Setting Priorities for Health Technologies Assessment: A Model Process (1992)

This chapter describes general principles that underlie the development of any priority-setting process and the implications of those principles for priority setting in the Office of Health Technology Assessment (OHTA).

A process is useful to the extent that it addresses four issues:

• The process should be consistent with the mission of the organization that is to use it. A process that does not incorporate the basic value system of the user cannot help an organization set priorities according to its values.

• The results of the process should be consistent with the needs of the user and should provide information in the form that is most useful. For instance, clinician users of technology assessments seek comparative information based on a given clinical condition to help them in decision making. This issue is discussed further in a later section of this chapter.

• The process should be efficient, especially in instances in which it must share resources with technology assessment itself.

• The process should be capable of operating in the real world of the organization. If the information it produces is to be used effectively, the process must consider not only what information is needed but also the political, economic, and social constraints that will affect how the information can be used.

These principles apply to any priority-setting process. The next section

considers how they should guide the development of a priority-setting process for OHTA in particular.

OHTA's priority-setting process must ensure that the priority rankings it produces are consistent with the agency's objectives. What are those objectives?

The goals and objectives of OHTA are those of a public agency charged with producing information about a medical technology. The information should support the public interest, and OHTA's process should provide the information as efficiently and effectively as possible.

Although specifying the public interest is not an appropriate task for this committee, there is little question that society has expectations of its health care system. An understanding of these expectations is relevant to the work of OHTA and should be incorporated in its proposed model of priority setting. These expectations are related to beliefs about what health care is to achieve and how the health care system is to achieve it—with beneficence, nonmaleficence, and fairness (the goal of distributive justice in the allocation of all resources).¹

Four elements of the public interest deserve consideration in determining which set of technologies should take precedence when assessment resources are limited:

• the extent to which health care services can reduce pain, suffering, and premature death; increase health, functional capacity, and life expectancy; or maintain the functioning of those who are permanently impaired;

• the extent to which expenditures for health care services that are ineffective or needlessly costly can be reduced or eliminated;

• the extent to which inequities in access to effective health care services or maldistribution across equally needy populations can be reduced; and

• the extent to which other special social issues can be informed by assessment.

The implications of these elements with respect to a process for priority setting are considered below.

A primary objective of health care is alleviating and preventing the pain and suffering that are part of illness and preventing premature death. Technology assessment—as a primary source of information about the extent to which health care services can effectively achieve these ends—plays a critical role in supporting this primary objective. It should be a priority to the extent that it can lead to the delivery of care that accomplishes these ends.

At a population level, these goals can be viewed as related to the current aggregate burden of illness (the number of people with the condition multiplied by the burden of illness), which is also a measure of the potential for improvement—the medical gain that would follow a change in practice that might follow an assessment. But technology assessment will not necessarily lead to a change in practice solely as a result of providing information on a technology's effectiveness. Such information is only one factor in determining how a technology is used; third-party reimbursement, the practice environment, and legal concerns also influence practice. Because, in the short run, technology assessment affects only the information base of health care decision making, evaluating a technology should not have high priority if increasing information is unlikely to lead to a change in practice. In other words, when any change in practice is unlikely to occur, resources for evaluation should be directed elsewhere.

Although reducing pain, suffering, and premature death is a primary objective of health care, it cannot be considered or accomplished outside of the context of public concern about the magnitude of current health care expenditures and the rate at which they are increasing. This concern suggests that the public would be additionally served to the extent that a technology assessment leads to appropriate reductions in the cost of health care services. Such reductions could follow when an assessment shows that certain health care services are truly ineffective or that competing technologies are potentially substitutable (with no important difference in health gains) at lower costs.

Expenditures for a health care service ought to affect its priority for technology assessment. The important factor for priority setting, however, is not the dollars spent but the potential for more appropriate expenditures. Any reduction in costs depends on the likelihood that cost-saving changes in practice will follow an assessment.

A reasonable goal of any health care system is to deliver the best possible health care to all citizens, regardless of their social, political, or financial condition. This goal is served whenever information is produced that leads to greater equity of health care delivery, especially in terms of the distribution of health care services to those who are underserved, or to more information about a problem that, because it affects a very small population, would not otherwise be the subject of investigation. The priority to be accorded an assessment depends not only on the magnitude of inequity but also on the sensitivity of that inequity to better information. Where there is little capacity to change practice through information, the problem of inequitable distribution of medical services or lack of information about technologies used in the care of a sparsely studied condition will be little affected by technology assessment.

Although the immediate user of the priority-setting process will be OHTA, the ultimate users are those whose decisions will be affected by an assessment. Thus, the committee considered the characteristics that would cause a priority-setting process to produce helpful information for those who use it.

The committee approached this issue by trying first to identify the users of OHTA's technology assessments and to understand how they use the information generated by those assessments. First, clinicians use comparative information about a technology; users are almost always interested in comparing the characteristics of one technology with another. Second, users are generally interested in the Characteristics of a technology with respect to some specific clinical condition. In some circumstances, users may be interested in assessment to help in deciding about the acquisition of an expensive technology that has the potential for use in a wide variety of conditions (e.g., an imaging technology, a multiphasic blood analyzer). Generally, however, users do not need information, for example, about positron emission tomography (PET) in isolation from the condition or conditions for which it is used. Rather, some users will need to know what information PET can provide for a patient with neurologic

disease; others will need to know what information PET can provide that will help a physician evaluate a patient with cardiovascular disease.

