How to Pay for Health Care

Michael E. Porter and Robert S. Kaplan | Harvard Business Review | July–August 2018

The United States stands at a crossroads as it struggles with how to pay for health care. The fee-for-service system, the dominant payment model in the U.S. and many other countries, is now widely recognized as perhaps the single biggest obstacle to improving health care delivery.

Fee for service rewards the quantity but not the quality or efficiency of medical care. The most common alternative payment system today—fixed annual budgets for providers—is not much better, since the budgets are disconnected from the actual patient needs that arise during the year. Fixed budgets inevitably lead to long waits for nonemergency care and create pressure to increase budgets each year.

We need a better way to pay for health care, one that rewards providers for delivering superior value to patients: that is, for achieving better health outcomes at lower cost. The move toward “value-based reimbursement” is accelerating, which is an encouraging trend. And the Centers for Medicare & Medicaid Services (CMS), to its credit, is leading the charge in the United States.

That doesn’t mean, however, that health care is converging on a solution. The broad phrase “value-based reimbursement” encompasses two radically different payment approaches: capitation and bundled payments. In capitation, the health care organization receives a fixed payment per year per covered life and must meet all the needs of a broad patient population. In a bundled payment system, by contrast, providers are paid for the care of a patient’s medical condition across the entire care cycle—that is, all the services, procedures, tests, drugs, and devices used to treat a patient with, say, heart failure, an arthritic hip that needs replacement, or diabetes. If this sounds familiar, it’s because it is the way we usually pay for other products and services we purchase.

A battle is raging, largely unbeknownst to the general public, between advocates of these two approaches. The stakes are high, and the outcome will define the shape of the health care system for many years to come, for better or for worse. While we recognize that capitation can achieve modest savings in the short run, we believe that it is not the right solution. It threatens patient choice and competition and will fail to fundamentally change the trajectory of a broken system. A bundled payment system, however, would truly transform the way we deliver care and finally put health care on the right path.

The Small Step: Capitation

Capitation, or population-based payment, is not a new idea. It was introduced in the United States with some fanfare in the 1990s but quickly ran into widespread criticism and was scaled back significantly. Today, a number of transitional approaches, including accountable care organizations (ACOs), shared savings plans, and alternative quality contracts, have been introduced as steps toward capitation. In the ACO model, the care organization earns bonuses or penalties on the basis of how the total fee-for-service charges for all the population’s treatments during the year compare with historical charges. In full capitation, the care organization absorbs the difference between the sum of capitation payments and its actual cost.

Under capitation, unlike in the FFS model, the payer (insurer) no longer reimburses various providers for each service delivered. Rather, it makes a single payment for each subscriber (usually per patient per month) to a single delivery organization. The approach rewards providers for lowering the overall cost of treating the population, which is a step forward. However, under this system cost reduction gravitates toward population-level approaches targeting generic high-cost areas, such as limiting the use of expensive tests and drugs, reducing readmissions, shortening lengths of stay, and discharging patients to their homes rather than to higher-cost rehabilitation facilities. As a response to the failed experience with capitation in the 1990s, current capitation approaches include some provider accountability for quality. However, “quality” is measured by broad population-level metrics, such as patient satisfaction, process compliance, and overall outcomes such as complication and readmission rates.

This all seems good at first blush. The trouble is that, like the failed FFS payment system, capitation creates competition at the wrong level and on the wrong things, rather than on what really matters to patients and to the health care system overall.

Providers are not accountable for patient-level value.

Capitation and its variants reward improvement at the population level, but patients don’t care about population outcomes such as overall infection rates; they care about the treatments they receive to address their particular needs. Outcomes that matter to breast cancer patients are different from those that are important to patients with heart failure. Even for primary and preventive care, which the concept of population health rightly emphasizes, appropriate care depends heavily on each patient’s circumstances—health status, comorbidities, disability, and so on. And managing the overall health of a diverse population with high turnover (as ACOs do) is extremely difficult.

