Printer Friendly

The Oregon priority-setting exercise: quality of life and public policy.

In 1989 the Oregon State legislature passed the Oregon Basic Health Services Act, which created a Health Services Commission charged with "developing a priority list of health services, ranging from the most important to the least important for the entire population to be served." [1] The goal of this legislation was to permit the expansion of Medicaid to 100 percent of all Oregonians living in poverty by covering only services deemed to be of sufficient importance or priority.

The Oregon Health Services Commission (OHSC) initially interpreted "for the entire population to be served" as suggesting the use of cost-effectiveness principles for developing the priority list. These principles are based on the utilitarian quest for "the greatest good for the greatest number" and tend to devalue adverse effects of a policy on specific individuals. [2] By the lights of cost-effectiveness, the "importance" of a health service depends not only on the expected outcomes of treatment (such as prolongation of life, reduction of pain), but also on the cost of that service and on the number of patients who can benefit from it. Thus, even very beneficial treatments might not be considered important if the costs of providing those treatments are high or if only a few people benefit from them.

In keeping with their interpretation of the statute, the OHSC initially conducted a cost-effectiveness analysis of over 1,600 health services ranging from appendectomies to treatment of colds and flu. Predictably, the resulting draft list rated outpatient office visits for minor problems as the "most important" services; the cost of these visits was estimated at $98.51. Indeed, the first 94 items on Oregon's initial list were for office visits, for often self-limiting conditions such as thumb-sucking and low back pain. By contrast, certain life-saving surgeries, such as appendectomies, were rated relatively low because of their higher associated costs.

This counterintuitive priority order (and negative public reaction to it [3]) led the OHSC to abandon cost-effectiveness analysis for purposes of developing its final priority list. [4] Instead, the OHSC developed a set of seventeen health service "categories," which described either a specific type of service (for example, maternity care, preventive services) or, more generically, the expected outcomes of care (for example, "treatment of life-threatening illness where treatment restores life-expectancy and return to previous health"). Commissioners formally ranked these seventeen categories in order of importance according to three subjective criteria: value to the individual, value to society, and whether the category seemed "necessary."

Each treatment was then assigned to the single most appropriate category, based on Commissioners' judgment. Services were ranked within categories according to the degree of benefit expected from treatment. Finally, the OHSC rearranged apparently misplaced services "by hand," for example, moving obviously important services rated low by the method higher on the final list.

This alternate methodology produced a much more intuitively sensible final priority list than the earlier draft list, although more work may be needed before the "final" list can serve as the basis for public policy, particularly with respect to better specifying treatments and indications for treatment. [5]

At the time of this writing, independent actuaries are estimating the costs of providing services on the final list. The Oregon legislature will then decide whether to accept the list as the basis for expanding the State's Medicaid program, as per the Oregon Basic Health Services Act. If so, the legislature will draw a line somewhere on the list to separate the services that will be covered under Medicaid from those that will not. Finally, if this step is taken, Oregon will appeal to the federal government for a Medicaid waiver, which must be granted if the plan is to proceed.

Role of Quality of Life

The Oregon priority-setting process is significant in many ways, particularly with respect to its implications for social policy. Some of the most important of these implications, discussed in the accompanying article by Charles Dougherty, relate to questions of distributive justice, including Rawlsian attempts to identify the least advantaged members of society and to assess how the Oregon process affects them. (For my money, the uninsured poor are the worst off and the Oregon process improves their lot.) In addition, the fact that cost-effectiveness analysis failed to produce a reasonable priority list has significant implications for future efforts to set health care priorities. [6]

Another important story concerning the Oregon priority-setting exercise remains to be told, however. This story concerns the critical role played by quality of life judgments in constructing Oregon's final priority list. Estimates of how treatments affect quality of life were by far the single most important factor in determining the priority order on that list. Most of the service categories that constituted the principal method of prioritization were explicitly defined in terms of quality of life or, in what was generally treated as an equivalent term, "health status." Furthermore, the secondary (within-category) rank-ordering was performed by reference to the "net benefit" from treatment, which, as well shall see, was an explicit numerical estimate of the impacts of treatment on quality of life.

