Every US policy maker knows that Medicaid presents vexing budget challenges, but a new analysis suggests that its costs are becoming truly unsustainable. McKinsey estimates that even after economic growth returns to a steady pace, this government health insurance program, which primarily serves the poor, will consume more than 75 percent of all new state revenues in 10 states, including Georgia, Indiana, and Oregon, by 2009. Medicaid will cost 11 other states, including Florida, New York, and Pennsylvania, from half to 75 percent of their incremental revenues (Exhibit 1). In an additional 22 states, this one program will consume 25 to 50 cents of each new tax dollar.¹

These results are obviously troubling for state leaders who have initiatives in education, public safety, transportation, and other critical areas high on their agendas, as well as for federal officials already grappling with record budget deficits. Yet they also present an opportunity. We believe that Medicaid's challenges have reached a critical point: the untenable nature of the status quo can be communicated compellingly to stakeholders, the press, and the public. As the risks of inaction—or of muddling through at the margins—come to be greater than those of action, creative leaders can communicate the facts in ways that open up paths toward a new consensus on the program's future.

Few governors or senators come into office with a deep desire to fix Medicaid, yet like all political leaders they must take the world as they find it. To stabilize spending, make room for other investments, keep taxes at levels consistent with a good business climate, and preserve a commitment to the neediest US citizens, every governor is destined to be a "Medicaid reform governor" in the years ahead. Many federal officials also will need to engage with these efforts.

The way forward starts with two steps. First, leaders must effectively marshal the facts to make the case for change; then they must develop and assess reform options that could make the program viable while honoring its traditional values. In an effort to contribute to the thinking of public officials on the future of Medicaid, in this article we will analyze its unsustainable spending trajectory, discuss opportunities for savings, and sketch some of the options for reform.

The program that ate the budget

Medicaid is now the health insurer for 15 percent of the US population. Thanks to recent coverage expansions, it has become the country's largest health insurance program, surpassing Medicare, a program mostly for retirees, in 2002 (Exhibit 2).

The taxpayers spent $305 billion on Medicaid in 2004, up from $74 billion in 1990. Enrollment rose to 41 million, from 23 million, over the same period. One striking measure of Medicaid's reach and of outsized US per capita health care spending is the fact that the United States spends more on Medicaid for 40 million poor people than the government of Britain spends on health care for its entire population of 60 million. And there's no end in sight: spending is expected to grow by roughly 7 to 9 percent annually over the next five years (Exhibit 3).²

A recent report from the nation's governors and state budget officers noted that in 2004, spending on Medicaid would, for the first time, exceed state expenditures for elementary and secondary education. Medicaid represents 21 percent of state budgets, on average, up from just 13 percent a decade ago. If current trends don't change, the program's cost will rise to a stunning 26 percent of state spending by 2009—and to as much as 46 percent of it in some states. The "crowding out" effect on other priorities will be severe.

Capturing savings opportunities

If these trends suggest that something has to give, the natural questions are what and how much? The short answer is that opportunities for savings in health care are large but capturing them can get complicated. This truth holds not only for Medicaid but also for other government health care programs and for private health care.

Consider the opportunities from a systemwide perspective. The US health care network today is radically inefficient. To begin with, the country spends 15 percent of its GDP on health care while many other advanced nations spend 9 to 10 percent. Those nations insure all of their citizens, while the United States has more than 40 million people without coverage. Despite a far greater investment, moreover, the United States doesn't report better public-health outcomes than other advanced nations do.

Within the US health care system, huge variations in practice patterns and medical spending bear no relation to quality. Medicare, for example, spends two and a half times more per senior citizen in Miami than in Minneapolis; health plans report that prices for identical procedures often vary by as much as 500 percent in different hospitals, depending on the market. These variations are not explainable by differences in the health status of patients or by regional differences in input costs. Instead, they suggest that a huge amount of unnecessary or ineffective care is being offered and that, in many local markets, some providers enjoy unusual market power, and thus pricing leverage.