Thus, capitated payments are not aligned with better or efficient care for each patient’s particular condition. Instead, capitation puts the focus on limiting the overall amount of care delivered without tying the outcomes back to individual patients or providers. The wrong incentives are created, just as is the case for fee for service, which reimburses for the volume of services but not the value.

Providers bear the wrong risks.

Because capitation pays providers a fee per person covered, it shifts the risk for the cost of the population’s actual mix of medical needs—over which they have only limited control—to providers. Some large private insurers favor capitation for just this reason. But bearing the actuarial risk of a population’s medical needs is what insurers should do, since they cover a far larger and more diverse patient population over which to spread this risk. Providers should bear only the risks related to the actual care they deliver, which they can directly affect.

A more fundamental problem is that capitation payments are extremely difficult to adjust to reflect each patient’s overall health risk, not to mention to correctly adjust for this risk across a large, diverse population. Risks are much better understood and managed for a particular medical condition—for example, the probable effects of age or comorbidities on the costs and outcomes for joint replacement—as is the case in bundled payments.

Because population-level risk factors are so complex, health systems under capitation have an incentive to claim as many comorbidities as possible to bolster their revenue and profitability. A whole segment of health care IT providers has emerged to help providers “upcode” patients into higher-risk categories. Such gaming of risk adjustment first became a problem during the era of managed-care capitation in the 1990s, and it remains one today.

Patient choice is limited, and competition is threatened.

Capitation creates strong incentives for a health system to deliver all the care within its system, because contracting for outside services reduces net revenue and results in underutilization of existing internal capacity. There is even a term for this in health care—“avoiding leakage”—and many systems explicitly monitor and control it. Capitated health systems encourage or require patients (and their referring doctors) to use in-house providers (the ultimate narrow network). Patients are often penalized with extra fees when they don’t use services within the system, even if outside providers have greater experience and get better results for treating the patient’s particular condition. Capitation creates, in essence, a monopoly provider for all the patients in the population. Consumers cannot choose the best provider for their particular needs.

Since providers now bear actuarial risk, they also have a strong incentive to amass the largest possible population. This will accelerate the recent trend of providers’ buying up other hospitals and physician practices and merging systems, which reduces competition. To offset health systems’ rising bargaining power, insurers will feel pressure to merge. The two dynamics will reinforce each other as provider consolidation begets even more insurer consolidation.

The end result will be the emergence of a few dominant systems—or even only one—in each region. This would be bad for patients. No one organization can have all the skills and technologies needed to be the best in treating everything. We need multiple providers in each region to ensure enough choice and drive innovation in care delivery.

The bottom line is that capitation is the wrong way to pay for health care. It is a top-down approach that achieves some cost savings by targeting low-hanging fruit such as readmission rates, expensive drugs, and better management of post-acute care. But it does not really change health care delivery, nor does it hold providers accountable for efficiency and outcomes where they matter to patients—in the treatment of their particular condition. Capitation’s savings also come at the high cost of restricting patient choice and inhibiting provider competition.

Let’s consider the alternative.

Paying for Value: Bundled Payments
For virtually all types of products and services, customers pay a single price for the whole package that meets their needs. When purchasing a car, for example, consumers don’t buy the motor from one supplier, the brakes from another, and so on; they buy the complete product from a single entity. It makes just as little sense for patients to buy their diagnostic tests from one provider, surgical services from another, and post-acute care from yet another. Bundled payments may sound complicated, but in setting a single price for all the care required to treat a patient’s particular medical condition, they actually draw on the approach long used in virtually every other industry.

Bundled payments have existed in health care for some time in isolated fields such as organ transplantation. They are also common for services that patients pay for directly, such as Lasik eye surgery, plastic surgery, and in vitro fertilization.

To maximize value for the patient, a bundled payment must meet five conditions:

Payment covers the overall care required to treat a condition.
The bundled payment should cover the full cost of treating a patient over the entire care cycle for a given condition or over time for chronic conditions or primary care. The scope of care should be defined from the patient’s perspective (“Delivering a healthy child”). Care should include all needed services, including managing common comorbidities and related complications. In primary and preventive care, bundled payments should include all the needed care for each defined patient segment (such as healthy adults or low-income elderly).

