Principles for Priority Setting in Mental Health Services and Their Implications for the Least Well Off

Per capita funding for mental health services in both private and public sectors has declined substantially in recent years (1,2,3,4,5,6). Studies of changes that occurred during the mid-1990s have identified reductions of 20 to 40 percent in inflation-adjusted annual mental health expenditures per patient in the private sector (1,2,3), the Department of Veterans Affairs (VA) mental health programs (4,5), and some managed care Medicaid programs (6,7), although considerable local variation exists within each of these systems.

These reductions have not been associated with evidence of a decline in either general demand for mental health services or in the needs of those who are least well off—those with the most severe symptoms and the most impaired functional capacity. For example, in VA, a system for poor and disabled veterans, the number of patients seeking mental health care has been increasing by 4 percent a year since 1990 (5). More generally, the epidemic of substance abuse among people with psychiatric illness continues unabated, increasing the need for specialized, intensive, and highly integrated services for these especially vulnerable patients (8). At the same time, reductions in state and city welfare programs, Aid to Families With Dependent Children (AFDC), Supplemental Security Income (SSI), and public housing programs, have made community survival increasingly difficult for people with low incomes, in general, and for people with serious mental illness in particular.

It has long been recognized that relatively small numbers of patients with the most serious illnesses consume a disproportionately large volume of health care services (8,9,10,11). Agencies and payers seeking to reduce expenditures will inevitably find the greatest opportunities for savings among patients with the highest levels of service use. Current circumstances, primarily fueled by the drive to reduce health care costs, thus threaten to reverse the substantial gains made during the 1980s in the design and implementation of specialized service programs for people with serious mental illness (12,13). Although a fully funded, nationwide health care system for people with serious mental illness has never been developed, an appropriate rehabilitation-oriented clinical philosophy has emerged, an array of successful demonstration projects has been implemented, and model service systems have been under development (13,14). The drive for cost minimization in more recent years has thus placed much that has been gained at risk.

Changes in the fiscal climate and in organizational culture of health care systems pose major challenges for decision making by clinicians, clinical administrators, program planners, and policy makers. What programs should be cut back or eliminated? Which programs, if any, should be expanded? What balance should there be between improving efficiency and maintaining intensive services for those with the greatest needs? If intensive programs are available on a limited basis, which clients should have access to them?

This paper attempts to articulate a series of principles and processes that should influence decisions about resource allocation and program development and considers their implications for providing intensive services to the most impaired. Efforts to justify program decisions are most often made on the basis of what treatments or programs are judged to be most cost-effective—that is, which treatments cost least and generate the most improvement in health status. For most conditions, however, data on the cost-effectiveness of specific treatments or programs are not available because appropriate studies have not been conducted (15). In addition, cost-effectiveness considerations are often not the only ones of relevance. This paper attempts to look both at, and beyond, the cost-effectiveness standard to systematically outline seven distinct principles that should, and often implicitly do, shape program decisions in mental health care.

Principles of resources allocation

Although each of the seven principles, which are summarized in the box at the top of the page, has merit, they may suggest conflicting preferences. Because they are not hierarchically ordered and no principle takes priority over others, their joint application does not always yield consistent determination of program priorities. Their value is to facilitate our understanding, from different perspectives, of how and why priorities may vary more than to provide a consistent formula for evaluating which programs should be given the highest priority and which individuals should have access to them.

Autonomy of individual patients' welfare.

Perhaps the most fundamental principle of medical ethics is that health care decisions must afford the greatest benefit possible for the individual patient untainted by considerations other than maximizing that individual's health and welfare (16). As articulated in the ultimate moral law put forth by the philosopher Immanuel Kant, no person should be treated as a means to accomplishing some other objective, but only as an end in himself or herself (16).

From this perspective one should never withhold treatment from a patient to achieve some other goal, such as to maximize personal profit or even to save resources for a patient with greater needs or for a patient with the potential to receive greater benefit from those resources. The imperative to fully serve the interests of each patient is reinforced, in practice, by the fact that patients typically lack the technical expertise to fully evaluate treatment options for themselves and must rely on health care professionals to be their single-minded agents. For a health care professional to do otherwise is thus not only a breach of Kantian ethics, but also a betrayal of implicit trust.

