TIP 14 — TIPs — Documents — Home

Chapter 3—Ethical Issues and
Outcomes-Based Accountability

The purpose of this chapter is to ensure that a discussion of ethics and the values choices raised in alcohol and other drug (AOD) treatment policy are part of the framework for monitoring the outcomes of treatment. This chapter will not resolve the ethical dilemmas that may occur when monitoring outcomes, but it will attempt to identify and describe some of these dilemmas.

Policymakers and practitioners in the field must be able to discuss ethical choices openly rather than burying them beneath more technical activities or ignoring them altogether. Values matter, and clarifying what values are under discussion can help make policy decisions easier to understand. Policy can be defined as the authoritative allocation of values, and it is important to be clear which values are under review.

Ethics as a field of study is useful in assessing both the rightness of decisions and the fairness of the decisionmaking process. Ethical guidelines can help order values when values are in conflict. The use of ethical concepts does not dictate final decisions, but can help substantially in clarifying what criteria are being used and how they are weighted in comparison with each other. A classic ethical issue, for example, occurs in the field of substance abuse treatment when there is a need to weigh the individual's freedom to choose against the rights of others (such as the individual's children or the larger society) not to be harmed by choices made by the individual in the role of parent or citizen.

Despite the frequent use of the phrase "the ends don't justify the means," ethical choices often involve exactly this kind of weighing of different ends and judging whether or not coercive means are justified by the probable results. Classical and recent ethical analysis has focused upon the neutrality needed for an observer to make such judgments. John Rawls's landmark work A Theory of Justice shows the importance of distributing goods to benefit the least advantaged in society, based upon a neutral observer's decision about which individuals fall into that group (Callahan and Jennings, 1983; Rawls, 1971; Rhodes, 1986).

Other ethical issues in human services include the point of view of the patient compared with that of the treatment staff and the funder; the competence of patients to make decisions about their own well-being; and the power of those who exercise coercive controls, and how doing so affects their perspective (Callahan and Jennings, 1983; Rawls, 1971; Rhodes, 1986).

Ethical choices involve the following:

• The rights and responsibilities of the patient, including the expectations society places upon those individuals whose behavior it is attempting to alter
• The rights and responsibilities of the treatment system and those who work in it
• The validity of claims made by funders and other stakeholders on both the treatment system and on patients, judged by some clear, specified standard of what is fair.

The last few years have brought a renewed interest in using outcomes measures to increase the accountability of human services programs. New Federal and State legislation has elevated the visibility of outcomes by linking funding regulations to outcomes achieved. But these are not merely issues of good government, program design, evaluation science, or technical measurement. These issues of outcomes measurement have deep ethical content.

The first set of outcomes-related issues concern who is acting and who is being acted upon in service delivery. There is a distinction between the person as subject and the person as object. The person as subject acts; the person as object is acted upon. One confers dignity when treating another as a subject, but to treat another as an object is demeaning. This distinction between subject and object argues powerfully for what some have called "empowerment" and expanded patient involvement. But as important as these concepts are, they result in changes that may be restricted to the processes of programs rather than to their content. Central to outcomes monitoring is the idea of patients as subjects. Such a focus argues that the content of AOD abuse treatment services should be measured by the criterion of what services do to improve the well-being of patients, with a primary assumption that those receiving help are themselves active participants in the helping.

If an agency accounts only for the costs and staff time involved in services, the patient tends to be treated as a case. However, if greater attention is paid to the changes in the condition of the person (the patient) as a result of services provided (the outcomes), then a fundamentally important shift has occurred. It is a shift from thinking about a patient as being acted upon by an outside agent to thinking of the patient as being an individual who is changed in part because of that patient's own actions. This shift has direct ethical content.

Beyond the renewed emphasis upon the needs of the patient in outcomes-based accountability, a second difficult set of ethical issues arises in determining what is fair to measure in holding an agency, a program, or even a staff person accountable for achieving certain specified outcome standards in treatment. The concept of fair measures has been used to underscore the importance of the patient's and staff person's perception that what is being measured is an appropriate indicator of what progress has (or has not) been made.

Fairness must have both process and substantive meaning. Procedurally, the staff or patients whose behavior or performance is being monitored need to feel they had a role in setting the outcomes, rather than feeling that the outcomes were forced upon them. To be sure, staff have to take a mandated standard seriously, but outcomes-based accountability clearly works best if the behavior being monitored is voluntary and the standard is seen as fair.

