Bending the Curve: Options for Achieving Savings and Improving Value in U.S. Health Spending

Interactive: Click policy options on left to view.

Executive Summary below the Full Report:

U.S. health spending is projected to increase from 16 percent of GDP in 2006 to 20 percent in 2016—from $2 trillion to $4 trillion. Meanwhile, the number of uninsured Americans continues to rise. In this report prepared for The Commonwealth Fund Commission on a High Performance Health System, the authors examine 15 federal policy options that have the potential to lower health spending relative to projected trends. They include policies that would: produce and use better information for health care decision-making, promote health and enhance disease prevention, align financial incentives with quality and efficiency, and correct price signals in health care markets. Combining policies would capture the synergistic benefits of individual changes. If implemented along with universal health insurance, a combination of selected options could save $1.5 trillion in national health expenditures over 10 years, while also improving value in terms of access, quality, and health care outcomes.

Executive Summary

Health spending in the United States is projected to increase from 16 percent of gross domestic product (GDP) in 2006 to 20 percent in 2016—from $2 trillion to $4 trillion in 10 years. At the same time, the number of people who are uninsured is rising sharply, including a growing proportion of middle-income families. While rising costs are putting all sectors of the economy at risk, the nation lacks a concrete, realistic plan for adopting a different approach that could achieve savings and improve value.

To inform national discussions and spur progress toward such a plan, The Commonwealth Fund Commission on a High Performance Health System sponsored this report, which examines 15 federal policy options and their potential for lowering health spending over the next 10 years, relative to projected trends.

These options are not presented as the recommendations of the Commission, but they represent a range of approaches that have been proposed to address the various factors that contribute to high and rising costs and represent sources of inefficiency in the current health care delivery and financing systems. The report focuses on federal policies for three reasons: the federal government accounts for the largest portion of health spending; changes at the federal level would probably have the broadest immediate effects on national health spending; and federal policies, particularly those adopted by Medicare, frequently serve as a model for policies adopted at the state and local levels and by the private sector. Nonetheless, many of the policy options could be applied by states and private payers as well. Indeed, collaborative efforts across public and private sectors will be essential for achieving higher performance and greater value.

The report's findings illustrate that it would be possible to reduce national expenditures over the next decade while simultaneously improving access, quality, and population health. Achieving significant savings, however, will require a combination of policies that span strategic areas amenable to policy action at the federal level. These include policies that:

• Produce and use better information for health care decision-making;
• Promote health and enhance disease prevention efforts;
• Align financial incentives with health quality and efficiency; and
• Correct price signals in health care markets.

By applying these policies collectively, the nation would be able to capture the synergistic benefits of specific changes that, if implemented individually, would yield more modest reductions in projected spending trends. Further, policies aimed at achieving savings while also improving quality would be even more effective in improving overall health system performance if they were combined with a policy to extend affordable health insurance coverage to everyone in the United States. On a foundation of universal coverage, payment and other policies could apply to a larger share of the population. Well-designed insurance also has the potential to lower administrative costs while ensuring access—both improve value. Combining selected options with affordable health insurance for all could yield $1.5 trillion in national health expenditure savings over 10 years, and enable a more integrated, systemic approach to health care delivery and financing.

Modeling the future impact of complex policy changes is inherently challenging and risky. The technical challenges include the uncertainty of estimating dynamic effects over time. Further, the estimates assume effective design and implementation, and therefore do not reflect the difficulty of achieving agreement on what changes are necessary, designing the often complex policies necessary to achieve those changes, or making the organizational adjustments required to implement them successfully.

What is certain is that the stakes a re very high if we continue on our current path of escalating costs and eroding coverage.

Options and Results

This report contains analyses of a set of 15 federal policy options that could ease health care cost pressures while at the same time either enhancing or maintaining access, quality, efficiency, equity, and the health system's capacity to innovate and improve. These options include federal policies targeted to produce and use better information, promote health and prevent disease, align incentives with quality and efficiency, and correct price signals in the health care market. The Commonwealth Fund contracted with the Lewin Group to estimate the potential effects of each option, with a focus on total national health expenditures and the distribution of expenditures across payers—the federal government, state and local governments, private employers, and households. (The Lewin Group is one of the leading health care and human services consulting firms in the United States, with more than 35 years of experience serving organizations in the public, nonprofit, and private sectors.) The estimates include effects on incremental and cumulative spending over a 10-year period, from 2008 to 2017. A summary of the options modeled in this report, their objectives, and the estimated effects on spending are described below.

