Warner Bros. Television/Everett Collection

Joely Richardson and Dylan Walsh in the television series Nip/Tuck


Most experts agree that the central problem with the US health care system is its high cost. We can’t afford universal coverage unless there is much better control of medical expenditures, which are now reaching over $2.5 trillion per year. What’s more, without effective control of health costs the federal budget deficit and the national debt will continue to increase. Nevertheless, our political leaders have decided to expand and improve insurance coverage first, while deferring any serious attention to costs. Moreover, as I will discuss in the second part of this review, the book by John E. Wennberg demonstrates that in many parts of the US, costs are driven up by an excessive supply of medical services.

In March, after more than a year of bitterly partisan congressional debate, a narrow majority of exclusively Democratic lawmakers passed the most extensive health care reform since Medicare and Medicaid were enacted forty-five years ago. As described by Jonathan Oberlander and Theodore Marmor in these pages,1 the main thrust of this extensive legislation is to provide federal aid for mandatory expansion of coverage by Medicaid and by private insurance plans, and to expand benefits under Medicare. It has also been promoted by its sponsors as a measure to control costs, but it is not. Oberlander and Marmor make very clear that there is little reason to expect it will do much in the near future to control the relentless rise in health expenditures—a task that must wait for future reforms.

Another major defect in the new health legislation is its failure to change our current dependence on private, for-profit insurance plans—a young but now dominant industry that scarcely existed a few decades ago. By mandating and subsidizing the purchase of private insurance for almost all those not eligible for such government programs as Medicare or Medicaid, the legislation has created a virtual monopoly for the private insurance industry. True, the law also restricts some of the industry’s worst practices, such as denial of coverage because of preexisting conditions or rescinding coverage because of an expensive illness. However, it imposes no effective controls on the price private insurers can charge for premiums, which undoubtedly will continue to rise. For example, insurers will be allowed to increase premiums by a maximum of threefold simply because a policyholder has reached a certain age.

Before paying doctors and other providers of care, investor-owned health insurance companies now spend as much as 15 to 30 percent of their premiums to cover their many overhead costs, which include extravagant salaries and bonuses for top management, dividends for shareholders, and retained corporate profit. The new legislation puts limits on this overhead, but the total bounty extracted from the health system by the industry will surely increase because of a greatly expanded market for private insurance.

Because of its overhead, as well as the expense of billing and collecting it imposes on doctors and hospitals, the investor-owned for-profit insurance industry probably adds at least $150–$200 billion to the annual cost of providing health coverage to the American population, as compared with government-run programs such as Medicare. These added costs will become even greater under the new law because the industry will grow. Furthermore, insurers greatly dislike government regulation of their business. So if the private sector of our health system continues to be dominated by for-profit insurance plans, the industry’s well-financed lobby and its political influence will probably prevent any future reform proposals that might threaten its income. I find it difficult to believe that this industry contributes anything that is close in value to what it costs the US health system.

Perhaps the most significant defect in the new legislation is that it does not change how medical care is organized, paid for, and delivered. Although stricter regulation or outright elimination of the investor-owned insurance system could save much money, broad reform of the medical care system itself could save even more. To understand what is needed and why requires a closer look at how US medical care differs from that in most other advanced countries, and what makes it so expensive.

The US devotes a much larger fraction of its GDP to health care than other advanced countries—nearly twice their average. We spend, in US dollars per person, two and a half times as much as our counterparts in Europe. The most important reasons for the uniquely high costs of the US health system are its commercialization and the effects of business incentives on the provision of care. The US has the only health system in the developed world that is so much owned by investors and in which medical care has become a commodity in trade rather than a right. In these pages last year and in a recently republished book2 I have described the industrialization of medical care in the US, and explained how this transformation has contributed to its rapidly growing expense. We spend more than enough to provide good care for all our people; the problem is with our system, not lack of money.