Jack Wennberg of Dartmouth and other experts in this area estimate that up to 30 percent of today's care could be eliminated if the system evolved toward best-practice, evidence-based treatment.³ Dr. David Brailer, the national coordinator for health information technology at the US Department of Health and Human Services, estimates that widespread modernization of the archaic IT infrastructure of health care could eventually reduce its cost by 10 percent through administrative and clinical savings. In a $1.7 trillion health care economy, such opportunities add up to real money—and a large portion of it flows through Medicaid.

Why are savings so hard to capture? One big difficulty is political: every dollar of health care waste is somebody's dollar of income. Beyond this, the health care system and Medicaid itself pose structural challenges:

• Insufficient data on outcomes. Policy makers have little choice but to focus on spending because they lack good data on outcomes. States and other payers have no way of knowing if they get more value for the extra cash they devote to health care. Certainly, some new expenditures on technology—to treat depression, cataracts, and heart attacks, for example—have generated benefits that far outweigh their cost. But reimbursement methods don't hold providers accountable for improving quality and performance.
• Rigid benefit rules that bar cost-effective tailoring. Federal benefit and enrollment rules make it hard for states to offer more cost-effective solutions to different populations. In some states, for example, the core set of federally mandated Medicaid benefits accounts for only 30 percent of the program's overall cost (Exhibit 4). States may add services, such as long-term care and coverage for prescription drugs, and federal regulations generally require that other covered groups receive the same benefits, even when they may not be necessary. (The benefits required for optimal service to the blind and disabled, for instance, are more elaborate than those needed by the broader population of low-income uninsured people.) Also, when all beneficiaries must receive comparable benefits, coverage for low-income workers who qualify for Medicaid can be substantial enough to induce employers to stop providing private health insurance for their employees.
• Lack of consumer involvement. Because states are sensitive to the burdens of their poorer citizens, they have exempted Medicaid recipients from playing any role in constraining costs. Co-payments for visits to physicians, for instance, are nominal—$1 and $3 for an appointment with a primary-care physician and a specialist, respectively—and haven't been adjusted for inflation in 20 years.
• Shrinking access to providers. In most states Medicaid already pays less than other payers do—sometimes less than the actual cost of providing care. At the same time, the program has expanded to cover a greater percentage of the population, so doctors, hospitals, and other providers have found it more difficult to recoup some of these lost revenues by charging more to patients covered by otherprograms. A growing number of providers thereforerefuse to take Medicaid patients.
• Federal rules that reward big spenders. Perverse incentives or unintended consequences often flow from the rules of the federal government. Medicaid regulations call for it to match state spending on the program, thus giving more support to poorer states. Yet because state coverage decisions ulti-mately drive the size of this match, federal formulas direct more absolute aid to higher-spending states, leaving large intrastate gaps. In 2001, for instance, New York spent $11,060 per recipient and received $5,520 (or 50 percent of the total); Mississippi spent $4,400 per recipient and received $3,380 (77 percent of the total).

Reforming an entrenched program like Medicaid is politically risky, so early efforts have been modest. With a few exceptions, the first instinct has been to muddle through in the traditional way, by slowing the pace of enrollment and reducing payments to providers. However, policy makers will quickly learn that the usual steps can't return Medicaid to a sustainable path.

Toward sustainable reform

Given the complexity of any effort to address the real drivers of medical costs and the structural problems of Medicaid, policy makers will likely need a combination of potential solutions and reform themes. These ideas fit within two broad categories: reforms that fiscally strapped states can begin to implement now, largely on their own, and longer-term changes requiring federal leadership or more extensive federal-state collaboration.

Getting started now

Five areas offer serious opportunities for near-term improvement. These proposed reforms may not fully cure Medicaid, but they can deliver major savings and improve the quality of care while the dialogue about longer-term reform begins.

Maximizing value from pharmaceuticals. No single reform prescription can control the growth of spending on pharmaceuticals, but real opportunities to get more quality for less money do exist. Clinical guidelines and carefully designed lists of preferentially covered drugs (formularies), for example, can encourage doctors and patients to select the lowest-cost drug that is safe and effective. Other techniques, such as reviewing specific clinical decisions (preauthorization), are available, too. The increased use of generic drugs can often reduce costs within therapeutic categories by 30 to 40 percent. Policy makers should emphasize three metrics: total drug costs (rather than maximum drug rebates), total health care costs, and clinical outcomes.