Producing and Using Better Information

These options are intended to address information barriers that contribute to the inefficiency of our health system and undermine care outcomes. The transparent availability of information and the incentives and ability to use it are critical prerequisites for effective, safe, coordinated care and the development of policies that encourage such care.

• Promoting Health Information Technology. Accelerate provider adoption of health information technology (HIT) with the capacity for decision support and to share patient health information across sites of care, financed by an assessment of1 percent on insurance premiums and Medicare outlays. After initial investment costs, estimated net health system savings could reach $88 billion over 10 years as HIT capacity is improved. Net savings would accrue by year 10 to all except private payers, which would realize cumulative savings in following years.
• Center for Medical Effectiveness and Health Care Decision-Making. Invest in the knowledge needed to improve decision-making and incorporate information about the relative clinical and cost-effectiveness of alternative treatment options into insurance benefit design. By generating the information and creating payment and cost-sharing incentives for providers and consumers to use it, this policy option could result in estimated health system savings of $368 billion over 10 years, shared by all payers.
• Patient Shared Decision-Making. Help patients decide between alternative treatment options by requiring providers to educate Medicare beneficiaries about alternatives through use of patient decision aids (such as videos and other materials). This option could save an estimated $9 billion over 10 years, primarily for the Medicare program. System savings would be greater if this policy were extended to Medicaid and private insurance.

Promoting Health and Disease Prevention

These options focus on the substantial costs to the health system of the care and complications of chronic diseases, such as diabetes or heart disease. They seek to lower the incidence of disease through public health initiatives and improved care.

• Public Health: Reducing Tobacco Use. Increase federal taxes on tobacco products by $2 per pack for cigarettes, with revenues going toward support of national and state tobacco control programs. If revenues were invested in effective programs, this option could yield $191 billion in health system savings over 10 years, shared by all payers. State savings would be largely offset by reduced state tobacco tax revenue, as consumption of tobacco products fell.
• Public Health: Reducing Obesity. Establish a new nominal tax on sugar-sweetened soft drinks of 1 cent per 12-ounce drink to finance national and state obesity prevention programs. If successful in reducing rates of increases in obesity and associated costs, the option could yield an estimated $283 billion in savings over 10 years, shared by all payers.
• Positive Incentives for Health. This option would use federal funds and incentives to encourage the federal government, state governments, and private employers to create positive incentives for individuals to engage in wellness programs and healthy behavior, and to cover preventive services. Such a focus on high value benefit designs could save an estimated $19 billion over 10 years, with a net investment by the federal government of $2 billion.

Aligning Incentives with Quality and Efficiency

These policy options are intended to address the misalignment of incentives in our fee-for-service payment system and the private insurance market. The options modeled include:

• Hospital Pay-for-Performance. Establish a Medicare pay-for-performance program for all hospitals similar to the current Centers for Medicare and Medicaid Services (CMS)/Premier Hospital Quality Incentive Demonstration. This option could result in health system savings of $34 billion over 10 years, with the major share accruing to the federal government through reduced Medicare payments, primarily from decreased hospital readmissions. If all payers adopted similar policies, estimated savings would be greater.
• Episode-of-Care Payment. Transform the current Medicare fee-for-service payment system to fixed prospective payments per episode of care (based on the current distribution of cumulative fee-for-service costs per episode). This policy would change Medicare payment methods to reward and encourage more efficient, coordinated care. When applied to hospital and ambulatory care, this could generate estimated net health system savings of $229 billion over 10 years. Other payers could avoid cost-shifting by emulating this payment approach.
• Strengthening Primary Care and Care Coordination. Change reimbursement to primary care physician practices to support enhanced primary care services, such as care coordination, care management, and easy access to appropriate care. Under this option, Medicare fee-for-service beneficiaries would be enrolled in "medical homes" that have this enhanced capacity. Mandatory enrollment could result in net health system savings of $194 billion over 10 years, with savings accruing to all payers. Estimated savings would be larger if this approach were adopted by all payers.
• Limit Federal Tax Exemptions for Premium Contributions. To provide incentives to enroll in high-value health insurance plans, cap the tax-deductibility of employer-sponsored insurance premiums. The option could reduce national spending by an estimated $131 billion over 10 years, with savings in federal tax expenditures exceeding that amount. However, to avoid putting sicker, older, and low- or modest-income families at increased health and financial risk, and to avoid potentially undermining current employer-sponsored pooled-risk group coverage, this change would have to be combined with universal coverage and changes in insurance market rules.

