Bundling Bandwagon: Hold Your Horses

Bundled payments are gaining momentum as providers and payers begin exploring how to implement these models. Yet some policy analysts urge caution as they look beyond “how-to” aspects to longer-term consequences.   

No more surprises

Bundled payments hold broad appeal to everyone from patients to payers by clearly defining the costs, treatment and period of care for a procedure such as CABG or a condition such as diabetes. Unlike the fee-for-service model, bundles reward efficiency rather than volume by tying payments to all services within an episode of care. Unnecessary testing bites into profits, as does avoidable complications or other practices that may jack up expenses with no improvement in quality.

Some successful early demonstrations such as Geisinger Health Systems’ CABG bundle have added to bundling’s rising reputation. The Pennsylvania system’s program, which included best practices to reduce complications and other approaches, reduced hospital charges  by 5 percent in its first year after implementation (Ann Surg 2007;246[4]:613-621). In 2012, Geisinger reported the program had reduced 30-day readmissions by 20 percent with a 17.6 percent increase in profit margin.

It’s a win-win that some stakeholders want to expand to the outpatient setting, for instance for PCI, and chronic conditions such as congestive heart failure and coronary artery disease. According to a survey conducted in October 2012 by the management consulting firm Booz & Company, 51 percent of hospitals reported exploring involvement in bundling while 30 percent already were pursuing a bundled model. The online survey included more than 150 hospitals and 400 physicians.

The Centers for Medicare & Medicaid Services continues to test models that extend the tentacles of an episode of care, including post-acute care. And while most agree that the fee-for-service model needs to be replaced, some health policy analysts warn that the change will be gradual with risks of repeating past mistakes.

“No matter what system you come up with, there will be downsides,” says Stuart H. Altman, PhD, an economist at Brandeis University in Waltham, Mass. As chair of the Prospective Payment Assessment Commission for 12 years, he advised Congress and the White House about diagnostic related group (DRG) initiatives and other reforms. “The fee-for-service system has turned out to be very expensive and hospitals have gamed it galore.”

Lagging practice

Bundling differs from the prospective payment system in significant ways, but some of the problems that arose in the past could resurface with bundling. Altman offers lessons learned with prospective payment as guidance for making bundling a success (Health Affairs 2012;31(9):1923-1930).
The prospective payment system based the payment on the patient’s diagnosis and average costs to the hospital for resources needed to treat the illness. It introduced DRGs, with weights that reflected the use of more or less costly services. DRGs applied to hospitals but not physicians, who remained under a fee-for-service payment model.

Bundling also relies on a DRG code at admission but it includes the physician services component in its payment. Bundling’s reliance of DRGs may make it vulnerable to potential disconnects between the actual cost of care and payment, though. That may be particularly acute in specialties such as cardiology and its subspecialties, where payments may lag behind improvements based on devices, pharmaceuticals or techniques.

“To the extent that new types of valves or stents or techniques are introduced and they are cost saving, over time the cost of that average of all hospitals will come down and therefore the cost of that particular intervention will come down,” Altman observes. The process applies to hospitals as a whole. “The problem develops, though, when the cost goes up.”

He cited the use of tissue plasminogen activators (tPAs), which were introduced as an alternative to the much less expensive streptokinase as a treatment for MI when Altman chaired the commission. “The average was still very heavily weighted toward the cheaper drug,” he recalls. “Every time the hospital gave the more expensive drug, it was incurring costs in excess of what the DRG was paying in the bundled payment.” While tPA nonetheless became the favored therapy and costs eventually aligned with payment, “in the short run, it was a problem.”

The update hiccup

Special adjustments can be made to DRG weights, but often they remain the same. Instead, Medicare applies a conversion factor that is updated over time to account for changes in the cost of care. But when the cost of one type of service rises or drops faster than for another—yet the conversion factor is applied uniformly—some services may end up being underpaid and others overpaid.

“The challenge is going to be tailoring payment policy reforms so that we maximize the use of clinically beneficial things and minimize the use of those that provide less or no benefit,” says Allison B. Rosen, MD, MPH, ScD, at the University of Massachusetts Medical School in Worchester. “It is going to be about moving from a one-size-fits-all payment policy to a more tailored payment reform.”

Rosen and colleagues examined commercial claims between 2003 and 2007 to identify which clinical conditions contributed the most and least to per capita spending growth, and to see if using a uniform rate across episodes led to an imbalance that they called skewness.

In that period, spending per capita increased 15 percent and accounted for 73 percent of total spending growth. The number of episodes per person rose, too, which contributed to the remaining total spending growth. The majority of spending growth—82.5 percent—was traced to only 10 percent of episodes. Breast cancer was among the top contributors to spending growth, at 5.3 percent, while spending for angina and acute MI dropped by 5.9 percent and 2.3 percent, respectively.

Applying uniform updates across episodes in a bundled care model would prove lucrative for hospitals treating MI patients and would be a money loser for facilities offering breast cancer treatments, according to their analysis.

“In the case of PCI, when we first started doing a lot of these cardiovascular procedures, it took a lot longer to do them,” says Rosen, whose research interests include resource use in cardiovascular disease. “They were more costly because there was a learning curve effect. More could be done over time but the weights didn’t go down, [so] then there was a relative overpayment.”

Unintended consequences

Bundled payments provide financial incentives for hospitals to reduce costs, which should lower healthcare spending. They also encourage physicians to look beyond their individual contribution and coordinate care, which should lead to improved outcomes. But potentially the allure of profits could overpower care.
“Theoretically it could push a hospital to do things they shouldn’t do to save money,” Altman says. “There could be a potential for hospitals to skimp on activities that they don’t think are important but are.”

Judith Feder, PhD, of the Georgetown Public Policy Institute at Georgetown University in Washington, D.C., shares that concern. Feder, whose work focuses on post-acute care, sees skimping and patient selection as possible downsides. In post-acute settings such as skilled nursing homes, where patient classification is weak and practice is not supported by established guidelines, facilities may favor patients who need little care or provide insufficient care to those in need.

“A lot of the bundles include post-acute care,” Feder notes. “The hope behind paying a single rate for a bundle of services is that it will assure appropriate use across the mix of services and thereby promote efficient delivery of quality care. But achieving that goal requires strong norms for appropriate care and precise classification of patients into distinct groups that reflect their care needs. If your norms and classifications are weak and your incentives are powerful, you create enormous pressure for selection and skimping.”

Researchers offer several suggestions to keep the bundled payment approach on a path to success.  Feder proposes a temporary hybrid payment system, a combination of fee for service with incentives for cost savings in which Medicare and providers share savings and risk, until a robust system exists for measuring patient needs and outcomes.

Rosen, using PCI as an example, recommends payment reform that recognizes clinical nuance.  “A technological advance has very different health benefits, depending on the setting in which it is done,” she says, arguing that primary PCIs save live while non-primary PCIs don’t. “The extent to which we can get more and more nuanced tailoring to make sure we are paying for those things that are likely to save lives, the better off we are.”