Lack of understanding about the role of quality of life in formulating the final priority list has already led to erroneous interpretations of the list. For example, a spokesperson for Children's Defense Fund in Washington, D.C., criticized the fact that treatment for extremely premature infants (less than 500 grams and less than 23 weeks gestation) was rated next to last on the list (just prior to treatment for infants born without a brain), saying, "If you're looking at it from a purely economic view, it makes sense not to cover those infants. Of course, we think it's completely unethical to do that." [7]

In fact, however, economic considerations had little or nothing to do with placement of this (or any other) treatment on the final list; rather, like most items near the bottom of the list, treatment of extremely premature infants was rated low because it had been assigned to the lowest-ranked service category: treatments offering "minimal or no improvement in QWB," or Quality of Well-Being, the OHSC's term for quality of life. It was a consideration of the outcomes (in this case, severe retardation and cerebral palsy) of treating extremely premature infants that led the OHSC to make this category assignment--and in turn determined placement on the final list.

Similarly, active medical or surgical treatments for terminally ill patients were rated near the bottom of the list by virtue of having been assigned to the same poor-outcome category just described. Terminally ill patients were defined (problematically, perhaps) as those with less than a 10 percent chance of surviving five years, even with treatment, and included patients suffering from "cancer with distant metastases" or "terminal HIV disease." Comfort care for these patients, including hospice programs and pain medication, was ranked relatively high, however--at 164 out of 709 total items--as was the longevity- and quality-of-life-enhancing drug AZT for patients with HIV disease, at 158.


Before describing the method used by the OHSC to obtain explicit estimates of quality-of-life outcomes (the basis for within-category ranking of services) a little background is required, both historical and philosophical. The focus on quality of life as a principal factor in health care resource allocation has a long history in Oregon, most of which concerns the activity of a community grassroots bioethics project known as Oregon Health Decisions (OHD). [8] For several years preceding creation of the OHSC, OHD had held hundreds of citizen meetings around the state to discuss health care and resource allocation issues. In September 1988 a Citizens Health Care Parliament was held in Portland in which fifty delegates met for a day and a half to develop "a set of public policy principles which are intended to be guideposts for the state legislature and other policy-makers concerned with health care resource allocation." [9] The principles developed by the parliament focused on quality of life to a remarkable extent. Indeed, the first six (of fifteen) principles contained explicit references to the importance of quality of life in making health care allocation decisions. Of particular interest are these:

1. The responsibility of government in providing health care resources is to improve the overall quality of life of people by acting within the limits of available financial and other resources.

4. Health care activities should be undertaken to increase the length of life and/or the health-related quality of life during one's life span.

5. Quality of life should be used as one of the ethical standards when allocating health care expenditures with insurance or government funds.

6. Health-related quality of life includes physical, mental, social, cognitive, and self-care functions, as well as a perception of pain and sense of well-being.

As part of the process of developing its priority list, the OHSC commissioned OHD to hold a series of public meetings to discuss people's values concerning the outcomes of care. OHD used a set of service categories similar to those ultimately adopted by the OHSC to elicit relative preferences for different treatments. Quality of life again emerged as a major priority.

Ethical Considerations

One final, critical clarification is required before describing the OHSC's approach to measuring quality of life. The use of quality of life information to develop public policy is potentially problematic from a couple of perspectives. First, the use of purely objective measures of quality of life, such as the degree of assistance required to walk or level of independence in self-care activities, does not correlate well with perceived quality of life. For example, Najman and Levine reviewed an extensive literature in which quality of life reports were obtained from patients who were "objectively" living restricted lives. Almost invariably, perceived quality of life was higher than might have been predicted. [10]

Second, judging others' quality of life may place us on a slippery slope. In the Encyclopedia of Bioethics, Reich notes that judgment of "unacceptable quality of life" are often determined by the "social acceptability" of various diseases or conditions. [11] Similarly, Harris worries about the use of quality of life for resource allocation policy:

If for example some people were given life-saving treatment in preference to others because they had a better quality of life than those others, or more dependents and friends, or because they were considered more useful, this would amount to regarding such people as more valuable than others on that account. Indeed it would be tantamount, literally, to sacrificing the lives of others so that they might continue to live....To discriminate between people on the grounds of quality of as unwarranted as it would be to discriminate on the grounds of race or gender. [12]

There is, however, a key distinction between underestimating quality of life or using the concept to bring about invidious discrimination, on the one hand, and Oregon's use of it on the other. Specifically, the concerns expressed in the previous paragraphs involve judgments made about a person's quality of life at a given point in time--independent of any medical or surgical treatments. Such judgments are inappropriate bases upon which to ground resource allocation policy.