Payment is contingent on delivering good outcomes.
Bundled payments should be tied to achieving the outcomes that matter to patients for each condition and primary care patient segment. Important outcomes include maintaining or returning to normal function, reducing pain, and avoiding and reducing complications or recurrences.

Payment is adjusted for risk.
Differences in patients’ age and health status affect the complexity, outcomes, and cost of treating a particular condition, as do their social and living circumstances. These risk factors should be reflected in the bundled payment and in expectations for outcomes to reward providers for taking on hard cases.

Payment provides a fair profit for effective and efficient care.
A bundled payment should cover the full costs of the necessary care, plus a margin, for providers that use effective and efficient clinical and administrative processes. It should not cover unnecessary services or inefficient care.

Providers are not responsible for unrelated care or catastrophic cases.
Providers should be responsible only for care related to the condition—not for care such as emergency treatment after an accident or an unrelated cardiac event. The limits of provider responsibility should be specified in advance and subject to adjudication if disputes arise. Bundled payments should also include a “stop loss” provision to limit providers’ exposure to unusually high costs from catastrophic or outlier cases. This reduces the need for providers to build such costs into the price for every patient (unlike in capitation).

How Bundled Payments Will Transform Patient Care
Decades of incremental efforts to cut costs in health care and impose practice guidelines on clinicians have failed. Bundled payments directly reward providers for delivering better value for the patient’s condition and will unlock the restructuring of health care delivery in three crucial ways that capitation cannot.

Integrated, multidisciplinary care.
Specialty silos have historically led to fragmented, uncoordinated, and inefficient care. With bundled payments, providers with overall responsibility for the full care cycle for a condition will be empowered and motivated to coordinate and integrate all the specialists and facilities involved in care. Clinical teams (the experts) have the freedom to decide how to spend the fixed bundled payment, rather than being required to deliver the services that are reimbursed by legacy FFS payments in order to receive revenue. Teams can choose to add services that are not currently covered by FFS but that provide value for patients.

Bundled payments are triggering a whole new level of care innovation. For example, hospital-based physicians are remaining involved in care after patients are discharged. Hospitalists are added to teams to coordinate all the inpatient specialists involved in the care cycle. Nurses make sure patients fill their prescriptions, take medications correctly, and actually see their primary care physician. (A recent study showed that 50% of readmitted patients did not see their primary care doctor in the first 30 days after discharge.) And navigators accompany patients through all phases of their care and act as first responders in quickly resolving problems. Bundled payments are also spurring innovation in the creation of tailored facilities, such as those of Twin Cities Orthopedics (Minneapolis), which performs joint-replacement care in outpatient surgery centers and nearby recovery centers, rather than in a traditional hospital.

Bundled payments will accelerate the formation of integrated practice units (IPUs), such as MD Anderson’s Head and Neck Center and the Joslin Diabetes Center. IPUs combine all the relevant clinicians and support personnel in one team, working in dedicated facilities. Joslin, for example, brings together all the specialists (endocrinologists, nephrologists, internists, neurologists, ophthalmologists, and psychiatrists) and all the support personnel (nurses, educators, dieticians, and exercise physiologists) required to provide high-value diabetes care. IPUs concentrate volume of patients with a given condition in one place, allowing diagnosis and treatment by a highly experienced team. Numerous studies show that this approach leads to better outcomes and greater efficiency (including less wait time and fewer visits). Bundled payments also encourage the formation of “virtual” IPUs, where even separate practices and organizations actively collaborate across inpatient and outpatient settings to coordinate and integrate care—something that rarely happens today.

Accountability for outcomes.
By definition, a bundled payment holds the entire provider team accountable for achieving the outcomes that matter to patients for their condition—unlike capitation, which involves only loose accountability for patient satisfaction or population-level quality targets.

Because bundled payments are adjusted for risk, providers are rewarded for taking on difficult cases. With a fixed single payment, they are penalized if they overtreat patients or perform care in unnecessarily high-cost locations. And because providers are accountable for outcomes covering the entire care cycle, they will move quickly to add new services, more-expensive interventions, or better diagnostic tests if those will improve outcomes or lower the overall cost of care. Specialists operating under a bundled payment, for example, have added primary care physicians to their care teams to better manage the overall care cycle and deal with comorbidities.