Substantively, fairness must be gauged in terms of external effects that may make the outcomes achievable or realistic. If unrealistic measures are set, staff will not take the outcomes seriously because they know they will not be able to ensure that patients achieve them. These issues of external impact will be discussed at greater length below.

Several other ethical issues arise in addressing outcomes:

• Who has the authority and legitimacy to define outcomes or outcomes standards? How do different perspectives of different stakeholders affect the definition of the goals and needs that underlie selection of outcomes?
• Who negotiates which outcomes indicators will be used as measures of performance?
• Who decides if the measures are fair?
• Who monitors to determine if the measures are achieved, and from what set of motives? How independent must that monitoring be from having an interest in the agency's or the program's survival?
• How are decisions made to reflect outcomes in funding determinations? Can outcomes be used to shift funding from programs with poorer outcomes to those with better outcomes? What are the ethical effects of closing down programs or shifting funding? What are the ethical effects of ignoring programs' effectiveness and not shifting funding to programs with better outcomes?
• How are external effects on outcomes taken into account?
• What is the legitimate role of treatment providers and the patients who are the intended beneficiaries? How much of a contribution is it fair to ask treatment staff to make? What responsibilities do patients have for improving their own well-being?
• What are the rights and responsibilities of funders, including taxpayers, in determining the outcomes by which to judge whether or not programs are successful?
• How do we assure that outcomes affecting people are given at least equal weight to those affecting dollars; in other words, how can patient outcomes be given standing in the face of the great importance given to fiscal outcomes
• How are the different groups affected by outcomes best taken into account in the process of defining and monitoring outcomes?

These are questions about power, fairness, and the willingness to make large changes in current ways of doing business. Each of these questions raises further questions of equity and effectiveness, and the answers to them are affected by the different perspectives from which each stakeholder judges equity and effectiveness.

Unfortunately, outcomes monitoring is often approached as though it were a far more technical process than it is. We look to researchers to develop the "clean data" needed for an objective assessment of whether programs are succeeding in their precisely described missions. Outcomes monitoring would probably be easier if it were as technical, impersonal, and dehumanized as it is often assumed to be—but reality does not present itself quite so neatly.

Outcomes monitoring is a way of setting and following up on goals, and the process of goal setting cannot be conducted as a technical exercise. Values choices come into play repeatedly in community or agency goal setting, and it is impossible not to take these values issues into account. To assign priority to one value over another, when both may involve the lives of children, families, or patients needing services, is to make choices among different goods. Sometimes these are very difficult choices in which each choice involves strongly held values that are in conflict. No matrix, gaps analysis, or needs assessment will make clear which values should be given priority in any given system or community.

So the first ethical task in working with outcomes is to be clear that choices about priorities involve ethics and not to try to obscure that reality with technical or analytical verbiage.

Very difficult questions arise in human services systems in making the set of choices that must be made among many different categories of patients who need various kinds of help. Sometimes the human services system seems to be trying hard to spread its resources widely across many different groups in need, rather than concentrating resources in ways that might result in denying any one group access to services. This allocation may lead to resources that are provided in "dosages" far too small to be effective. Providing many patient groups with token-level, symbolic allocations of resources may be politically effective—but it is not often programmatically effective.

A thorough approach to outcomes monitoring, as argued elsewhere (Young et al., 1993), requires decisions about priorities. Often, human services systems allocate resources by earmarking a portion of funding for high-priority groups, based upon a national or local sense of the crisis affecting those groups. Thus, some drug treatment constituency groups have demanded a set-aside of funds for pregnant users, HIV-positive patients, and other patients with certain conditions deemed more urgent than others.

There are ethical consequences of setting priorities when there are more patients than available resources can help. The first ethical issue is identifying the values underlying the establishment of priorities and assessing the results of those choices. If political lobbying by the advocates for one group is successful, does that success come at the expense of other groups with needs which are arguably as great or greater? Advocates have every right to seek the best possible treatment for "their" patients, but if there is no standard other than what works best politically, power becomes the only criterion for decisionmaking.

Another set of ethical concerns arises in the question of the roles patients themselves have in setting priorities and in designing programs that flow from those priorities. If patients do not have a role in the debates over resources, a critic may be justified in labeling such a system agency-centered rather than patient-centered. Though the focus on outcomes may force a greater emphasis upon patients, there is no guarantee that, simply because large amounts of data are being compiled about patients, patients themselves will contribute to making the decisions that flow from the collection of that data.

Ethical concerns cannot answer these questions—but they can help in posing them as a part of the decisionmaking process about how scarce resources are allocated.