Measuring outcomes is important because drugs are both overused and underused. To target overuse, state Medicaid agencies should conduct prospective and retrospective utilization-review programs. Behavioral drugs, such as antidepressants and antipsychotics, would be a good place to start, since their cost is exploding and cases in which they were prescribed inappropriately have been widely reported. Underuse is an issue when taking medicine today could prevent more costly problems tomorrow. Raising the proportion of Medicaid enrollees who take medication to treat high blood pressure, for example, would probably deliver excellent health returns for relatively modest investments in education and outreach.

Innovating in care for the disabled. Many Medicaid beneficiaries suffer from a number of medical conditions, and an integrated approach could improve outcomes and manage costs. What's more, our analysis in Tennessee suggests that this population's turnover rate is quite low. In that analysis, more than 40 percent of all Medicaid beneficiaries, unlike people with commercial insurance, remained in the high-cost group from year to year (Exhibit 5). Cost and quality improvements from disease-management programs are therefore likely to be significant and enduring.

Recognizing that truth, Medicaid agencies and managed-care companies are beginning to replace stand-alone efforts for each medical condition with more specialized and integrated case- and disease-management programs targeted at this population. Promising innovations include collecting better information about enrollees, using improved predictive-modeling techniques, training Medicaid staff members in the social and cultural issues of serving a diverse group of consumers, and developing a more integrated approach to coordinating care for this challenging population.

Moreover, the cost of hospitalizing disabled beneficiaries who are also institutionalized in long-term-care facilities could be reduced. Medicare currently pays for these patients' hospital costs, while Medicaid covers long-term care. This division of expenses creates a perverse incentive encouraging unnecessarily long hospital stays. Better care-management practices—such as the use of hospitalists (specialists who primarily serve inpatients in hospitals)—should reduce their length and raise the level of expertise, thereby generating substantial savings.⁴

Developing standards for providers. States are large purchasers of health care and also regulate commercial insurance to a significant degree, so they have a unique position to craft standards for provider practices and patient care. These standards should include meaningful measurements of outcomes, pay for performance, and evidence-based medical guidelines that reduce variations in care and the associated costs. The creation of these standards offers opportunities for public-private collaboration. In California, for instance, six commercial health plans—covering eight million people treated by 30,000 physicians—started working together in 2002 to develop standard metrics for outcomes. The metrics cover immunization, screening for cancer, the monitoring of diabetes, asthma medication rates, patient satisfaction, the use of technology, and more. If California's Medicaid program joined such commercial efforts, it could accelerate the adoption and increase the impact of these standards.

In particular, states should take the lead in establishing evidence-based practice guidelines—an area in which a great deal of underlying work has been completed. Professional societies, managed-care organizations, and academic medical centers have already developed a wide set of guidelines for most conditions and symptoms. What has been missing is a commitment by public and commercial health care purchasers to endorse a specific standard of care, hold providers to it, and show the public how well the system is performing.

Creating tiered benefits. Medicaid serves several populations with special needs, including pregnant women, children, and disabled adults, as well as additional "optional" populations that states can choose to add. A tiered benefit structure would help states match coverage to the needs of each group while also saving money; an example might be higher co-payments and less extensive benefits for the optional population and lower co-payments and richer benefits for the disabled one. Likewise, changing the levels of co-payments for drugs (along the lines of the three- or four-tiered co-payment programs of commercial insurance) could provide greater coverage at the same cost.

While critics may argue that all Medicaid patients deserve the same coverage, policy makers could explain that the suggested changes are part of a necessary effort to preserve benefits for the broader low-income population over the long term. A move toward tiering requires federal support; the federal Centers for Medicaid and Medicare Services (CMS), which help administer Medicaid and Medicare, would have to give states more flexibility in developing their programs. Governors, in a spirit similar to that of their participation in welfare reform during the late 1980s and early 1990s, may present CMS with a united front promoting new solutions.

