Afew years ago, the State of Massachusetts passed legislation that facilitated near-universal health insurance coverage for its citizens.1 Massachusetts now enjoys health outcomes comparable to those of other industrial countries and generally superior to the rest of the United States. Despite these achievements, its level of health spending and rate of health care inflation remain stubbornly high (at about the national average). As a result, the state is now considering legislation to reform dramatically how health care is delivered. The Massachusetts experience thus provides a cautionary lesson for the U.S. Congress as it seeks to achieve similar objectives in reforming the national health care system.
The challenge for both Massachusetts and the country will be to curtail the significant inefficiencies that make the United States the highest spender on health among all industrial countries. In order to be viable, health care reform must not only expand the scope of coverage but must also significantly slow the growth of spending, which, in the words of President Obama, places an “unsustainable burden on the taxpayer.”
The reforms likely to pass may fail to tackle either the sources of existing inefficiency or the causes of medical cost inflation. But they also will almost certainly fail to acknowledge and address the ethical and economic policy tradeoffs that confront all countries in a world where technological progress in medicine is allowing greater sophistication in improving health but at the cost of rising health care spending.
This discussion has three purposes: first, to briefly suggest the factors underlying both the inexorable rise of health care costs and the sources of U.S. health care inefficiency; second, to assess the various reforms now being considered by the Congress; and third, to offer some advice about how to think about the problem we face.
Structural Causes
of Cost Escalation
All advanced countries confront the challenge of containing health care costs. U.S. annual real health care cost growth per capita exceeded real GDP growth per capita by roughly 2.7 percent per year during the 1995–2005 period. Some industrial countries experienced even higher growth (4–6 percent in the Netherlands, Norway, Australia, Belgium and the United Kingdom); some roughly comparable rates (as in Japan and Italy); and a few somewhat lower rates (under 2.5 percent in France, Canada and Germany). This suggests that medical care cost growth arises from fundamental forces independent of the peculiarities of any single nation’s health care system.
Ironically, the most important of these forces is scientific and technological progress. Unlike most sectors of an economy, where technological progress allows for higher productivity and cost savings, this hasn’t proven true in the health sector. In most sectors, technological innovation yields cost savings with the substitution of abundant capital for relatively expensive labor. But in health care, advanced technology or machinery typically cannot substitute for medical professionals. So a greater demand for medical capital does not necessarily reduce the demand for medical labor; often, it increases it. The more ways there are to diagnose and treat illness, the more people will be diagnosed and treated, and the more caregivers will be involved.
One must also recognize that while technological progress may give doctors new ways to diagnose and treat illnesses, they may be costly to procure regardless of capital/labor ratios. Imaging equipment or genetic tests are obvious examples. They allow doctors to diagnose medical problems that previously would not have been detected, but also create new demands for therapeutic services. Even when a technological innovation reduces the overall cost of addressing a health problem—say, robotic surgery that halves recovery time in the hospital—overall costs can rise. For the individual patient being treated, the innovation may save money. But the innovation will allow many individual patients to be treated who previously would not have sought or been provided with surgical treatment. Thus, the new technology may save resources and yield health benefits at the individual level, but its provision to a larger population adds to overall health spending.
Finally with regard to medical science and technology, we must acknowledge that our cost problems are probably just beginning. We must assume that technological progress in the medical sphere will continue and new approaches to diagnostics and treatment will be developed that will be costly under any reasonable patent regime. Society will want to adopt these if at all possible. Not only is there little we can do about this; there is little that any reasonable person would want to do.
The precise shape of future innovations, however, will depend on what motivates academic researchers, the research funding decisions of national health institutes, and the profit-driven investment decisions of private pharmaceutical and medical equipment producers. There is no guarantee that new medical technologies will address the most urgent or costly medical problems, or that they will focus on cost reduction as an objective. Some new medical techniques will surely prove highly effective at solving serious problems; others will fulfill desires of little social consequence. Indeed, in the United States at least, the private health insurance system implicitly supports research on profitable new products or processes. If a product is effective and is prescribed by physicians or demanded by consumers who are exposed to heavy marketing, its use will be financially underpinned by the insurance system. If it is more profitable for manufacturers to direct their efforts toward modest discretionary enhancements instead of what we think of as medically necessary treatments, then overall costs might rise without achieving proportional health benefits.