Much of the outcry against the changes in health care service systems in recent years has been based on the violation of the principle of the absolute priority of patient autonomy. To many observers, private managed care companies seem to have violated this principle by mixing the mission of restoring each patient to optimal health with the mission to enhance their own personal gain or the profits of shareholders. Applying the Kantian imperative to the most disabled patients obligates us to address their needs fully and without compromise, regardless of the higher cost, as we would address the needs of any other patient.

Involvement of local networks of patients, providers, and other stakeholders as arbiters of health system objectives.

Even if one accepts as absolute the value of maximizing each individual's health and recognizes that most health care professionals have greater technical knowledge than their patients, it does not necessarily follow that physicians or other health care professionals are the ones to define the needs, interests, and priorities of their patients or communities. Even in the absence of competing incentives for personal financial gain, health care professionals may not always make decisions that are consonant with patient preferences. The principle that medical decisions must ultimately be based in the informed judgment of the patient—or if the patient is incompetent to judge his or her own best interests, by someone who represents those interests exclusively—explicitly recognizes that the arbiter of the best choice for each patient must be that patient or the patient's representative, not the provider.

Resource allocation decisions often affect service delivery to large populations or subpopulations of patients, not just to specific individuals. Decisions that concern the discontinuation or expansion of special programs such as outreach programs to the homeless, assertive community treatment programs, or psychosocial rehabilitation programs affect many patients and communities. These decisions are best made by teams of relevant stakeholders, including professional experts, consumers, and members of the local community. The Kantian injunction to serve each patient's needs cannot stand by itself, but must be guided and supplemented by a consensus derived from the community of concerned citizens.

Maximization of benefit and minimization of cost.

We live in a world of scarce resources. In addition, there are few universally accepted psychiatric treatment indications and no universally accepted treatment protocols that fully specify necessary and sufficient treatments for each of the many situations encountered in clinical practice.

As a result, the principle that medical decisions should be guided exclusively by each patient's best interests—even if the decisions are based on a consensus of stakeholders as described above—cannot be realized. Resources are inevitably inadequate to finance every treatment that would benefit every patient. Best interests and best treatments are imprecisely specified. Thus blindly providing all possible services to patients selected on a first-come, first-served basis would be neither efficient, as some services would yield limited benefit, nor fair, as this strategy would eventually penalize the person next in line after the last person to receive all the services he or she needed.

As vividly demonstrated over the past 30 years of health care policy debate, we must take resource constraints into consideration in planning how the health care system will work. We are thereby brought to the principle of maximizing cost-effectiveness—of giving priority to programs that maximize the amount of improvement in health-related quality of life per dollar expended.

This cost-effectiveness perspective derives historically from the utilitarianism (17) of Bentham and Mill, a philosophical tradition quite different from that of Kant. In utilitarianism, societal decisions are to be guided by the goal of achieving the greatest good for the greatest number of citizens. Because the health care system cannot meet the needs and desires of every individual, its goal must be to maximize the total health care value for society within the available resources. This goal inevitably entails restricting expenditures for some patients and treatments because more health care benefit can be obtained by expenditures on other treatments for other patients. The Kantian imperative must be modified to accommodate resource limitations and the needs of society as a whole.

The practical problems in maximizing the cost-effectiveness of the health care system are formidable. First, the cost of medical treatments must be determined and their effectiveness measured on a common scale such as Quality-Adjusted Life Years (17). Such a scale must take into consideration not only how many additional years of life each treatment offers, but the quality of life enjoyed during those years.

Next, the cost-effectiveness ratio of each treatment must be computed by determining the dollars spent for each treatment that is required to save a year of life, adjusted for quality of life. Treatments could then be arrayed in order from most to least cost-effective.

Finally, depending on society's total health care budget or its valuation of a year of life, a subset of the most cost-effective treatments could be funded (18) to generate the maximum health care improvement possible for the allocated dollars. Although priority setting of this type is challenging to implement, it has been done on a large scale in Oregon (19) and New Zealand (20), and rougher rule-of-thumb methods have been proposed for use when comprehensive cost-effectiveness data are not available (21).

A crucial distinction in cost-effectiveness analysis must be made between comparison of mutually exclusive treatments, such as conventional versus atypical antipsychotic medication for treatment of schizophrenia, and comparison of nonmutually exclusive treatments, such as treatment of depression with selective serotonin reuptake inhibitors versus assertive community treatment for people with severe mental illness (18). In identifying the better of two treatments for the same type of patient, benefits invariably accrue to those patients, and no loss is experienced by any other group. However, in evaluating treatments for different groups of patients, one of those groups, and society as a whole, may realize increased benefits, but the other group—especially those with more serious illnesses—may experience reductions in care.