Correcting Price Signals in the Health Care Market

These options seek to address the tendencies of the current pricing mechanisms to send the wrong signals to participants in the market. These include signals for higher rather than lower costs and pricing mechanisms that support inefficient care and wide variation in costs without corresponding differences in quality and outcomes across geographic areas.

• Reset Benchmark Rates for Medicare Advantage Plans. Modify the current Medicare Advantage payment methodology by setting the benchmark rate for plans in each county at a level equal to the county's projected per capita spending under traditional Medicare. The current mechanism for setting the benchmark rates, which results in payments to plans that are higher than what costs would have been in traditional Medicare, sends a price signal through the market that encourages higher rather than lower costs among those plans. By recalibrating benchmark rates, this option could result in an estimated health system savings of $50 billion over 10 years and reduce federal spending by $124 billion over the decade. However, spending by Medicare beneficiaries would increase by $74 billion, in the form of additional private premiums for those individuals who replace the additional benefits available under the current Medicare Advantage payment rates, as well as extra payments required for those who enroll in plans with bids that exceed the benchmark rate.
• Competitive Bidding. Establish competitive bidding among Medicare plans and traditional Medicare. This option would replace the current administered pricing mechanism in Medicare Advantage with a system that would determine prices through increased competition on the basis of quality and efficiency. The option could result in estimated health system savings of $104 billion over 10 years and substantial reductions in federal spending over the same period. Spending by Medicare beneficiaries would increase by $178 billion, as they may choose to stay in more expensive plans (or remain in traditional Medicare where it is more expensive than the available alternatives), or may choose to pay to restore extra benefits they currently receive under Medicare Advantage. This option would have to be designed carefully to avoid drawing healthier beneficiaries away from Medicare and putting elderly and disabled beneficiaries, as well as those with modest incomes, at risk.
• Negotiated Prescription Drug Prices. Give the U.S. Secretary of Health and Human Services the authority to negotiate or set price limits for Medicare prescription drug plans for their enrollees. This option could result in a net savings of $43 billion over 10 years, with a focus on dual eligibles and prescriptions within monopolized seller markets. Without provisions to prevent cost-shifting, all payers except the federal government could experience a net increase in spending. The potential impact of this option on innovation in the development of new effective drugs would have to be assessed.
• All-Payer Provider Payment Methods and Rates. Require all payers to adopt Medicare payment rates and methods for hospitals and physicians. This option would provide higher payments for Medicaid patients and reduce the pressure on the prices paid by private insurers to offset Medicaid and other shortfalls. It would also address the fragmented system under which providers must deal with numerous payment mechanisms and reporting rules. The option could result in net system savings of $122 billion over 10 years, with the savings accruing to the private insurance industry. This estimate presumes the resolution of the payment cuts projected under the current sustainable growth rate mechanism for physician payments, and so does not reflect the projected cost of those changes. Moreover, to avoid undermining safety net providers, it would be necessary to redirect some of the savings to finance insurance expansion or uncompensated care pools.
• Limit Payment Rate Updates in High-Cost Areas. Reduce Medicare spending growth by basing annual hospital and physician payment updates on cost per beneficiary in relation to a national benchmark. This option would focus more cost control pressure on high-cost regions and avoid across-the-board adjustments that otherwise would apply equally to low- and high-cost geographic areas. Limiting payment growth in high-cost regions could save $158 billion in health system spending over 10 years, with savings accruing to the federal government. Unless they followed Medicare's lead, other payers in the affected regions could see increased costs as a result of cost-shifting.

The estimated net effects of each of these options on national health spending, and on spending by payer group, are shown in exhibits ES-1 and ES-2, respectively.