A second major necessary source of cost escalation arises from demographic factors. Many worry that the aging of the baby-boom population, coupled with increases in life expectancy, will trigger increased medical spending. Health spending by the elderly, on average, does substantially exceed that of younger age groups. This higher spending in part reflects the high costs associated with dying. But it is also true that the elderly are not only living longer but living in better health, and with health outlays not that much greater than other age groups. Indeed, the high costs attendant on death principally relate to those who die early—say in their sixties or seventies—because of poor health behavior (due to obesity or smoking) rather than to those who die much later. While there is no clear consensus, on balance OECD and European Commission studies do suggest that, even under conservative assumptions, the larger number of baby boomers in retirement years will increase pressure for higher medical outlays, especially the need for long-term care. The incidence of mental dementia rises sharply among those in their eighties (by a factor of three) and further doubles for those over ninety. More than a third of those over ninety have Alzheimer’s disease or another form of dementia. Thus, the aging of the baby boomers could be accompanied by a substantial increase in long-term care outlays unless there is a pharmaceutical breakthrough for dementia.
A third likely factor in cost acceleration is counterintuitive, but real: wealth. Health spending typically rises more rapidly than household income. Assuming that per capita income will continue to rise, this would constitute another reason why health spending may rise more rapidly than GDP.
Idiosyncratic Causes of High Spending in the United States
The United States spends far more on health care than any other industrial country. Some of this excess reflects cultural proclivities rather than structural or policy factors. It is worth acknowledging these factors since they would be very difficult to change rapidly, even if one were to undertake dramatic changes in the health care system. Four factors in particular should be noted.
First, the United States has a relatively high prevalence of chronic health conditions. In part, this reflects a higher than average obesity rate as well as a larger proportion of the over-fifty population that smokes or has smoked in the past. One estimate suggests that 5–7 percent of U.S. health care costs in the late 1990s was attributable to obesity, compared to 2–3.5 percent in Canada, Australia and New Zealand. The cost of health care services is a third higher for obese people than for normal-weight people in the United States, and the cost of their medications 77 percent higher.2
Second, the U.S. medical system more aggressively screens for diseases than other countries and is more likely to treat less severe cases of a disease. This may contribute to a higher diagnosed prevalence rate for many illnesses, particularly cancer, as well as a higher rate of treatment procedures, without suggesting that Americans are actually less healthy than in those in other countries.
Third, regional health-cost differentials are substantial for Medicare patients. There are various ways to explain this, but these differentials are unlikely to be easily reduced.
Finally, while America appears to buy less health than other countries per dollar spent, it does buy amenities that Americans value: the ability to choose among practitioners (both primary care physicians and specialists) and medical institutions; the ability to obtain sophisticated medical treatment when needed; the easy availability of many pharmaceuticals; widespread access to sophisticated diagnostic equipment; rapid access to many forms of medical care; the right to obtain legal redress for perceived medical errors; and a fundamental role for the private sector in the financing and provision of medical care. Americans, evidently, value choice and are willing to pay for it.
But when analysts disentangle why the United States spends so much on health care—16–17 percent of GDP compared to 8–12 percent in many other industrial countries—several more policy-related issues usually stand out. These include:
Clearly, this list suggests that many agents in the U.S. health care system profit from its structure relative to what might result from less costly alternative approaches. This is not a simple case of greed. These manifestations of “inefficiency” are rather direct consequences of the structure of incentives embedded in how the U.S. health care system is organized. Some result directly from U.S. government policies. While it may be difficult to address many of the structural factors propelling health care costs upward, our ability to deal with these idiosyncratic factors will be principally determined by whether we have sufficient political will to do so.
But summoning that will is not easy. Whether the profits or earnings derived from current practices are appropriate or not, they are real to those who receive them and significant in the U.S. economy. Many people—not only those directly benefiting from the present system but also others who have invested in companies associated with it—would stand to lose, through no fault of their own, from policies that change the way medical care is provided. If we must have bitter arguments in the process of reform, we should at least fix the obvious problems, rather than merely shifting them around by means of offsetting political compromises. That means that we must find the torque points in the present system that can eventually make a difference.