Because of the severity and chronicity of the illnesses of severely ill patients, clinical gain among these patients may appear limited, and the cost of their care is often substantial. As a result, the gain in quality of life per dollar spent for treatment of more severe illnesses may be lower than that for less severe illnesses, or at least appear to be so.

Because methods have not been developed for empirically comparing the cost-effectiveness of treatments for different illnesses, it may be incorrect to assume that treatment of less severe illnesses is more cost-effective. In the absence of appropriate data, however, decision makers in many organizations seem to be adversely evaluating the relative cost-effectiveness of psychiatric treatment in general and of intensive psychosocial treatment for severely ill patients in particular.

Thus cost-effectiveness analysis as currently practiced has two important limitations. First, refined methods of comparing the cost-effectiveness of treatments for different mental illnesses have not been developed. Second, even if these methods were to be developed, they might support decisions in which efficiency as measured by total societal benefit per dollar spent would be maximized but in which considerations of equity or fairness, to be discussed in the next section, would be neglected.

Equity.

A major limitation of an approach favoring the greatest good for the greatest number is that it makes no provision for assuring equity or fairness in the distribution of health care resources across individuals. If our only goal is to maximize aggregate benefits in society, it makes no difference which individuals receive those benefits, as long as the total pool of benefits increases.

Thus from the perspective of cost-effectiveness, if one person happens to have an extensive array of illnesses, each of which is amenable to highly cost-effective treatments, that individual might have all his or her health care needs met while someone who has an illness that is less treatable and more expensive to treat might get no services at all. Resource allocation guided exclusively by the goal of maximizing societal welfare is indifferent to the distribution of benefits across individuals; it makes no difference whether all benefits go to a small number of people or are shared equitably by many people (22).

One of the basic principles of a democratic society is that social benefits ought to be distributed fairly and equitably among citizens. America is not a winner-take-all society, but rather one that has been committed from its beginnings to maximizing equality of liberty and opportunity among citizens and to assuring a minimum quality of life below which no citizen should be allowed to fall. It is here that the special obligation to the least well off—those patients with the most severe, disabling, and treatment-resistant conditions, regardless of the specific diagnosis—emerges as important regardless of the relative cost and effectiveness of their treatment.

In his widely respected, although controversial, theory of justice John Rawls (23) argued that just societies should operate on what he called the "maximin" principle. This principle asserts that in a society committed to fairness and equal opportunity, the first obligation, after guaranteeing liberty for all, is to provide a minimum standard of living for the least-well-off citizens.

In a related body of work, the philosopher-economist and 1998 Nobel Prize winner Amartya Sen (22) suggested that what is to be "equalized" in such a society is each person's capability to fulfill his or her potential. In this perspective, which is an important corrective to simple cost-effectiveness analysis, psychosocial and rehabilitative programs for people with the most disabling mental illnesses would receive especially high priority. Thus, in considering how benefits are to be distributed among members of society, attention to equity or fairness is needed to counterbalance the insensitivity of approaches based on cost-effectiveness or welfare maximization alone. Although the simple goal of maximizing cost-effectiveness may tend to displace support for more intensive programs for severely ill patients, principles of fairness suggest that the person with schizophrenia ought to have as much, if not more, opportunity for treatment as the person with less severe problems (23).

Responsibility of consumers.

It has becoming increasingly common in discussions of social policy to suggest that those who benefit from public services incur an obligation to make responsible use of the benefits they receive (24). In this view, the social safety net has become a trampoline requiring effort and commitment by its beneficiaries (25). The imposition of work requirements on welfare recipients is perhaps the most publicized example of this trend, but the growing interest in compliance with psychopharmacological treatment reflects a parallel concern with maximizing the value of treatments and empowering consumers to make the best use of them.

Daniel Callahan (26) has recently extended this line of thinking into the health care realm. With due regard for the risks involved in expecting people to take responsibility for medical events over which they may have little control, Callahan has proposed that wide recognition of the health consequences of personal habits such as smoking, drinking, and overeating suggests it is time to foster a public ethos of individual responsibility for one's health.