The committee recognized that questions concerning coverage for specific technologies drive most of OHTA's current assessments and those of other payers. Nevertheless, the committee believes that OHTA needs to reformulate a question such as, "Should PET scanning be reimbursed by HCFA?" to one that is more useful to clinicians (and ultimately third-party payers)—for example, "What is the optimal management or care for the patient with new-onset angina?" This is the kind of question asked by clinicians, and an assessment will have the greatest impact if it can supply the answers that clinicians seek. If technology assessment efforts are to strive to produce patient-specific recommendations, each candidate for assessment must be specified precisely enough for the assessment to serve a clinician's needs.

Thus, the answer to the above question might eventually be tied to a specific patient population—for example, "PET scanning is the most cost-effective diagnostic test to perform for a 68-year-old male, type-II diabetic patient with new onset of exercise-induced substernal burning." Specification at the level of individual patient characteristics may seem only a distant goal of technology assessment, but the health care system must attempt to achieve it if the assessment and its products are to be useful to the clinician and his or her patients (e.g., McNeil and Abrams, 1986). Such detail was unthinkable 20 years ago; it is now, however, possible to state, for instance, that every patient with a head injury who has temporarily lost consciousness does not need a PET scan, magnetic resonance imaging, a CT (computed tomography) scan, a brain scan, and a lumbar puncture. Instead, a protocol derived from decision analysis can specify the most cost-effective diagnostic test or sequence of tests for a given patient.

Yet great care must be taken, in developing clinical practice guidelines with this level of specificity, to ensure that innovation will continue. There is often a learning curve with new technologies, and because data on such technologies are frequently limited, early assessments may provide incorrect or misleading conclusions. Further, data for good evaluations depend on some diffusion of the technology prior to assessment. Any clinical practice guideline needs to include an explicit statement about the quality of supporting evidence and to make allowance for clinician latitude in the face of poor evidence (Eddy, 1989, 1990a,b,c; IOM, 1990c).

Efficiency requires that the priority-setting process accomplish three objectives. First, it must ensure that important issues are addressed. (If an

important issue related to the use of a particular technology is never recognized, then that technology may never come to the attention of the priority-setting process.) Second, the process must ensure that relatively unimportant issues are excluded as quickly and as inexpensively as possible. The first objective implies a process that is open enough to minimize the risk of excluding important issues. On the other hand, openness implies that many candidate topics will turn out to have relatively low priorities. An efficient process will eliminate low-priority issues at an early stage of evaluation before substantial resources have been invested in their evaluation. Thus, the process must include a method to reduce the number of topics if both openness and efficiency are to be achieved.

The third objective is to minimize the cost of data collection. A process for priority setting that requires a large amount of highly detailed information is not reasonable, especially given that a high level of precision in setting priorities is probably not necessary. Given OHTA's access to public data sets and to content experts, a process that uses available data and supplements it with expert opinion will be more cost-effective than a process that requires primary data collection. The eventual implementation of a computer-based patient record might allow much more accurate data gathering than is currently possible at a feasible cost (IOM, 1991a).

A priority-setting process must be acceptable to those whose decisions are to be influenced by it. In the long run, the acceptability of the priority-setting process will depend largely on the validity of the priority-ranked list of conditions and technologies. The design of the process, however, should include elements that will make the process acceptable and credible. First, it must be understandable; people will mistrust any "black box" process. Second, the logic of the process must be open to inspection, and the logic must be clearly and reasonably articulated. Third, the process must be defensible. Because of the competing demands for technology assessment resources, those who assign priorities must be able to justify the process. Finally, the process must be, and must appear to be, objective and fair. If it appears to be sensitive to the influence of special interests, the product of the process will have no credibility and therefore no power. The process should, therefore, be open to input from a broad array of constituencies. Openness and broad input are the most effective means to ensure objective, fair priority setting.

The committee formulated several general principles to direct its development of a priority-setting process. The first such principle is that a priority-setting process should be consistent with the mission of the organization that uses it. For a public agency, the values of the public that the agency serves need to be incorporated into the process. For OHTA, such a process would require the assembling of information about the potential to improve health outcomes, to reduce inappropriate expenditures, to redress inequity among those receiving health care, and to inform special social issues.

Second, the priority-setting process must consider the information needs of users. The process designed for OHTA should, in general, focus on technology assessment for specific clinical conditions and for alternative approaches to those clinical conditions.

Third, the priority-setting process must be efficient so that scarce resources for technology assessment are not needlessly consumed in the process of setting assessment priorities. OHTA should seek broad input at the outset but also have some relatively simple mechanism to reduce a large set of candidate topics to a smaller one. The process should also take advantage of available data, or, where data are lacking, of subjective judgments, rather than require the collection of new data.

Finally, the priority-setting process must be capable of motivating decision makers in a politically complex environment; it must be—and must appear to be—objective, open, and fair; it must also invite input from a variety of interested parties and present the logic of the process clearly and carefully to others. The chapter that follows presents a process that the committee hopes can be understood as logically deriving from consideration of these issues.