One of these torque points concerns the way the U.S. health care system is financed. The United States has a fee-for-service system in which many hospitals and nearly all physicians work on an individual for-profit basis. This partly explains the high density of sophisticated technology in many urban areas, as hospitals compete with one another to market their competencies. And it is why this technology is often used to excess, for only by using and charging for it can they afford it. Other countries do things differently and obtain satisfactory outcomes, but Americans tend to resist the limitations on individual choices that these systems impose. While we need to do things “our way”, this could still include creating new incentives for doctors to work in complementary medical groups or those in which they receive salaries rather than a fee for services.
There is more to the financing dilemma than that. Most Americans get their health insurance through their place of employment. There are alternatives to this approach. But this is unlikely to change, judging from the legislation emerging from Congress. Because we have chosen to make health insurance benefits a form of nontaxable compensation, employees have pushed hard to increase the generosity of these benefit packages in terms of covering the most likely conditions as well as having small deductibles and low co-payments. These features reduce the incentives of insured consumers to be prudent in their health care choices, fostered at times by physicians because a “third party” (the insurance plan) will underwrite the cost of tests, referrals and prescriptions. The phenomenon of third-party payment is not by itself a cost driver; other markets function efficiently with third-party insurance payments (for example, auto, veterinary care, and fire insurance). But the unique linking of health insurance to employment has created incentives that have driven up costs.
Moreover, efforts by private insurers to limit fraud and abuse often underpin the high costs of the system. They do so, for example, by limiting treatment alternatives (requiring hospital care as opposed to home care, say, in order to prevent non-medically related charges from being billed as medically related), and by refusing to provide incentives for preventive care—addressing obesity, for example—that are hard to monitor and are more easily abused than other medical services.
Fixable legal problems represent another torque point. Currently, patients must resort to litigation if they believe a doctor has made an error. Patients are not made aware of what constitutes a “reasonable standard of medical care in accord with accepted clinical practice guidelines”, and peer reviews by medical associations rarely cast unfavorable judgments on fellow physicians. In other markets, governmental regulatory bodies enable consumers to know what would constitute a reasonable standard of quality and provide mechanisms for recourse before the need for litigation.
Another deficit in our regulatory regime is the absence of restrictions against various forms of conflict of interest. Too frequently, physicians receive indirect forms of compensation from pharmaceutical companies in return for prescribing their products. Self-referral by physicians to diagnostic centers they own is another problem. In other markets, government regulations limit these kinds of market distortions.
Other government policies contribute to higher costs and influence both the structure of supply and demand in the health care market. For example, by legislative action, Medicare has been precluded from negotiating better prescription drug prices, thus raising the cost of sole-source drugs and other medications. Medicare is also limited in its ability to import generic drugs. Also, since we seek a market that allocates by price, stemming cost escalation requires more incentives to increase the role of supply factors in health care markets as well as influencing demand. Government inaction also reinforces information asymmetries in health care. Patients have to rely on what medical professionals tell them. Their access to a government-level resource giving them the ability to make comparisons among treatments, drugs, hospitals and so forth is essentially nil. The Internet helps some, but it is a resource of unknown reliability. We can address this problem, too; we just haven’t done so.
On the Table
To what extent does the health care reform legislation emerging from Congress reflect a comprehensive understanding of the problems and provide effective solutions? Is debate moving us closer to an “efficiency point” where we will see improved or at least unchanged performance of the health system in terms of what it can achieve, but with spending levels much closer to what we observe in other industrial countries? And are we approaching this efficiency point while still satisfying the values Americans have affirmed—values like choice, freedom from health care rationing, and a heavily privatized medical care system intermediated by private insurance?
To answer these questions, let’s focus on the intersection of two key points—insurance coverage and cost containment, recognizing that the issues are so complicated, the legislative process so complex, and the sheer length of any potential bill so immense that it is impossible to do more than highlight some major themes.
In the U.S. context, the thrust of the legislative efforts for achieving universal coverage has focused on satisfying the following objectives:
The most likely legislation now appears to require individuals to have insurance and large businesses to provide insurance, or the means for insurance, for their employees. Subsidies would help low-income households finance the purchase of the minimum required insurance, and penalties would be levied on individuals who fail to purchase it. Various ideas have been advanced about how to use taxes and subsidies to ensure that the minimum plan is purchased or provided by employers. Individuals and employers could still buy more coverage than the required minimum. Regional exchanges would be established to facilitate households and small businesses “shopping around” to obtain competitive insurance rates. Insurance companies would be required to provide the minimum health insurance policy independent of the health status of the individual applicant. Finally, under the House reform bill, a new government insurance plan is being considered as a “public” option. It may serve other objectives, but is not likely to be effective at promoting equal competition with private insurers because the playing field would not be level. The private companies need to hold capital to stave off failure, but the public entity does not; being public, it has recourse to government coffers.