The process of selecting outcomes as goals to guide agencies' work requires making some important tradeoffs that have ethical content. One of the most important of these is the tradeoff between cost effectiveness and treatment effectiveness. The former is funder driven, while the latter is patient driven. Both funders and patients have rights (and responsibilities), but often the cost-effectiveness issues related to fiscal outcomes take precedence over patient outcomes.

In the long run, this emphasis may lead to choices that are not only ethically weak, but also ultimately fiscally and programmatically weak. For example, if a short-run savings is achieved by selecting patients who may need less help than more difficult patients, two results may follow that can jeopardize both fiscal and patient outcomes. First, the fiscal outcomes may eventually be worse because the harder-to-serve, chemically dependent patient who is left untreated will continue to require aid from a number of systems (for example, criminal justice or social services) at much higher short- and long-range costs. Second, the patient effectiveness of such short-range choices may also be limited, since treatment will be restricted to those methods that work only for the "easier" patients. As a result, there will be little pressure on agencies to improve their performance to help the patients who need help most.

This issue of "creaming"—the effect of performance-based contracting in creating incentives for agencies to select the patients easiest to serve—has been treated extensively in the literature on outcomes. If patients with greater needs receive lower priority in an outcomes-based system, the accountability this instills will be destructive of public policy goals in some important respects. Ethically, the use of outcomes requires some consideration of the incentives created by outcomes-based funding and the impact of those incentives on different patients.

In a somewhat different context, discriminating against patients who may have the same claims to public services but differ in their severity of abuse may be justified based on some screening of patients' willingness to work on their own behalf. Typically, however, it is the agency's assessment of the patient that determines the priority given, rather than a fair test of the patient's willingness to assist in treatment.

The use of outcomes measures as a basis for funding raises another question about fairness: How fair is it to base all or part of an agency's funding upon its performance? The question, in part, concerns external factors that may have a great deal to do with whether the agency can achieve its goals. In the AOD treatment field, an example would be expecting a program to succeed where length of stay was limited to 1 week. Therefore, it is important to evaluate outcomes measures against the capacity of the agency to achieve certain outcomes in the context of local setting and conditions.

Capacity also involves the self-sufficiency of the agency. If funds for a treatment program serving teenaged parents are linked in part to its success in reenrolling teen parents in school, but the school system has restrictive policies that make it very difficult for teen parents to attend, the treatment provider may be penalized because of policies it does not control. Similarly, if the treatment provider needs childcare and transportation resources to achieve its programmatic goals but controls neither, it is neither sensible nor fair to penalize the treatment provider for not achieving results that are beyond its control.

The process of assessing whether the measures are fair can rely in part upon the ethical concept of the "neutral observer," but it also has to take into account the subject-object distinction explained earlier. A disinterested observer can argue that a 75 percent recovery rate, measured in an objective, verifiable way, is "fair." But from the points of view of the staff whose performance is being judged—and the patients whose lives are even more fully involved—the fairness of the measures may be perceived differently. Both the disinterest and determination of patients and staff who will be affected by outcomes have ethical meaning, and both need to be considered in assessing whether outcomes are fair measures.

The ethics of the effectiveness of treatment raises additional issues. If programs are ineffective and yet scarce resources are spent on them, what are the obligations of systems to redirect resources toward programs that really help people?

Are results being taken seriously if it is known that a treatment program or method does not achieve the outcomes it is being funded to achieve—and nothing is being done to reallocate resources away from such an ineffective program or method? More specifically, if patients are on waiting lists for programs that have been proven effective and funds are being allocated to programs that have been proved ineffective, is there a moral responsibility to seek to reallocate funds toward the programs known to be effective?

If outcomes monitoring is intended to improve resource allocation and accountability for effective use of funds, it is necessary to make the hard choices that reallocation demands. While funded agencies and programs may argue against reallocation proposals, their arguments must be weighed against other uses of the funds which may shift resources to programs of more proven value to patients. Whether a system is ultimately patient centered or agency centered is revealed in part by whether funds are ever redirected to programs that have proven more effective in serving patients, as contrasted with meeting the needs of funded agencies operating programs that have not proven their effectiveness.