Exhibit ES-1. Policy Options and Net Cumulative Impact onNational Health Expenditures over One, Five, and 10 Years
	One-Year Impact on NHE (billions)	Cumulative Five-Year Impact on NHE (billions)	Cumulative 10-Year Impact on NHE (billions)
Producing and Using Better Information
1. Promoting Health Information Technology	$8	$14	-$88
2. Center for Medical Effectiveness and Health Care Decision-Making	-$18	-$125	-$368
3. Patient Shared Decision-Making	-$1	-$4	-$9
Promoting Health and Disease Prevention
4. Public Health: Reducing Tobacco Use	-$5	-$64	-$191
5. Public Health: Reducing Obesity	-$3	-$61	-$283
6. Positive Incentives for Health	$0	-$5	-$19
Aligning Incentives with Quality and Efficiency
7. Hospital Pay-for-Performance	-$2	-$14	-$34
8. Episode-of-Care Payment	-$17	-$96	-$229
9. Strengthening Primary Care and Care Coordination	-$5	-$60	-$194
10. Limit Federal Tax Exemptions for Premium Contributions	-$10	-$55	-$131
Correcting Price Signals in the Health Market
11. Reset Benchmark Rates for Medicare Advantage Plans	-$3	-$20	-$50
12. Competitive Bidding	-$7	-$42	-$104
13. Negotiated Prescription Drug Prices	-$3	-$16	-$43
14. All-Payer Provider Payment Methods and Rates	$2	-$23	-$122
15. Limit Payment Updates in High-Cost Areas	-$4	-$43	-$158
Note: A negative number indicates spending decreases compared with projected expenditures (i.e., savings); a positive indicates spending increases.

Exhibit ES-2. Policy Options and Distribution of 10-Year Impact onSpending Across Payer Groups (in billions)
	Total NHE†	Federal Gov't	State/ Local Gov't	Private Payer	Households
Producing and Using Better Information
1. Promoting Health Information Technology	-$88	-$41	-$19	$0	-$27
2. Center for Medical Effectiveness and Health Care Decision-Making	-$368	-$144	-$49	-$98	-$107
3. Patient Shared Decision-Making	-$9	-$8	$0	$0	-$1
Promoting Health and Disease Prevention
4. Public Health: Reducing Tobacco Use	-$191	-$68	-$35	-$39	-$49
5. Public Health: Reducing Obesity	-$283	-$101	-$52	-$57	-$73
6. Positive Incentives for Health	-$19	$2	-$12	-$4	-$5
Aligning Incentives with Quality and Efficiency
7. Hospital Pay-for-Performance	-$34	-$27	-$1	-$2	-$4
8. Episode-of-Care Payment	-$229	-$377	$18	$90	$40
9. Strengthening Primary Care and Care Coordination	-$194	-$157	-$4	-$9	-$23
10. Limit Federal Tax Exemptions for Premium Contributions	-$131	-$186	-$19	-$55	$130
Correcting Price Signals in the Health Care Market
11. Reset Benchmark Rates for Medicare Advantage Plans	-$50	-$124	$0	$0	$74
12. Competitive Bidding	-$104	-$283	$0	$0	$178
13. Negotiated Prescription Drug Prices	-$43	-$72	$4	$17	$8
14. All-Payer Provider Payment Methods and Rates	-$122	$0	$0	-$105	-$18
15. Limit Payment Updates in High-Cost Areas	-$158	-$260	$13	$62	$27
Note: A negative number indicates spending decreases compared with projected expenditures (i.e., savings); a positive indicates spending increases. † In some cases, because of rounding, the sum of the payer group impact does not add up to the national health expenditures total.

Combining Individual Options with Affordable Coverage for All

In addition to the individual options described above, we modeled the effects of several of those options under a scenario intended to provide affordable health insurance coverage for all. The universal coverage scenario is based on a policy that would expand affordable coverage through a blend of private and public group health insurance. Similar to the savings options above, this scenario is presented as one of a range of potential approaches aimed at accomplishing this goal, rather than as the particular policy favored and recommended by the Commission. We refer to this scenario as the Insurance Connector approach.

The Insurance Connector approach builds on and connects current public and private group insurance through the creation of a new national entity that would offer a structured choice of private health plans as well as a Medicare option to individuals and small employers. Enrollment in some plan would be required. The availability of both publicly sponsored and private plans would help induce competition not only among private insurers, but also between private insurers and the public plan. This would put pressure on all plans to operate more effectively and efficiently. The expansion would achieve near-universal coverage.