These policies could affect overall health spending in several ways. Insurance companies would no longer need to spend resources to vet applicants to minimize their exposure to insurees in poor health, since the overall health risk would be the same for all carriers. Presumably, too, companies would have incentives to offer policies on competitive terms, lowering prices in a market with a larger pool. But there is no guarantee that companies would save on resources currently dedicated to marketing, since they would still be competing for clients, and present expenditures for administering benefits would grow with a larger number of clients, notwithstanding the attempt by the legislation to limit spending on administration and profits. The presence of a government insurance option, if adopted, might provide an additional mechanism to contain costs, though this is by no means certain. Efficiency will be gained, and some cost savings might even arise if those presently uninsured no longer have to seek care from expensive hospital emergency rooms but rather go as outpatients to primary-care physicians.3 This would also enable more preventive care, and facilitate the provision of other medical services at lower cost.
Less clear is whether these reforms, valuable in themselves in terms of enhanced coverage, would affect cost growth. The proposed legislation does not appear to include mechanisms to tighten eligibility and business practices for particularly high-cost or less-than-cost-effective drugs, tests or medical procedures. Moreover, the provision of government subsidies to facilitate coverage for those presently uninsured will require Federal spending. These increases will affect overall health spending, but to the extent the legislation is effective at making the private insurance markets function, they will not necessarily be reflected as higher Federal outlays.
Moreover, if health insurance is available to most of the 47 million currently uninsured, this should increase effective demand for health services, even if it lowers average insurance premiums and produces the efficiencies noted above. Unless measures are taken to increase supply and change incentives and practices that have lowered the value-for-money quotient, further cost pressures are likely to arise from this additional effective demand, no matter what mix of current proposals is eventually adopted.
One approach not included in the emerging legislation, unfortunately, is coverage in the mandatory plans for catastrophic conditions and government support for the reinsurance of catastrophic claims. Such reinsurance would effectively allow the splitting of the insurance pool into two levels. Individual insurers would thus be exposed only to the most probable and predictable health risks, and would be free to manage these risks as they wish. No one insurer would be at a competitive disadvantage with regard to its exposure to particularly expensive health risks. All insurers would share proportionally high-cost, low-probability catastrophic claims that require expensive surgical procedures or extremely high-cost drugs. These claims might result from unusually high-cost and rare diseases, chronic conditions that require continuous care (severe stroke or neurological conditions) or accidents requiring expensive life-saving surgery. Under such an approach, once a private insurer would reach the catastrophic cutoff for a claimant, all excess claims for that claimant in that year would be shared in a re-insurance pool. If such a structure and re-insurance pool membership were mandatory for every health insurer, insurance companies would no longer have to try to compete for the “best” health insurance risks to offset the risk of catastrophic cases, because they would all share that higher risk category equally.
Beyond insurance-related factors, the emerging legislation includes several policy ideas to limit the pace of cost growth in the health sector, especially in regard to Medicare. These include adjusting Medicare reimbursement policies to encourage health care providers to improve productivity; trying to link hospitals’ Medicare payments to the “quality of care received” (so-called value-based purchasing); imposing penalties on hospitals with high readmission rates; improving coverage for preventive and educational actions that hospitals could take to prevent readmissions; and providing incentives for doctors to spend more time on primary care or to expand prevention and wellness activities. Another proposal would seek to improve Medicare quality and extend its solvency through policies based on the recommendations of a new independent Medicare Commission. Not included in the package of Medicare reforms, unfortunately, would be some linking of reimbursement rates to the practices of physicians in states where costs are relatively low.
The likely impact on spending from such reforms is difficult to judge, given their present lack of specificity. There is no guarantee that reductions in Medicare spending will not be offset by higher spending by the elderly through private insurance coverage. The link between most of these proposed reforms and actual medical practices and procedures with respect to the non-elderly is even less clear.