Applying this line of reasoning to mental health resource allocation suggests that more systematic attention might be paid to patients' adherence to the treatment regimen for high-cost interventions. If limited numbers of patients have access to intensive case management, atypical antipsychotic medication, or vocational rehabilitation programs, the interests of fairness, efficiency, and effectiveness might be well served by requiring a periodic clinical review to determine whether treatment has been effective and well used. Important issues that would have to be considered include who would conduct such reviews, what standards would be used to define treatment adherence, and how many chances each patient would receive to show acceptable adherence.

Research evaluation of novel treatments is typically based on short-term trials, and the long-term value of these treatments is largely unknown when they are first used. Efforts should be made to develop standards of responsible use of treatment opportunities without imposing unrealistic or punitive expectations for outcomes that are not under anyone's control. Appropriate clinical evaluation of each individual's capacity to contribute responsibly to his or her treatment is also necessary.

Industry innovation and marketing.

In the years since the demise of the "command and control" economy of the Soviet Union, there has been a flush of confidence in the ability of free-market incentives to guide private industry in the production and marketing of goods and services that the public wants (27,28). To some observers, the free market should be the ultimate evaluator of cost-effectiveness. Although it has long been clear that the market is not an optimal tool for organizing the delivery of heath care (29), the idea of managed competition (30) and the reality of managed care in its various forms represent efforts to introduce market-like incentives into the delivery of health care. These incentives drive contemporary health care organizations to limit high-cost services for people with severe mental illness, sometimes in ways that conflict with the values of individual autonomy or societal equity.

However, private markets also affect the delivery of mental health services and the allocation of treatment resources by creating financial incentives for technological innovation, especially in pharmaceuticals. The past decade has seen the introduction of new medications for depression, bipolar disorder, and schizophrenia that have well-demonstrated clinical advantages but that also cost up to ten times as much as conventional treatments.

Market incentives thus affect resource allocation processes in two ways. On the one hand, they have stimulated the development of improved treatments that would not have been available otherwise. It now costs approximately $200 million and takes ten years to bring a new medication to market (29). In the absence of the profit motive, it is unlikely that such gargantuan development efforts would be undertaken. On the other hand, market incentives expose decision makers to hard-selling advertising campaigns that are designed to sell as much product as possible, often at the highest price the market will bear.

Although research studies funded by both private industry and the government provide much objective data on the value of new treatments, few program managers have the mastery of the subtleties of the scientific literature that is needed to judge the effectiveness of the treatments. In addition, general professional opinion may be biased by the hope for better treatments. Private industry thus plays a pivotal role in generating demand for improved treatments. Clinicians and program managers must weigh the value of new treatments carefully and independently against existing treatments—often psychosocial or rehabilitative services—that are not proprietary and therefore receive far less commercial promotion.

Technical quality and skills of the local workforce.

Finally, although in principle resources should be allocated rationally, in ways consistent with both the values presented here and with scientific findings, the portfolio of programs provided by any agency or in any geographic area must be influenced by staff skills and local population needs. A mental health center in which the staff has extensive training and experience in assertive community treatment is likely to serve its clients best by providing a different array of services than one with a strong local network of psychosocial clubs. Likewise, a community with a large population of monolingual immigrants will need a different array of programs than will a community located near a state hospital in the upper Midwest. Program managers must adapt broad principles like those presented here with due respect for local circumstances.

From principles to practice

The seven principles outlined here are not hierarchically ordered, and as a result their joint application will not necessarily yield a consistent, ordered determination of program priorities. Decisions based on patient autonomy and preference, for example, are likely to conflict with decisions based on cost-effectiveness, and decisions based on cost-effectiveness may conflict with concerns for equity, local population needs, or provider skills.

Perhaps the most useful application of these principles is as a frame of reference for understanding why priorities may legitimately vary and conflict. Providing a formula for evaluating which programs should be given the greatest priority is attractive, but it is not realistic. These principles will not make decision making easy. However, they may find wide applicability in private group practices that must decide whether to offer rehabilitative services for severely ill patients or in public-sector mental health centers that are considering whether to invest in assertive community treatment teams or outreach programs to reach homeless mentally ill persons. By mapping the terrain of our conundrums, we may be able to do a better job of understanding our options and of making balanced and broad-based decisions.

Dr. Rosenheck is director of the Department of Veterans Affairs Northeast Program Evaluation Center (182), 950 Campbell Avenue, West Haven, Connecticut 06516 (e-mail, [email protected]). He is also clinical professor in the departments of psychiatry and public health at Yale University School of Medicine. This paper is one of a series of three papers in this issue on care of the least well off in mental health services.