Paying for care by the episodes may be the shortest path to value-based care. But administration is tricky, and early results raise questions about the savings achieved.
Health care researchers are typically a sober bunch, not given to angry online contretemps on the intricacies of their often dry and arcane topic. So it was a little surprising to see some strongly worded back-and-forth on the Health Affairs blog last year debating whether the concept of bundled payments has proven its worth.
A pair of Rand Corp. researchers had just published a review of a bundled payment demonstration in California that they said had “failed” because it was not able to attract enough physician participants. Their dismissal of the program got under the skin of the nation’s foremost proponent of bundled payments, François de Brantes, whose nonprofit Health Care Incentives Improvement Institute (HCI3) has been a leading innovator in payment reform.
De Brantes complained in his own blog that the Rand review reflected “vague conjecture and the deliberate choice to ignore the truth and spread falsehoods.”
The Rand researchers felt compelled to fire back on the Health Affairs blog about “ad hominem” attacks. De Brantes got in the last word, responding in the comments that the review was unfairly harsh about the California program paying retrospectively rather than prospectively—most bundled payment programs at this point have not yet advanced to the risk-taking prospective model. He also complained that the Rand reviewers left the impression that bundled payments have not yet been successfully implemented, arguing that “health plans have now sufficiently operationalized bundled payments to manage thousands of bundles.”
DRGs, but better
The episode may reflect the fact that the stakes are high for the concept of bundled payments, the success of which is essential to the fundamental financial reforms of the health care system envisioned by the ACA. It is one of a handful of value-based payment initiatives that the Obama administration is counting on to show that the politically fraught law is working to reduce costs and improve quality of care, not just expanded insurance coverage. Last month, CMS proposed to require hospitals in 75 areas of the country to use bundled payments for hip and knee replacements.
A bundled payment moves beyond fee-for-service payment by combining the services for a given diagnosis, procedure, or event, known as an episode of care. It is similar conceptually to the diagnosis-related group (DRG) that CMS has used for years, but the DRG includes only inpatient care, while bundles can encompass a variety of providers. Bundles can be designed in many ways, by nearly any player in the health care system—insurers, hospital systems, medical groups, even employers.
In contrast to other value-based payment concepts such as ACOs and patient-centered medical homes, the idea of bundling the provider payments for an episode is more targeted and flexible because it is on a smaller scale, involving a given procedure or diagnosis and a smaller group of patients. Bundled payments are seen as a transitional payment method that moves the American health care system toward more efficient coordination of care without physicians having to learn complex financial models for managing all the care for a group of patients.
“Bundles are little bite-sized nuggets of episodes that you can plug into an ACO or a medical home,” says Elaine Daniels, who worked with bundles for 7 of her 20 years at Blue Cross Blue Shield of North Carolina before joining Aver, an informatics company. “To me, it’s an easy segue into managing costs in ACOs and medical homes.”
Rob Lazerow, a practice manager for the Advisory Board Co., agrees that the methods can be used together, noting that they have different purposes: “An ACO tries to reduce or avoid utilization, whereas bundled payments try to make an episode more efficient.”
Lili Brillstein, director of the episodes of care program for Horizon Blue Cross Blue Shield of New Jersey, sees the value-based payment models—bundles, ACOs, medical homes—as synergistic. “What’s really nice about the episodes approach is that it’s a microcosm of these bigger models, and it’s easier to get up and running. Episodes don’t require the same kind of infrastructure and you can get there fast, but the concepts are exactly the same. You create this medical neighborhood and have this cadre of providers who can deliver services effectively and efficiently across the spectrum of care.”
Proponents of bundling tend to be enthusiastic, even evangelical about the potential for bundling to finally take the health care system where it needs to go. “I would bundle everything if I could,” says Daniels, only partly in jest. They find it frustrating that some published studies have painted a less optimistic picture of how bundling works in actual practice.
Skepticism about bundled payments centers on concern that bundles won’t save much money if they are variations on fee for service and don’t require providers to take on risk. Critics also argue that bundling is administratively difficult. It requires one party to take financial responsibility and have the expertise to understand how much the whole episode should cost and how much each participant should earn. That expertise can be hard to come by, as can the information systems and algorithms required to carry it out.
