest to compile, said Dr. Ashish Jha, a Harvard University health policy professor who studies quality improvement. “We often don't pay attention to quality,” he said. “And when we do, we don't measure the right thing.”

Jha's research measured mortality rates for hospitals inside and outside the Premier test of quality incentives and found no difference among patients hospitalized for pneumonia, congestive heart failure, heart bypass surgery or acute myocardial infarction. Nevertheless, the CMS expanded the effort nationally.

Similarly, Jha said Medicare's test of accountable care does not measure the results that matter most to patients, such as how well patients function in their daily lives after treatment. That's also true for the CMS incentive effort using bundled payments for hip and knee surgery. “They want to be able to return to the things they used to do,” he said. “We shouldn't have a bundled payment program unless we measure what matters.”

In the past, Medicare has had more potent and successful programs changing financial incentives. Until the early 1980s, Medicare paid hospitals based on their costs, giving providers little incentive to improve efficiency. In 1983, Medicare shifted to paying hospitals fixed prices through diagnosis-related groups, which later were adjusted for patients' severity of illness.

Hospitals responded. The length of hospital visits dropped. “The decision of how we treat people is incredibly responsive to price,” said David Cutler, a Harvard University economics professor and health policy expert.