Merely setting up a new management information system to monitor outcomes will not ensure that appropriate resource reallocation occurs. Human services agencies have collected massive amounts of data for decades, through both manual and electronic means, without causing any appreciable impact on shifting resources from ineffective programs to effective programs. A sense of urgency—and a healthy skepticism about current claims of effectiveness—are required to undertake the difficult, often controversial task of redirecting funds. Ethical content arises in comparing the claims of currently funded agencies and programs against those that are underfunded or unfunded. But once knowledge of ineffectiveness is in the hands of decisionmakers, issues of the misallocation of resources become directly policy relevant. If planners and agency leaders are not aware of this dimension of information system building—or do not want to involve themselves in the tangled thickets of reallocation—the use of outcomes data will be limited to operational adjustments that will often end up as little more than tinkering at the margins of ineffective programs. In times of scarce resources, such tinkering is at best ethically questionable and may represent serious misallocation of vitally needed resources for patients.

Information in Chapter 8 about turning the empirical results of outcomes monitoring into policy, including fiscal policy changes, should be considered in the context of this discussion of ethics.

Additional hotly debated ethical issues in the field of substance abuse treatment go to the heart of public and personal perceptions of addiction and recovery. Many individuals' beliefs about addiction have grown out of personal experience and observations, and ethical judgments are likely to follow from these perceptions and biases. A central issue is the responsibility of the addicted person for his or her own life and whether or not the causes of the disease of addiction is more the fault of the patient than any other disease might be. Other ethical issues arise around whether or not society is ethically entitled to insist that the addicted person seek treatment and whether or not society is ethically required to provide a treatment opportunity for every addicted person who seeks to go into rehabilitation and recovery. Another ethical issue is the perceived difference between legal and illegal substances, which can lead to very different value judgments about patients.

Further ethical issues in treatment arise from the different resources provided for different patients, based on personal ability to pay for services or differing reimbursement systems. Thus, addicted celebrities and star athletes have access to the best treatment, while an unemployed individual with the same condition is treated very differently and may even have to meet eligibility criteria to qualify for treatment. Waiting lists for some programs require callbacks to remain on the list; in contrast, some private hospitals will send cabs to make it easier for people to enroll. This disparity may result in very different outcomes, and ignoring these issues will obscure important problems in AOD treatment access.

Other issues specific to the AOD system include the fairness of the expectations society places upon patients in defining when they are "well." Should the expectation (and thus the measured outcomes) be that no more AOD abuse will take place; that less use will take place; that abstinence will follow treatment; and that society will benefit as a result of less crime, lower health costs, and increased worker productivity? What is fair to expect? Different systems and agencies draw these lines in different places. An employee assistance program (EAP) cares about workers' performance and productivity, a treatment program for homeless people wants to ensure stable housing, and the child welfare system wants to know if parents are ready to be responsible for their children. This measurement of effectiveness of treatment in fundamentally different ways results in a loss of "horizontal equity"—the social policy criterion that people in similar situations should be treated the same.

What is apparent is the enduring power of the categorical funding system to reduce the patient to "pieces of a person"—to whatever eligible definition of need will trigger funding, rather than considering the patient as a whole person in need of combined services across systems. Some portion of the AOD outcomes issue, then, demands a point of view that incorporates integrated services; the categorical system should not be allowed to fragment treatment into whatever the treating agency defines as its mission—rather than what patients (and their children or families) want and need.

There is no ethical calculus that can easily be applied to determine whether a decision is ethically valid. Rather, there are ethical checkpoints that may be used to frame the issues for a thoughtful analysis of the moral dimension of policy choices, which can be added to the social, economic, cultural, and political dimensions. Some of these checkpoints are:

• Is this choice fair to those involved? How would they respond to that question? How would a neutral observer respond?
• Has the impact of this choice on all parties been weighed carefully? These parties include providers, patients, and the community.
• Is there reciprocity in the responsibility for the choice; that is, have both those providing help and those receiving help accepted obligations to make it effective, according to what they are able to do?
• Have those with the least power to raise issues and implement policy around those issues been given help in that task by those with more power?
• Are decisionmakers driven by motives that are narrow and largely self-interested?
• What is known about the results of previous policy choices similar to this one, and how did those choices affect their intended beneficiaries?
• Is there evidence that decisionmakers have or sought to have empathy with those most affected by their decisions?
• Have the values underlying policies been identified and debated openly as values issues, or are they concealed beneath the economic, political, or technical issues?
• Has an attempt been made to identify the interests of the whole community—the values that bind it together—as well as the more special interests that may seem to conflict with the policy choice?
• Has there been an attempt to understand the full context of the family setting in which treatment is being provided? Have the needs and strengths of children, parents, and other family members been added to the assessment of what the patient needs?
• Is there an institutionalized, independent channel through which patients can regularly evaluate the programs that serve them?

1.This chapter was written for the consensus panel by Sidney L. Gardner, M.P.A., M.A.

Last Updated 11-7-02