Factors that are common to many advanced countries, like the aging of populations, the increasing prevalence of chronic diseases, and the development of expensive new technology, also drive up costs. These broad trends explain why other countries also are contending with inflation of their health costs. But only in the US are there such strong business incentives to supply ever-increasing amounts of expensive technical care, even when it is not medically needed. Relatively unconstrained by the kinds of government regulation that exist in other countries, these incentives cause US physicians and hospitals—both for-profit and not-for-profit—to act like businesses and to maximize their income, thus increasing health expenditures. Elimination of these incentives, with appropriate change in the organization of care and the behavior of health care providers, would undoubtedly produce huge savings even while making medical care more efficient and effective.3

Of all the providers of medical care, physicians are most important in determining how much will be spent. In the US they account for only about 20 percent of medical expenditures, of which about half they use for expenses. But in treating patients, physicians call on the facilities and services of all the other providers of care—hospitals, imaging centers, diagnostic laboratories, manufacturers of drugs and equipment, etc.—and thus they control most medical expenditures. Unlike most other sectors of the economy, where interactions between buyers and sellers of goods and services determine the allocation of resources, physicians—the key providers of medical care—make most of the allocation decisions on behalf of the consumers of medical care. In short, most patients agree to whatever services doctors say will help them.

A growing number of physicians and providers in the US are giving patients more information that will assist them in participating in decisions about their own care. This is a salutary development, but it will not change the basic role of the physician as the adviser and expert on whom sick or injured patients depend. The more serious or worrisome the illness or injury, the greater the dependence of the patient. It follows that if we are going to reduce the cost of medical care significantly, physicians will have to change the ways they now use all medical resources, whether drugs, diagnostic tests, surgical procedures, or hospital facilities.4

The legislators who voted for the new health care laws either did not understand the part played by physicians in overall health care expenditures or felt unable to do anything about it, because the legislation does nothing much to change their behavior. Despite its many provisions, the law mainly concerns the insurance system, which still pays physicians largely on a piecework basis, by fee-for-service. This form of payment is an incentive to provide more services than necessary and thus increase the use of resources. The legislation calls for experiments with new types of physician payments and new organizations of providers called “accountable care organizations” (ACOs) that might encourage more efficient medical practice, but its immediate and major effect is simply to expand coverage in a basically unchanged health care system.

In this respect the federal reforms resemble the legislation passed in Massachusetts some four years ago that mandated near-universal coverage but made essentially no provisions for containing the costs that would inevitably ensue. Massachusetts is now struggling with its costs and is being forced to curtail health services.5 Last year a special state commission for controlling costs suggested replacement of the fee-for-service payment method by global payments to groups of providers in ACOs that would have to be created. Although many experts in Massachusetts agree that fee-for-service is a major impediment to the control of costs, the current payment system is so profitable for most medical providers that they are not inclined to change it. The state legislature therefore recently decided to postpone any action on this idea.6

The Massachusetts experience is likely to be repeated nationally in the next few years. The seemingly optimistic reports of the Congressional Budget Office (CBO), which were important in supporting the Democrats’ legislative proposals, should not be misunderstood. The deficit reduction predicted by the CBO referred to covering only the added costs of the legislation for the first ten years, not to the likelihood of stemming the continuously rising costs of existing federal programs. Furthermore, the CBO is concerned only with the federal budget, not state budgets, and not the financial condition of the entire health care system, private and public. But even considering only the government’s burden, the current rate of inflation in health costs—between 4 and 6 percent per year—obviously cannot be sustained.


Why is this the case? First, the present trajectory of federal health expenditures predicts continued rapid growth of Medicare expenses and the exhaustion of the Medicare Part A fund—which partly covers hospital costs—within a decade or so. Second, most economists agree that the states will not be able to pay the rising costs of Medicaid in future years, when millions of beneficiaries will be added to the rolls. And looking at the private sector, there is increasing evidence that the inflation in the cost of health insurance cannot be supported by employers and employees much longer. In sum, the whole health system, if not radically transformed, seems headed toward bankruptcy.

