By David Blumenthal, M.D. and David Squires, The Commonwealth Fund
Twitter: @commonwealthfnd
The Obama administration recently announced that it plans to expand bundled payment models in the Medicare program. Bundled payments are one of several new payment alternatives in Medicare and Medicaid designed to hold health providers accountable for the cost and quality of care, and thereby encourage and reward better health care value.
All these so-called value-based payment initiatives, which include accountable care organizations (ACOs) and enhancements to older managed care programs under Medicare and Medicaid, are to some degree experimental and their full effects are still uncertain. However, the spread of bundled payments deserves particular scrutiny because of potential unintended consequences.
Why Bundled Payments?
Under bundled payments, a single payment is made for all of the services associated with an episode of care, such as a hip or knee replacement or cardiac surgery. Services might include all inpatient, outpatient, and rehabilitation care associated with the procedure.
There are a number of potential advantages to this payment approach. Bundled payments give providers strong incentives to keep their costs down, including by preventing avoidable complications. Bundled payments also can encourage collaboration across diverse providers and institutions, as well as the development and implementation of care pathways that follow evidence-based guidelines. In addition, bundles target the work of specialists, who have been less likely up to now than primary care physicians to participate in value-based payment arrangements.
More conceptually, health care economists are drawn to bundled payments because a bundle of care constitutes a clinically and intuitively meaningful “product” — in this case, the clinical episode. Defining clear products in health care helps create markets in which providers directly compete on quality and price. One barrier to effective health care markets has been that prices, when available, tend to relate to inputs into clinical care — such as pills, bandages, bed days, or X-rays — that are not meaningful to consumers of care and that don’t necessarily predict the total costs of care. For example, a health system that charges a lot for X-rays may still be more efficient because it uses fewer of them or saves money on other inputs. Bundles bring all these inputs together into a single price for a single basket of services.
What Are the Potential Problems?
Yet bundled payments have drawbacks. First, it can be complicated to define and track the type of care that should be included in the bundled payments for which a given provider is at risk. Knee and hip replacements are well-suited to bundles because they often involve comparatively young patients who are physically active (often the source of their joint damage) and want to remain so. But when patients have multiple chronic conditions that interact with each other, it becomes less clear whether the bundle should include the costs of caring for all those problems. When a hip patient has asthma and diabetes, for example, there may be a tendency for orthopedists to try to shift costs from a hip bundle to another specialist’s asthma or diabetes bundle. Monitoring the fairness of these interactions could become burdensome and increase administrative costs.
Second, as the hip patient example suggests, bundles could inhibit certain types of care coordination, even as it encourages other types. On the plus side, bundles may encourage hospitals to work more closely with rehab centers. On the negative side, bundles may encourage specialists’ already strong tendency to see patients not as whole individuals, but as single disease problems or procedures, and to diminish their sense of responsibility for costs of illnesses not included in their particular bundled payment.
Third, bundled payments could encourage destructive competition for patients with profitable bundles. The otherwise healthy patient needing a knee replacement may prove more profitable than a knee replacement patient with complicating problems such as heart, lung, or kidney disease. While risk adjustment could somewhat compensate for cherry-picking, such adjustments have not proven foolproof in the past, and an entirely new fleet of risk adjusters that are specific to given clinical episodes will likely be required. Monitoring the work of multiple risk adjusters and possible gaming by providers could become yet one more administrative expense.
Finally, bundled payments may make it harder for population-based payment methods like ACOs to be successful. Providers who participate in ACOs assume responsibility for all the care their patients need during a given period of time, including specialty care. This general accountability for their patients’ health encourages efforts to coordinate care, especially for complex patients. Still, to be financially viable, ACOs must generate savings from existing services. If independent specialty providers capture the elective procedures for which savings are easiest to generate through bundled payments, it could be harder for ACOs to find those savings within their own service mix.
Proceed with Caution
So how to move forward? At a minimum, the Centers for Medicare and Medicaid Services should make sure that the current bundled payment experiments don’t undermine ACOs and other population-based models. This will mean taking a close look at the rules governing who is responsible when a patient is attributed to both an ACO and a bundled payment.
The ideal solution may be to encourage and support the use of bundles within ACOs and other risk-bearing organizations that assume broad responsibility for their patients’ health. One of the most well-known examples of bundled payments — ProvenCare, which was originally marketed as a “warranty” on coronary artery bypass graft surgery — was developed and implemented within the integrated Geisinger Health System. If ACOs increase the number of provider organizations with financial incentives like Geisinger’s, bundled payments may spread rapidly on their own — and without many of the current drawbacks.
This article was originally published on The Commonwealth Fund Blog.