Improving Value-Based Payment

The principle of “paying for value rather than volume” has been touted by health policy experts as one of the most important ways to improve health care and lower the rates of spending growth in the United States, where spending on health care at nearly 18% of GDP is higher than almost anyone wants it to be.

Despite the country’s history of high spending on health care—in the aggregate and per person—the health status of citizens has lagged behind that of other countries. This is not especially surprising since medical care doesn’t have nearly as much impact on health as the social determinants of health—education, income, employment, early childhood development, general living conditions, etc. However, changing the conditions in which people live, work, and die is even more challenging than changing health care.

Finding strategies that improve the value for money spent on health care has been a focus of both government and the private sector for the past decade. Despite all of these efforts, the results have been modest, although changes in hospital behavior have been easier to produce than changes in the behavior of physicians who practice inside and outside of them. Hospitals have been shown to change some processes—investments and programs—with only a 1% to 2% increase in their annual update factor, as occurred with the voluntary reporting of quality of care data, once hospitals were incentivized to do so. Physicians, however, seem to require much larger increases in payment in order to change their behavior. In both cases, it has proven even more challenging to bring about changes that affect either health outcomes or aggregate spending levels and growth.

Beyond a common invocation of value-based payment as a needed replacement for fee-for-service (FFS) payment, there has been nonpartisan consensus that a decade of more discussion and experimentation has yet to produce major improvement. Ashish Jha has argued for higher incentive payments to “reboot” efforts, while Alex Azar in 2018 opined that “I don’t want to spend the next several years trying to build the best joint replacement bundle—we want to look at bold measures that will substantially recast how Medicare and Medicaid will pay for care.”

A Biden administration that hopes to make effective changes in payment—most obviously for public payers but perhaps extending to consumers with private insurance—should, we argue, contemplate bolder and transformational approaches. That boldness will require two things: incentive rewards or penalties that are substantial, and a hard, risk-adjusted specification of quality goals.

In many past value-based payment programs, only a small percentage of the total payment was related to a value measure. This made it hard to know how much of the payment actually was at risk for low-value activities. For example, though it is reported that 34% of hospital payments are tied to value-based care, it is unclear how much of any individual payment to a hospital is at risk if lower quality of care is provided. Hospital managers typically translate incentives to physicians using internal payment systems still built on a volume or fee-for-service chassis. Direct payment to physicians for changes in how they treat insured patients seems to require even larger financial differentials (of the magnitude of the commercial insurance-Medicaid gap, for example), to get them to increase access and volume. Even for incentive programs that have been regarded as successful, such as the Medicare Hospital Readmission Reduction program, the readmission revenue received often exceeds their additional cost, so that there is still profit at the margin. Moreover, research has shown that changes in risk adjustment rather than actual readmission reductions were about half of the estimate of successful reduction. Finally, even the bundled payment program for joint replacement, one of the more successful changes, saved at most only a few percentage points of baseline cost.

Nevertheless, the desire to move more payment to a value basis continues at the Centers for Medicare and Medicaid Services (CMS). In September 2020, CMS provided a roadmap for states to increase the adoption of value-based care for Medicaid beneficiaries. However, the challenges of convincing states that it is in their interest to adopt such strategies in Medicaid is even more difficult than for other payers, especially because CMS is unable to provide evidence that any of the possibilities work. Another reason is that most of the effort of moving to value-based care will be incurred by the states and by providers while most of the savings, if any, would accrue to the federal government.

The potential for savings in the commercial sector from encouraging physicians to practice high-value care and to meet cost-efficiency standards remains significant, even if achievement of that potential remains elusive. According to estimates from UnitedHealth Group, the largest US health insurance company in terms of covered lives, if all physicians who care for commercially insured patients practiced high-value and cost-efficient care, there could be savings of over $70 billion a year over the period 2019-2028, that could be shared by their insurance customers and stockholders. Measuring cost and quality performance for 16 specialties using National Quality Forum and National Committee for Quality Assurance cost and efficiency measures from local market benchmarks, they estimated that higher-value physicians have lower rates of complications. This conclusion was based on stent-placement heart procedures and the experiences of patients who had more than 75% of their care provided by high-value physicians. These patients had 28% fewer emergency room visits, 33% fewer hospital admissions, and 24% lower risk-adjusted health care spending. The highest impact in terms of savings was seen for orthopedic and general surgery, rheumatology, and nephrology. However, because the volume of primary care is so high, the opportunities for savings there was $41 billion in 2019 and $507 billion over ten years.

The potential for savings, so attractive and yet so far out of reach, is frustrating to insiders and challenging to reformers. There is consensus that it most likely can be achieved by moving away from current FFS reimbursement, but to what? Medicare’s 1983 adoption of a prospective payment system for hospitals, which reimbursed hospitals according to the diagnosis of the patient at discharge, moved away from per-diem reimbursement. It changed the definition of a “service,” but not the underlying structure, and did nothing to affect much criticized hospital quality. At the other end of the spectrum, capitation or partial capitation, where a portion of the payment is fixed and a portion varies according to usage of care, dramatically changes the incentives that physicians face. However, this type of “fee for no service,” after 50 years of discussion, remains risky enough to be more of a principle than a practice—particularly in terms of how physicians are reimbursed, because of the challenge of determining and implementing quality and access guardrails against the incentive to stint on whatever efficiently produced care is supplied.

The question remains of how to best move the US health care system from one that pays primarily for volume, especially for physician care, to one that encourages greater focus on higher value, efficiently produced care. Whether that can happen is unclear and will require bold action by the new Biden administration.

Here is what that bold action might entail. In terms of payment for quality, the payer or “principal” needs to decide in advance what outcomes it wants to incentivize the agents (hospitals or doctors) to achieve, and find hard metrics that will trigger substantial enough rewards or penalties to meet those targets. A payer, for example, might define a target rate of risk-adjusted performance for joint replacement for an accountable care organization (ACO) population—and then find the dollar amount per payment bundle needed to get the ACO hospitals and physicians to do just that amount. Or Medicaid might pay health systems much more to take care of social problems like housing or food adequacy for poor and disadvantaged populations.

Penalties often are harder to implement than rewards (as CMS found with the readmission reduction program) because they can drive some suppliers out of network or out of business, especially if risk is not fully controlled for. However, in addition to the obvious budgetary challenges, paying hefty rewards for desired behavior to hit defined targets raises another issue. It may well be that the amount needed to get physicians, for example, to change behavior enough to improve on some metric exceeds the gain to the payer of hitting the metric. This “principal’s curse” arises when the value of the health improvement at any level of payment to agents is less than the additional cost of the improvement. Though it is obviously hard to attach money values to quality, that unrewarding task is unavoidable once you start incentivizing with money. The Biden administration, therefore, needs to be prepared for the possibility that in many cases it may not be desirable to pay for value if value costs too much. Humble, but critical, analysis of a type that avoids excesses of boldness may be the best counsel.

About the Authors