As for the sanguine CBO estimates of the affordability of the new legislation, Douglas Holtz-Eakin, a former director of the CBO and chief economist for the President’s Council of Economic Advisers under George W. Bush, recently published, along with Michael J. Ramlet, an analysis that challenges those estimates.7 They predict that the legislation will aggravate, not lessen, federal budget problems. They doubt that the new reform will contain costs or generate the predicted additional revenues, and they conclude that it will probably produce additional federal deficits of about $562 billion in the first ten years. This, of course, like the original CBO analysis, is merely educated guesswork, and it underscores the uncertainty of all estimates of the future costs of health care—though there is broad consensus that they will continue to increase.

Barack Obama
Barack Obama; drawing by John Springs


The cost crisis now facing the US health care system urgently calls for more effective control than the new legislation provides. That is why a new book by Dr. John E. Wennberg, Tracking Medicine, is so important and timely. It offers a perspective on the health care system quite different from that reflected in the new legislation. It focuses on the supply of medical services, not on the lack of insurance coverage, and it provides convincing evidence that oversupply of services throughout the US adds greatly to the cost of care. This evidence rests on an ingenious analytical approach devised by Wennberg and his colleagues, which compares health care expenditures in many different regions of the country.

Wennberg founded the Dartmouth Institute for Health Policy and Clinical Practice and has spent his career studying the distribution of health care. Nearly forty years ago he made the seminal discovery that there were large differences between neighboring Vermont towns in the frequency (i.e., incidence per population) with which tonsillectomy was performed. These differences could not be explained by medical need or by the sociodemographic characteristics of the populations involved, and were not accompanied by any discernible differences in the health of the children in these towns.

These initial observations were followed by the discovery of similar regional disparity in the provision of other medical services elsewhere in New England. For example, differences of 50 to 100 percent between Boston and New Haven in rates of hospitalization for certain common medical conditions could not be accounted for by differences in prevalence of the medical conditions in the two populations.

Such large variations in medical services occur widely, as Wennberg and his colleagues later found when they began to analyze the masses of data collected by the federal government on payments for the care of Medicare patients. They devised new methods for studying the frequency of selected common medical services for the treatment of chronic diseases among the elderly as they neared the end of life (this helped correct for any regional differences in the average severity of illnesses). Their results were published in many articles in peer-reviewed medical journals and in a series of richly documented volumes from the Dartmouth Institute (The Dartmouth Atlas of Health Care).

The data revealed as much as two- or threefold variations in the use of services when the country is divided into 306 hospital referral areas. The variations in Medicare services resembled those Wennberg first found in New England. They closely coincided with the availability of hospital beds and physicians in those areas, but not with the sociodemographic characteristics of the populations served or their need for medical services. Furthermore, he found no evidence that these variations in medical services resulted in different overall outcomes in the health of patients.

The general implications of the research by Wennberg and his Dartmouth colleagues are now widely accepted. Other researchers repeating some of the Dartmouth studies have come up with slightly smaller estimates of the extent of the area variations, and there have been a few technical criticisms of the Dartmouth methodology, but there is general agreement with Wennberg’s central conclusions, and he has not changed his basic message. There are indeed significant regional variations in expenditures on medical services, unexplained by differences in medical need or health outcomes, but correlated with the numbers of specialist physicians and the availability of hospital beds in each area.

Wennberg believes there probably is excessive supply of specialist doctors and hospital beds in the high-service areas—leading to a higher total of payments—rather than undersupply in the low-services areas. This latter conclusion has been challenged by some high-expenditure teaching hospitals that claim the Dartmouth group has not taken sufficient account of their more advanced care for the sickest patients. But Wennberg’s colleagues believe they have properly controlled their studies for severity of illness and they believe this criticism is invalid. Furthermore, much of their data refers to regions rather than specific hospitals.

In any case, basing his analysis mostly on unchallenged observations, Wennberg draws important conclusions about the US health care system. First, he argues, expenditures are largely driven by the supply of services. His data convincingly show that the use of medical resources in the US is not determined by the interplay between supply and demand, but mainly by supply. However, he is ambiguous about what he means by “supply.” Sometimes he refers simply to the supply of facilities, particularly hospital beds available for the care of acutely ill patients. At other times he seems to be focused on the supply (and specialty mix) of physicians, thus acknowledging that physicians make most of the deci- sions to use medical facilities and other resources.