How bundles work
Many of the early bundled payment initiatives have been retrospective with upside risk. The providers bill just as they have under fee for service, but when the episode of care is over there is a reconciliation process. If financial goals are met or exceeded, a bonus is paid. If they aren’t and the bundled payment arrangement included downside risk, a penalty is assessed. Andrei Gonzales, MD, a value-based payment expert at McKesson Health Solutions, says there’s often a lead provider, or “quarterback,” in bundled payment arrangements who earns the bonus or pays the penalty. For a bundled payment for a surgical episode, the quarterback would typically be the surgeon; for congestive heart failure, it would probably be a cardiologist. One issue with retrospective bundled payments is that the bonuses (and sometimes the penalties) may come many months after the actual care is delivered. The delay tends to diminish the psychology of reward, so the worry is that the bonuses may not be a strong enough inducement for providers to improve the quality of care or the outcome.
Better, most agree, is prospective payment, which offers a set amount for a given episode of care, and leaves it to the providers involved to manage it. This, however, requires the provider to be the “banker” that manages both the care and the payments to any other providers involved in the episode of care. The banker can be a hospital, a provider group, or even the insurer, depending on how the bundle is designed. Many physicians have been hesitant about stepping into that role.
Bundles of variety
Not every health care service is appropriate for bundling, says de Brantes. Some low-cost services don’t really need management to squeeze inefficiency out of them; others are more complex or unpredictable and are too risky to manage. That leaves a large middle ground to work with. The concept of bundling payments is fairly simple to grasp, but the design has many variables, so bundled payment pilots come in many varieties. Creating a bundled payment program requires a few essential design choices:
- Which diagnosis, procedure, event, or chronic disease is involved, and how are eligible patients defined.
- What period of time is covered—for a procedure, this could be just a week prior and a month after, or it could be a complete year of care for a chronic condition.
- Whether to include an insurance (stop-loss) component for unusually high but necessary spending.
- What services and providers are included.
- Whether providers are paid prospectively or retrospectively.
Bundles have the potential to be used for more long-term, chronic conditions, so long as the patient population and included services can be defined. De Brantes sees value in taking them further into more sophisticated incarnations, such as bundling patients with multiple comorbidities together—the connection being not their specific diagnoses, but the way providers need to organize to care for them.
Bundled payments make up less than 1% of all commercial insurance payments, according to a 2014 report by the Catalyst for Payment Reform, an employer group pushing for value-based payment. It’s unclear exactly how many bundles are currently in use. Lazerow at the Advisory Board is tracking a couple of dozen commercial insurer bundling pilots around the country, though there are undoubtedly more running under the radar. Blue Cross Blue Shield insurers in North Carolina and New Jersey are both enthusiastic adopters of bundles; the North Carolina Blue reported saving 8% to 10% on a bundled payment arrangement for knee replacements back in 2011.
In 2013, CMS launched the Bundled Payments for Care Improvement initiative, which offers 1 of 4 models and 48 different episodes to choose from. The agency says providers are currently participating in its bundling pilots. An evaluation of the initial participants released earlier this year was more an outline for future evaluation and drew few conclusions. Jeff Goldsmith, a health care consultant and member of the Managed Care editorial board, likes the “elegance” of the ideas behind bundled payments, but found the CMS results troubling because he didn’t see them as showing any cost savings. If you add in the report from the Rand researchers, says Goldsmith, there’s enough recent bad news about bundled payments to sow serious doubts about whether an attractive idea can be implemented successfully.
State Medicaid agencies are also innovating with bundled payments, particularly in Arkansas, which has launched an initiative involving bundled payments and medical homes with its Medicaid program, along with two major commercial insurers. Initial results showed that more providers gained from the program than lost, though there were substantial penalties for those that did not meet financial goals. It is hoped that will change as providers adjust their practice to meet financial and quality metrics.
