Among the three challenges of health care—providing access, ensuring quality, and controlling costs—the latter seems the most daunting. Perhaps this is because efforts to rein in costs date back at least 70 years.
In the 1920s, concerns about the cost and distribution of medical care resulted in the formation of the Committee on the Costs of Medical Care. In 1932, this committee gave the first estimate of national health care expenditures: $3.66 billion, or 4 percent of the gross national income.
In 1971, President Nixon announced a "crisis" in medical costs and said that, unless costs were controlled, the system would collapse. That year, national health expenditures were $81.0 billion, representing 7.2 percent of the gross domestic product (GDP). The Nixon administration imposed price controls and supported legislation aimed at encouraging health maintenance organizations, health planning, and professional standards review organizations to monitor quality and utilization and reduce costs.
During the 1990s, the share of GDP made up by personal health care was steady or even declining slightly at just under 12 percent. Analysts attribute that period's slower growth to three factors: health plans' successful bargaining with providers over prices, managed care plans' use of strategies to control the volume of services, and competition among plans that restrained premium growth.
In the current decade, with health plan premiums jumping and health spending back at more than 13 percent of GDP, pressures to control costs are once again front and center. The projected national health expenditures for 2005 are $1,920.8 billion, or 15.7 percent of GDP. And yet medical costs are not evenly distributed among patients, as the work of Wennberg, Fisher, and their colleagues at Dartmouth reveals. Regional variations in costs, adjusted to account for age and sex, show twofold differences, while quality measurements are as good or even better in some lower-cost areas. Equally startling variations occur among academic health centers in different parts of the country. Regional variations have been called "the Achilles heel of American medicine."
Some argue that disease prevention efforts can restrain costs. Such efforts are certainly important, but promoting longer life may only defer expensive final illnesses, with additional consumption of health care services in the intervening years. Some have suggested that increased competition among health plans and providers could be an effective strategy for cost containment. But this theory ignores the fact that competition and marketing often increase the demand for services. While competitive forces may restrain prices, volume drives costs. Experience in the Medicare program demonstrates that costs go up as volume increases, even though prices are controlled.
In the New England Journal of Medicine, Paul B. Ginsburg, Ph.D., outlines "four basic options for slowing the trends in health care spending: one can increase the efficiency of health care delivery; increase the financial incentives for patients to limit their use of medical services; increase the administrative controls on the use of these services; or limit the resources available to the health care system." The American public would likely have difficulty accepting aspects of all of these options, except the first—increasing efficiency, which I define as the reduction of unnecessary services and care of marginal value. For this reason, the nation should focus efforts to control costs on increasing the efficiency of health care delivery.
Adherence to clinical practice guidelines can improve quality and at the same time raise the value of health spending. I expect that pay-for-performance programs, as recommended by the Medicare Payment Advisory Commission and in place for a growing number of private plans, will initially increase costs as efforts are made to identify underuse and encourage delivery of preventive and other appropriate services. At the outset, efforts to increase efficiency will likely take the form of measuring and reporting use of resources. However, as more research into cost-effectiveness programs guides the design of medical benefits, regional variations in utilization should decrease, and efforts to reduce cost and simultaneously improve quality should bear fruit.
Professional liability reform, especially toward alternate dispute resolution, such as a no-fault system, offers some opportunity to reduce defensive medicine costs and to direct legitimate payments toward injured patients rather than attorneys. But liability reform alone is not the answer to escalating costs.
More and better use of information technology to provide decision support and facilitate coordination of care ultimately will increase efficiency, although the providers of care must somehow spread out the initial investment.
Another way to increase efficiency is to reward primary care providers for coordinating care among the various specialists who treat patients with multiple chronic illnesses. Care coordination can help to avoid duplication of effort and deter use of specialty services with marginal value, such as unnecessary imaging. It will have to be implemented in such a way to avoid the "gatekeeper" stigma associated with managed care. True care coordination can prevent the fragmentation of care that lowers quality and wastes resources.
Finally, it is important to rationalize the capacity in the system, both in terms of facilities and human resources. Health planning at the community level to improve efficiency is a logical step. Antitrust laws should not interfere with this logic.
What about steps to involve the patient as a more prudent consumer of services? Certainly, informed patients are critical to enhanced value. The jury is still out on the effectiveness of health savings accounts and high deductible plans as cost-containment strategies. Such arrangements have the potential to worsen adverse selection, leaving sicker patients to draw resources from the traditional insurance system. Inducements to patients to seek care from more efficient providers are being tried, but have yet to show success. In any case, these will remain controversial until risk-adjustment methods are refined.
Notably, all of these strategies aim to increase efficiency and value in the system, rather than impose global budgets and price controls. In my view, cost increases are to some extent inevitable until the baby boom generation moves through the health system. Boomers have always been "the pig in the python" in our society—the bulge in an otherwise skinny age distribution that affects the school system, job market, and, as we are now seeing, the health care system.
Because of this, and because of the growth of technology that makes new medical services available, I believe health care costs will continue to increase. This will likely be acceptable to our society only if value increases as well. I believe that most Americans will prefer this to explicit rationing of care, at least in the near term.
Alan R. Nelson, M.D., is special advisor to the CEO of the American College of Physicians and is a former president of the American Medical Association. His views are entirely his own and do not necessarily represent the positions of any organization.
The views presented in this commentary are those of the author and should not be attributed to The Commonwealth Fund or its directors, officers, or staff.
 Health Spending Growth Slows in 2003. Cynthia Smith, Cathy Cowan, Art Sensenig et al., Health Affairs 24 (Jan/Feb 2005): 185–194; Centers for Medicare and Medicaid Services, National Health Expenditures Statistics.
 MedPAC Report to the Congress, March 2005, p.7.
 Health Spending Projections Through 2013. Stephen Heffler, Sheila Smith, Sean Keehan et al., Health Affairs Web Exclusive, February 11, 2004.
 Use Of Medicare Claims Data to Monitor Provider-Specific Performance Among Patients With Severe Chronic Illness. John E. Wennberg, Elliott S. Fisher, Thérèse A. Stukel, Health Affairs Web Exclusive, November 7, 2004.
 Controlling Health Care Costs. New England Journal of Medicine 351 (Oct. 14, 2004): 1591–3.