Creating fixed budgets. Medicaid budgets are now open-ended, so cost savings might come from some form of cap. Block grants are one way to provide it. Another method, which could work under the current federal matching system, would be for states to create their own internally set budget targets. A state might, for example, enact a budget trigger that forced action when Medicaid (or total health care expenditures) reached a certain percentage of overall spending. Exceeding this budget threshold could prompt a study by an advisory commission that would recommend adjustments to benefits, eligibility, and coverage—as happens routinely in private-sector companies when they exceed cost targets for their employees. Such devices have been discussed at the federal level to control long-term spending on entitlements, and one version was adopted in the 2003 Medicare legislation.

While a fixed budget isn't a solution in itself, it forces policy makers to confront critical trade-offs. These are essentially a matter of political values: how should the desire to cover the broadest number of people be weighed against the level (and thus the cost) of benefits for each insured person? Such wide-versus-deep choices are always made implicitly; one virtue of fixed budgets is a more explicit consideration of the trade-offs.

Longer-term reforms

Other reforms involve a fundamental rethinking of the system, greater federal-state collaboration, and more active federal leadership. These approaches are better seen as part of the longer-term debate. Policy makers may want to address the following questions:

Should Medicaid alter its eligibility criteria? Like many government programs, Medicaid has extended coverage to different groups at various times, more for political reasons than from any grand or consistent design. As a result, more than 8 million people who have incomes upward of 150 percent of the federal poverty level are now covered, while roughly 18 million poor working people (especially unmarried adults and childless couples) who have incomes less than 150 percent of the federal poverty level are not.⁵ Some analysts say it is time to change the rules so that coverage is more closely tied to financial need.

Should Medicaid promote consumer-directed health decisions? Medicaid has traditionally been paternalistic. But in some contexts, giving patients choice and even requiring them to share costs to a modest extent may be appropriate ways to enhance the quality of health care and to cut spending on it. States could, for example, increase funding for programs that give consumers information about the opportunities for home-based alternatives to long-term care. In addition, consumer-directed health savings accounts, though now in their infancy, may be relevant for some Medicaid patients if they are designed carefully.

Should governments create Medicaid vouchers? States or the federal government might give Medicaid recipients vouchers they would use to purchase insurance on an appropriately regulated private insurance market. A variant of this approach has been used in the 40-year-old Federal Employees Health Benefits Program, which covers nine million federal workers and retirees. Consumers would purchase insurance—such as fee-for-service or managed-care plans, as well as health savings accounts—from a variety of state-approved companies, and since consumers would own the policies, they could be fully portable.

While such reforms are often associated with the Republicans, former senator Bill Bradley, a Democrat, offered a similar framework during his 2000 presidential campaign. In addition, the health economist Victor Fuchs is working on a proposal for universal health care vouchers to make the system more fair and efficient.

The road to reform

Our work helping leaders address such challenges suggests how vital it is to begin by making the case for change. The late US senator Daniel Patrick Moynihan famously said, "Everyone is entitled to their own opinions but not to their own facts"—counsel that may offer the best starting point to fix Medicaid. As stakeholders spar over the program, the facts about its future can provide a wake-up call for a dialogue about reform. At the federal level, the inevitable quest to cut the budget deficit will ensure that Medicaid reform has a central place. Each state, meanwhile, will have its own particular situations raising the stakes. A thorough analysis of the program's local economics should offer insights that can alter perceptions and lay the groundwork for a constructive engagement with citizens and the health care community alike. An analysis of Tennessee's program, for instance, revealed that the state spends more on prescription drugs through Medicaid than on the entire state system of higher education. Reform is never easy, but such facts can begin to influence the terms of debate.

Medicaid's future is interwoven with broader questions about the cost and quality of health care. They will dominate business and government debates for the next two decades as 76 million baby boomers retire. In a country that spends nearly one in six dollars of its income on health care, remarkably little is known about the value this investment actually delivers. In addition, if US spending on health care rises from 15 percent of GDP now to 20 percent by 2020, as currently forecast, those expenditures will challenge other public priorities, not to mention profits, wages, and US competitiveness.

The economist Herbert Stein had a wonderful maxim: "If a trend is unsustainable, it tends to stop." He was right, of course, but we can control spending on health care while boosting its quality; the question is how. In Medicaid's case, the path to sustainability begins with creative political leadership.