Another reform being discussed involves strengthened use of cost-effectiveness analyses (CEAs), at least on a pilot basis. In principle, the impact of CEAs on medical spending decisions would hinge on whether the results of such assessments are actually used in setting Medicare reimbursement levels, and whether they also inform doctors and private insurance companies in their decisions with regard to the non-elderly. In the United States, the resistance of pharmaceutical companies, device manufacturers and medical practitioners to CEAs seems to have ensured that the emerging legislation will not apply such restrictions. Moreover, even if CEAs were more rigorously used, their application to medical decisions is not easy and would be unlikely to prevent cost increases associated with technological progress.
A further reform that appears politically untouchable relates to eliminating the current exemption in the U.S. tax code for employee health insurance benefits. If adopted, the Senate proposal for an excise tax on generous plans may offset some of the tax losses and reduce demand to some extent (unlike the House bill, which simply raises marginal income tax rates on the wealthy and without any link to the sources of health care demand). But continuation of the exemption means that households would still lack a clear picture of what health insurance and medical care actually cost, although the legislation does try to enhance transparency in this regard.
Finally, the forthcoming legislation only tepidly encourages greater use of accountability care organizations (ACOs) of the kind now being considered by Massachusetts. In ACOs, doctors would work within a network and be reimbursed accordingly rather than on a patient-specific fee-for-service basis. (Massachusetts is again a trend-setter in promoting this idea, and its experience will attract much attention.) This idea draws on the success of the Mayo and Cleveland Clinics. In some contexts, doctors would effectively be paid salaries based on such factors as the nature of their specialties, their seniority, how many patients they treated and how many hours they worked. Primary-care physicians would serve as “gatekeepers”, determining whether a patient would see a specialist within the network. The gatekeeper might be charged with making judgments as to the necessity or “value for money” of providing a high-cost diagnostic test or treatment. These efforts also seek to foster collaboration among doctors at different phases in the treatment process, particularly in relation to chronic diseases. In principle, such an approach should realign incentives to be more compatible with cost containment. But the current legislation promotes ACOs only on a pilot basis, and does little to encourage doctors to join such networks.
The experience this year with realizing health reform highlights the political challenges and difficulties of changing a system so peculiarly American, and with so many vested interests clinging for dear life to resist any significant reform that would cut their earnings or profits. Perhaps this explains why the legislative emphasis thus far has largely been about insurance coverage expansion rather than about cost inefficiencies, or cost acceleration. And make no mistake about it: Success in increasing health insurance coverage, while reducing some inefficiencies in spending, such as the overuse of emergency rooms, will almost certainly lead to an increase in the demand for health care.
In contrast, the pro-efficiency aspects of the legislation appear either inadequate in size or uncertain in effect. Many new incentives and penalties being considered may affect the supply side, particularly as it relates to Medicare reimbursement policies and potentially, though far more indirectly, the non-Medicare market. But these linkages are still too indirect to provide much confidence that current inefficiencies in spending will be significantly contained. Most facets of the present market structure, prevailing business practices, failures in the regulatory structure, inappropriate government policies and controls, and misaligned provider and consumer incentives remain entirely unaddressed by the reforms being considered. Indeed, some of the legislation being discussed in Congress would prevent effective restrictions on some of the most questionable practices—notably, tie-ins with already existing physician-owned facilities. Proposals for more Federal control would also constrain experiments by the states, arguably the most useful laboratories for solving the problems before us.
We can illustrate the problem simply by asking a series of straightforward questions: Do any of the efficiency reforms incentivize the wide-scale establishment of accountability care organizations or other alternatives to fee-for-service financing? Do any address unjustified disparities in health care spending across different regions or duplicative spending on costly diagnostic equipment in many localities? Do any offer financial incentives for medical students to pursue training in primary care, pediatrics and geriatrics rather than specializations for which the provision of care may do less to address the sources of poor health in our population? Do any create incentives to increase the supply of health care professionals to meet the demands of the millions of newly insured? Do any limit self-dealing or tie-in relationships with medical device and pharmaceutical companies? Do any significantly address the defensive medical practices associated with concern over malpractice, and provide physicians with confidence that following acceptable clinical practices will keep them from being sued? Do any allow consumers to judge whether their physicians are performing well? Do any facilitate access to less costly generic pharmaceuticals? Do they ensure that the results of comparative effectiveness studies underlie policies on health insurance reimbursements? Do they encourage the allocation of NIH research dollars toward the problems that are likely to mushroom health care costs in the future? For all these questions, the answer is “no”, leading to the conclusion that Congress and the Administration have largely been having the wrong conversation about health care.