Large national employers are experimenting with the idea, focusing on contracting with providers for specific services. For instance, Wal-Mart has a program for employees who need cardiac, spine and certain other surgery that encourages them to use 1 of 6 providers with high volumes and good outcomes.
To-do list for bundles
So, what needs to happen to bring bundling into the mainstream and convince skeptics? Several things:
Providers must become willing to take risk. That way, these models can go from retrospective to prospective and reap more savings. They also need to move ahead with hiring care coordinators, collecting data and sharing information with other providers in an episode of care.
Definitions must be agreed upon. Differences in how bundles are defined could cause trouble down the road, as providers have to manage several versions of a given episode of care. HCI3’s Prometheus project has provided a starting point for episode definitions, says Daniels, adding “but payers and providers need to work collaboratively to come to a comfort level on what a definition should be and what services should be included in that time period.”
Payers must invest in analytics systems to manage bundles and track savings and quality. Seventies-era information systems used by some Medicaid administrators and other payers are a real roadblock to managing innovative payment systems, argues de Brantes. And insurers run the risk of becoming irrelevant if providers contract directly with employers to act as centers of excellence.
Insurers and employers need to work together to design benefits to encourage consumers to use services packaged for efficiency and quality.
The insurers that have led the way into bundling see several advantages, beyond just the need to take part in an inevitable change in the marketplace. One important plus, says Brillstein, is that the process of designing bundles gives insurers and physicians an opportunity to work closely together to support high-quality, patient-centered care, which can improve a relationship that has often been adversarial. “These programs are changing the spirit of the relationship between providers and payers,” says Brillstein. “It’s very collaborative.”
Once providers get on board, payers have analytic and payment systems in place to manage bundles, and employers encourage patients to use bundled services (through steerage mechanisms), “the sky’s the limit,” says Daniels, who led early adoption at Blue Cross Blue Shield of North Carolina. “We’re right on the cusp, and you’re going to see a lot of stuff happening in the bundled space. The future of health care is going to be bundled payment.”
Texas Medicaid insurer bundles pregnant moms, bundles of joy
When Community Health Choice (CHC), a regional HMO in southeast Texas, decided to try a bundled payment pilot, pregnancies were the natural place to start. The nonprofit insurer covers 300,000 lives, two-thirds of them in Medicaid, and maintains a specialty with Children’s Health Insurance Program (CHIP) perinatal patients. CHC manages more than 25,000 births each year.
Karen Love, the HMO’s senior vice president for strategic planning and partnerships, had been hearing about bundled payments as a concept of increasing interest to the state Medicaid agency, and she started attending a yearly national bundled payments summit to learn more.
CHC ended up hiring the summit’s sponsor, the Health Care Incentives Improvement Institute (known as HCI3), to help launch a pregnancy-related bundle with two physician groups with an interest in trying new payment models. HCI3 helped the insurer examine claims data to provide target budgets and develop the pitch to the providers. “It’s helpful to have a third party facilitate when you’re having those conversations,” says Love.
The two-year program was launched in March with the first year including just upside risk, meaning that CHC will share any savings 50–50 with the physician groups at the end of year one if they come in under budget. The second year will involve some type of downside risk that has not yet been decided. It will also include quality metrics (such as rates of sepsis, breastfeeding, and postpartum depression screening) that are being benchmarked during the first year. The physicians helped choose the quality metrics and select their own methods of adjusting clinical care to meet the financial and quality goals.
CHC was interested in pursuing bundled payments as a new way to reduce medical costs while at the same time improving the quality of care for its pregnant patients. “We consider ourselves cutting-edge and like to do things that are novel and out there in the literature,” says Love.
If the pilot is successful, the insurer would like to have it inform future contracts with the provider groups, with a shared savings component built in; the concept could be expanded to other provider groups that do high volumes of particular medical services. CHC will also monitor quality measures and patient experience through surveys.
One unique thing about this bundle, Love says, is that it includes care and outcomes for the baby as well as the mother. “We really like that focus,” she says. “Those kiddos become ours down the line. We believe in that whole thread…improving birth outcomes leads to better health outcomes. The kid is healthier, gets a better education, a better job and someday doesn’t need Medicaid services at all.”