Most important, the accountability built into bundled payments will finally bring to health care the systematic measurement of outcomes at the condition level, where it matters most. We know from every other field that measuring and being accountable for results is the most powerful driver of innovation and continuous improvement.

Cost reduction.
There have been repeated efforts to control health costs for decades without success, and top-down cost reduction initiatives have sometimes increased costs rather than reduced them. The core problem is that legacy payment models such as FFS have given providers no incentive to cut costs or even to understand what their costs are for treating a given condition. Bundled payments, by contrast, directly reward and motivate cost reduction from the bottom up, team by team. At the same time, they encourage accurate cost measurement not only to inform price setting but to enable true cost reduction.

Bundled payments will be the catalyst that finally motivates provider teams to work together to understand the actual costs of each step in the entire care process, learn how to do things better, and get care right the first time. By encouraging competition for the treatment of individual conditions on the basis of quality and price, bundled payments also reward providers for standardizing care pathways, eliminating services and therapies that fail to improve outcomes, better utilizing staff to the top of their skills, and providing care in the right facilities. If providers use ineffective or unnecessary therapies or services, they will bear the cost, making bundled payments a check against overtreatment.

The result will be not just a downward “bend” in the cost curve—that is, a slower increase—but actual cost reduction. Our research suggests that savings of 20% to 30% are feasible in many conditions. And, because bundled payments are contingent on good outcomes, the right kind of cost reduction will take place, not cost cutting at the expense of quality.

Overcoming the Transition Challenges
Despite the now proven benefits of well-designed bundled payments, many hospital systems, group purchasing organizations, private insurers, and some academics prefer capitation. Bundled payments, they argue, are too complicated to design, negotiate, and implement. (They ignore the fact that capitation models continue to rely on complex, expensive fee-for-service billing to pay clinicians and to set the baseline for calculating savings and penalties. Bundled payments are actually simpler to administer than the myriad of FFS payments for each patient over the care cycle.)

Skeptics raise a host of other objections: The scope of a condition and care cycle is hard to define; it is unrealistic to expect specialists to work together; the data on outcomes and costs needed to set prices are difficult to obtain; differences in risk across patients are hard to assess, which will lead to cherry-picking; and bundled payments won’t rein in overtreatment.

If these objections represented serious barriers, we would expect to see little progress in implementing bundled payments and plenty of evidence that such programs were unsuccessful. To the contrary, bundled payments have a history of good results and are currently proliferating rapidly in a wide range of conditions, organizations, and countries.

In 2007, for example, the Netherlands introduced a successful bundled payment model for treating patients with type 2 diabetes, and, later, for chronic obstructive pulmonary disease (COPD). In 2009, the County of Stockholm, Sweden, introduced bundled payments for hip and knee replacements in healthy patients, achieving a 17% reduction in cost and a 33% reduction in complications over two years. More recently, Stockholm introduced bundled payments for all major spine diagnoses requiring surgery, and extensions to other conditions are under way there.

In 2011, Medicare introduced the voluntary Bundled Payments for Care Improvement (BPCI) program, which currently includes more than 14,000 bundles in 24 medical and 24 surgical conditions. Numerous physician practices have embraced the BPCI model, a transitional bundled payment approach that covers acute-care episodes and often a post-acute period of up to 90 days to promote better management of post-discharge services. According to participating providers, BPCI bundles have achieved significant improvements and savings an order of magnitude greater than savings from ACOs. Building on that success, CMS launched a mandatory bundled payment program for joint replacements in 2016, which covers 800 hospitals in 67 U.S. metropolitan areas.

Bundled payment contracts involving private insurers are also finally beginning to proliferate. For example, Twin Cities Orthopedics offers a bundle for joint replacement with most of the region’s major insurers at a price well below the traditional hospital models. The practice reports better outcomes and cost reductions of more than 30%.