Warning: Tradeoffs Ahead
If we have to choose between spending less and getting less or spending more and getting more, how should we think about the choice?
In our view, it is not self-evident that American society should want ultimately to spend less on health care, or even less per capita. Indeed, we might wish to spend more. There is no way to assign a priori an optimal level of spending on health, for health is not a binary outcome. It defines a wide range of factors, around which one can imagine a wide range of cooperative behaviors between patients and caregivers. A society that chooses to increase the share of its income devoted to health would not be making an irrational decision. The benefits of much of what is achievable from today’s health technologies far exceed their cost, and this may prove even more the case as technologies in medical science further advance. As long as there is a social consensus on how to pay for such higher spending, it would be difficult to argue that we are spending “too much” on health.
But therein lies the problem: We don’t agree on how to finance what we desire, and we sense that the way we do it now is neither efficient nor fair. And as long as we lack a consensus on that basic question, vested interests within the system will have a much easier time getting their way and preventing major change. One of the reasons we have not had the key debate that must be prelude to achieving some consensus is the illusion that achieving efficiencies will make basic tradeoffs unnecessary.
Note that one key constituency in the real debate we need to have is the Federal government itself. When scoring the cost of reform, the focus of budget policy analysts has largely been on the impact on government spending and the fiscal deficit. This explains the concern that the legislative package be budget neutral—that the cost of extending coverage, which would lead to increased government outlays, is fully financed by various revenue-raising or cost-reducing measures. Given the prospective size of the U.S. fiscal deficit and the role of health spending in contributing to out-year deficits, this would seem a reasonable concern. But the criterion for a satisfactory health reform must necessarily be broader than concern for out-year Federal budget deficits. The right metric must also determine whether the reform limits the growth of total health spending, private as well as public. And if it does not, does it at least achieve significantly better health outcomes for the money spent?
Moreover, even if fully financed by higher taxes, a health reform package should not be allowed to effectively crowd out so much future fiscal space such that it limits the potential to finance other vitally important spending needs (for example, for social security, transforming America’s infrastructure, adapting to climate change and reforming education policy). What is striking in the policy debate to date is how little these broader fiscal concerns are being addressed.
It is not hard, however, to address them. If we decide that containing medical costs is our highest priority, then reform should include a “health spending rule” and a new governance structure in the form of an independent agency that facilitates adherence to it. Such a rule would mandate that action be taken in the event that the health care cost curve is not “bending” downward and that either Federal spending levels or total national health spending is on an unacceptable trajectory. The Senate Finance Committee version of the legislation does include language that is a tentative step in this direction, giving the independent Medicare Commission responsibility for recommending what to do if the cost trajectory exceeds certain targets.
This could mean de facto triage, or the potential rationing of health care. Many observers have pointed out that, for practical purposes, some rationing is inevitable. Just as not everyone can drive a Cadillac, not everyone can have every expensive health care service on demand. In fact, that is the status quo—some do not even have basic health care service. There are rational ways to guarantee everyone basic care, and leave the market to decide the rest, with more efficiency than we now have, but American politicians don’t want to go there. The result is that no one in the current debate has proposed the several steps outlined here that could both produce better “value for money” and limit overall health spending, and so we have to conclude that we are not really serious about containing costs.
This leaves us, then, with a prospective piece of legislation that is too small to deal with the real problems we face, while being based on the fantasy that we can avoid real tradeoffs through the deus ex machina of “efficiency.” Something’s got to give, and it’s anyone’s guess what that will be. Most likely, we will have to accept that, having moved closer to universal coverage, we will, soon enough, have to confront the even more difficult task of cost containment.
2See Samuel Preston and Jessica Ho, “Low Life Expectancy in the United States: Is the Health Care System at Fault?” Population Studies Center, University of Pennsylvania, 2009.
3But the resistance to inclusion of illegal immigrants means that this group, plus those still remaining without coverage, would still place some cost burdens on hospital emergency rooms, and thus continue to shift some of the cost burden onto the insured population.