Wennberg’s second conclusion is that since the medical care in the low-expenditure areas is not discernibly different in quality from that in the high-expenditure areas, a huge amount of money could be saved if all of the country were to receive care the way it is provided in the low-expenditure areas. Wennberg estimates the savings would be about 30 to 40 percent of the total now being spent on health care services. Some idea of the magnitude of that estimate can be gained from the fact that we now spend over $2 trillion on medical care.

To realize such savings, he says, would require the development of organized systems of health care delivery, because there is evidence that organized systems, particularly multi-specialty group practices such as the Mayo Clinic and some well-managed hospital networks such as Intermountain Care in Utah, can deliver excellent care at lower cost. Since Wennberg believes that about 20 percent of all care is primarily determined by patients’ choice, elimination of unnecessary services would also require, as he explains, “establishing informed patient choice as the ethical and legal standard for decisions surrounding elective surgeries, drugs, tests and procedures, and care at the end of life.” And finally, it would require a major investment in developing the “science of health care delivery”—the study of how best to use resources to maximize desired health outcomes, so that physicians can base more of their decisions on solid experimental evidence.

In an epilogue written with Shannon Brownlee (the journalist and author of the excellent 2007 book Overtreated8), Wennberg assesses the recent health care legislation, focusing not on its provisions to expand insurance coverage, but on its prospects of improving medical care. He applauds the law’s proposal to study changes in the way care is paid for and medicine is practiced. But in another section of his book he acknowledges that “these changes will take time and experimentation before we get it right. In the meantime…health care costs must be held in check.”

To do that he suggests five strategies for improving the present system without waiting for more extensive reform, three of which involve the manipulation of Medicare fee-for-service payments to facilities and physicians. A fourth would involve controlling the size and specialty composition of the physician workforce to limit the increase in specialists and expand the proportion of primary care physicians. And the fifth would involve giving physicians and hospitals feedback information on their performance, with the expectation that this would improve their effectiveness. Since none of these strategies changes the basic economic structure of the system, I doubt they would be very effective in controlling costs. We need major reform, not tinkering with the present system.

Nevertheless, Tracking Medicine is a significant book because Wennberg assembles good empirical evidence showing that the dysfunctional delivery of medical care is a major cause of the crisis in health costs. He shows that correction of abuses in the insurance system alone, even the achievement of near-universal coverage, will not solve the crisis. Expansion of coverage will probably exacerbate the problem because insurance premiums will surely increase. Wennberg’s studies and the work of his Dartmouth colleagues make a convincing argument that a sustainable health care system in the US will require not only insurance reform but fundamental changes in the way medical care is organized, delivered, and paid for. Although the book does not spell out what changes will ultimately be needed to improve the delivery of health care, by identifying overuse of services as a key culprit it does suggest the direction future reforms should take.

But what puzzles me is Wennberg’s failure to highlight the extent to which economic incentives for health care providers, particularly physicians, lead to overutilization. He seems focused simply on the evidence for the overuse of services and the underlying excess capacity of our health system. However, excess capacity is mainly the result of unregulated entrepreneurialism by providers, in response to the incentives of a commercialized system. When most medical institutions—hospitals, ambulatory facilities, nursing homes, etc.—are managed like businesses (and so many of these facilities are investor-owned), and when most physicians are paid on a piecework, fee-for-service basis, unnecessary expansion is inevitable. Doctors and other providers have a vested interest in continually increasing the amount of medical services they provide; and as they do so insurance companies also increase the premiums they charge.

Wennberg’s painstaking documentation of overuse as a cause of excessive costs greatly helps our understanding of the US health care problem because it shows that costs could be controlled by eliminating unnecessary care, without rationing medically appropriate services. The clear implication of his work is that we could afford good care for all if we made our medical system more efficient and less wasteful.

What he does not emphasize is that to produce such change will require elimination of the economic forces that have made medical care a commodity in trade instead of a social service, and have transformed our health care system into a profit-seeking industry. Until we join other advanced countries in treating medical care as a right and not a business, we will have to wait for control of health costs.

This Issue

September 30, 2010