With insurance changes alone, total health system costs would increase by an estimated $15 billion in the first year and $218 billion over 10 years, as a result of improved access for those who are currently uninsured or underinsured. (This scenario offers the option of selecting Medicare. The modeling estimates that lower administrative costs and other features would partially offset costs of coverage expansion.) As modeled, this scenario would increase spending by private employers and the federal government, because of financing provisions to make coverage affordable. State and local governments and households would experience net reductions in spending.

To illustrate the potential of policies focused on better information, public health, improved incentives, and price signals in the context of universal coverage, we modeled the Insurance Connector approach together with the following options that were described above: Promoting Health Information Technology; Center for Medical Effectiveness and Health Care Decision-Making; Public Health: Reducing Tobacco Use; Public Health: Reducing Obesity; Episode-of-Care Payment; Strengthening Primary Care and Care Coordination; Competitive Bidding; and Negotiated Prescription Drug Prices.

In the context of universal coverage with a national insurance connector as described above, several of the Medicare-focused policies that are combined in this option would apply to a larger number of people and therefore would have a larger estimated effect. Although not included in the modeling, in addition, the synergistic effects of implementing a combination of policies aimed at improving health system performance with universal coverage could be expected to lead to even greater savings. For example, the potential savings from electronic medical records (as well as the improvements in the quality and effectiveness of care) would be augmented if physicians also had expanded information on clinical effectiveness.

This combined approach could lower national health expenditures by 1 percent initially and 6 percent after a decade, compared with baseline projections. These annual savings add up to cumulative 10-year savings over the current system baseline of more than $1.5 trillion, as shown in exhibits ES-3 and ES-4. As illustrated, the cumulative effect of the combination of options grows rapidly over time: the estimated reduction in national health expenditures in the first year is $31 billion, while the reduction over 10 years is more than 50 times greater; similarly, the net costs to federal government diminish rapidly over time as savings offset federal costs of insurance expansion. Further, by the end of a decade, the net federal costs could be negligible if bundled with options that focus on improving both the effectiveness and efficiency of care.

Exhibit ES-3. Distribution of the Effect of Combination of Selected Individual Options with Insurance Connector Approach on Spending over One Year, Five Years, and 10 Years Across Payer Groups (in billions)
	Total NHE†	Federal Gov't	State/ Local Gov't	Private Payer	Households
Combining selected individual options with Insurance Connector approach††
after 1 year	-$31	$31	-$14	$24	-$71
after 5 years	-$407	$111	-$119	$87	-$486
after 10 years	-$1,554	$158	-$380	$72	-$1,404
Note: A negative number indicates spending decreases compared with projected expenditures (i.e., savings); a positive indicates spending increases. † In some cases, because of rounding, the sum of the payer group effect does not add up to the national health expenditures total. †† Selected options include: Promoting Health Information Technology; Center for Medical Effectiveness and Health Care Decision-Making; Public Health: Reducing Tobacco Use; Public Health: Reducing Obesity; Episode-of-Care Payment; Strengthening Primary Care and Care Coordination; Competitive Bidding; and Negotiated Prescription Drug Prices.

As shown in Exhibit ES-5, these estimated savings succeed in "bending the curve" to about halfway between the currently projected trend and the amount that would maintain the current proportion of GDP devoted to health spending. A policy of guaranteeing health insurance for all combined with selected savings options yields savings in national health expenditures of $1.5 trillion over 10 years. This represents an extremely large amount of resources that could be available to address other societal needs or wants, either within the health system or elsewhere. Moreover, an environment in which affordable health care is available, markets operate with better information, and payment reform offers potential for even greater savings, can produce dynamic, synergistic gains over the longer term. The first-order estimates are thus likely conservative compared with the potential gain over time.

Toward a Higher-Value Health System: Cross-Cutting Themes and Conclusions

When one considers the results presented in this report, some key themes emerge:

• Improvement is possible, and it is urgent to start now. The consequences of continuing the status quo, with respect to both human and economic costs, are very significant. The numbers of the uninsured are up sharply and moving up the economic ladder as middle-income families lose coverage. Costs are squeezing households, businesses, and the public sector. The option estimates illustrate that cost savings are achievable in the context of a high performance health system. They also demonstrate that early enactment of even modest changes has the potential for substantial cumulative benefits over several years. On a base of more than $2 trillion, even small percentage changes add up quickly.
• Better information is a key to improved performance. It is difficult to improve the health system without information on current performance at the national, local, and individual provider levels. Transparency of information on quality and price is essential to the effectiveness of a number of policies that aim to achieve higher performance. A valid, publicly available database on provider performance, appropriately adjusted for patient conditions, is critical for focusing providers on improving both quality and efficiency, enabling payers to construct rewards and other mechanisms that encourage such behavior, and providing patients with the information they need to make appropriate choices. Data on the patterns and causes of variations in spending across geographic areas is essential for developing policies to narrow such variations and providing consistently effective and appropriate care, regardless of location.
• Addressing total health system costs, not shifting costs among sources of financing, should be the focus of policy action. Many of the policies proposed in the past have simply shifted costs from one payer source to another—between government and employers, or from payers to beneficiaries and patients. Narrow policies that cut governmental budget outlays by simply displacing those costs onto Medicare beneficiaries, or by paying substandard rates to providers under Medicaid, are stopgap measures that do not fundamentally address underlying health care cost trends.
• There are no magic bullets that by themselves fully address rising costs and key sources of inefficiency. Just as the steady increase in costs relative to incomes—which is projected to worsen over time—represents the cumulative effect of multiple and interacting factors, tackling cost levels and trends will require a coherent set of policies aimed at the misaligned incentives and structural flaws that plague our health system and produce the cost pressures we face. The design and effective implementation of policies matter. The solutions are not simple, and will require risk-taking and a willingness to invest, learn, and allow time for health systems and system capacity to improve through innovation.
• A multifaceted approach that is combined with health insurance coverage for all can be designed to achieve substantial reductions in future spending growth. When combined with universal coverage, a bundled approach focused on system performance should be able to reduce the growth of spending significantly over the next decade while maintaining and enhancing the value of our health care dollar. But we need to start now, with a strategic, coherent set of goals, policies, and incentives designed to address the underlying factors that add to costs without adding commensurate value.
• Value means more than savings in national health expenditures. Higher value includes improved performance on quality, equity, access, and healthy lives, in addition to savings. A policy proposal that generates a modest savings but achieves substantial improvement in access or health outcomes may be as valuable, or more so, than one that generates larger savings but makes minimal progress toward other health system goals. Options that extend health insurance to all, promote the public health, improve information and lead to more informed patient decisions, enhance quality and care coordination, and eliminate waste, duplication, and unnecessary care all contribute to value and performance.
• Reaching consensus will require a focus on the potential gain for the nation. The Commission has sought to identify options that are win-win—that is, that both achieve savings and contribute to improving key dimensions of health system performance. Yet, approaches that substantially reduce projected expenditures over time will by definition decrease revenues for some segments of the health care sector.
• Achieving high performance will require that every stakeholder take part in finding solutions. Across the individual options, the estimated distribution of savings or net new cost varies among major payers—the federal government, state and local governments, employers, and households. Achieving national health system savings may require a shift in payment sources and an increase in federal outlays. Doing so will also require that providers be willing to address payment inequities where providers that care for the uninsured and the poor receive lower compensation that those with privately insured patients. Narrow self-interest is a major barrier to changes that have the potential to benefit all.
  Constructive approaches will also require political compromises, and a willingness to forsake ideological purity. As a nation, we will need to move beyond the point where everyone's second choice is the status quo.
• Leadership is critical. Building consensus will require leadership and public/private collaboration, and a coherent set of goals, policies, incentives, and tools. Options will work better if public and private policies align toward a common aim of achieving a high performance health system. Consensus will also require a whole-system view: aiming for improved cost trends while improving population health and achieving continuous improvement over time.

The range of options considered in this report illustrates strategic approaches that could, in combination, ease cost pressures and create a path toward a higher performing, high-value health system. The goal of the analysis is to spark discussion and development of constructive national policies that could reduce costs and enable a more efficient, effective, and equitable health system.

With cost pressures mounting and coverage eroding, the stakes are high. As a nation, we will all gain if we focus on improving the value we obtain for the $2 trillion we are now spending on health care—a sum that will continue to consume a greater and greater share of our nation's economic resources, without yielding proportional gains to society, if we fail to act.

View more presentations from brianahier.

Citation

C. Schoen, S. Guterman, A. Shih, J. Lau, S. Kasimow, A. Gauthier, and K. Davis, Bending the Curve: Options for Achieving Savings and Improving Value in U.S. Health Spending, The Commonwealth Fund, December 2007

via commonwealthfund.org