By contrast, the appropriate, non-discriminatory way to deal with quality-of-life information (and the approach adopted by the OHSC) is to focus on the change in quality of life expected with the use of a specific treatment or procedure. How much better or worse (if at all) is a patient's quality of life likely to be with application of a particular health service? This focus permits appropriate consideration to be given to the important impacts of treatments on quality of life which, as described earlier, are of considerable importance to the public in determining fair and rational systems of resource allocation. At the same time, the potential for discrimination is eliminated because treatments for handicapped or "poor quality of life" patients are evaluated on the same basis as are treatments for everyone else. It is the change in quality of life, or net benefit, realized from a treatment that matters, not the point-in-time quality of life of a patient.

Quality-Adjusted Life Years

After consulting with advisors at Oregon Health Decisions, the OHSC decided to incorporate quality of life considerations into the priority-setting process using the "quality-adjusted life year" (QALY) approach. [13] This method permits integration of the quality-of-life effects of treatment with its associated impacts on life expectancy. Some treatments, such as appendectomies, are valued not for any improvements in quality of life, but rather for their substantial positive effects in life expectancy: in the case of appendicitis, going from perhaps two weeks to normal. Other treatments have significant impact on quality of life, but little or no effect on life expectancy--such as medication or surgery for arthritis, or prostatectomy for benign obstruction. Still other treatments involve trade-offs between quality and quantity of life, where a longer life expectancy may come at the expense of various side effects from treatment, resulting in a possible decrease in quality of life.

Use of the QALY approach requires the explicit estimation of "percentage of normal quality of life." One


year of "normal" quality of life is considered equivalent in value to two years of "one-half normal" quality of life: moreover, two treatments offering these respective outcomes would be valued equally (other things being equal). The QALY concept is useful primarily because it reminds us of a few key principles:

Necessity. We have no choice but to consider both quality and quantity of life in some integrative fashion to properly evaluate health care services.

Common Sense. Other things being equal, treatments offering fewer net benefits in terms of either quality of life, longevity, or both, should be valued less highly than treatments offering more such benefits. Thus, if two treatments each offer about a year of additional life, the one that offers greater benefits in terms of quality of life should be favored.

Proportionate Value. Building on the last concept, treatments should be valued roughly in proportion to the degree of benefit they offer to patients. Thus, we should be able to distinguish between treatments that offer highly valued outcomes, such as comfort and the relief of pain that is characteristic of hospice programs, from treatments that provide less-valued outcomes, such as the marginal prolongation of life with severe side effects characteristic of many aggressive treatments for terminally ill patients. The QALY approach in theory permits this sort of distinction to be made.

Several problems with the QALY concept have so far caused it to remain merely a heuristic device, by limiting real-world application:

Questionable Assumption. The QALY method assumes that people see no difference between, say, one year of normal-quality life and ten years of life at one-tenth quality (whatever that is). The QALY approach assumes that a short, good life is of equal value to a long, ailing one. This assumption seems unlikely to be valid.

Equity Problems. The QALY approach suffers from a limitation common to any purely utilitarian construct: our intuitive rejection of conclusions to the effect that one person should be treated rather than many. In particular, QALY logic supposes that if a choice comes down to treating one person who stands to gain ten QALYs, or nine people who each stand to gain one QALY, then the single person should be treated.

Measurement Problems. What does it mean to speak of someone having a "one-tenth normal" quality of life? How could we come up with such an overall numerical estimate? Anyway, how can we quantify a quality, especially one so amorphous and ill-defined as quality of life?

The first two of these problems speak to the fact that QALYs cannot be used as the sole basis for resource allocation decisionmaking. This conclusion is hardly new; even the staunchest advocates of the QALY concept realize that the conflict between individual and societal preferences and issues of justice and equity must also be entered into the resource allocation equation.

It is the measurement problem that is responsible for the fact that QALYs have had so little impact on the health care system since their introduction almost twenty years ago. And it is here that the OHSC made its greatest methodological contribution to the goal of setting health care priorities. As such, the Oregon process represents the first large-scale effort in the United States to operationalize the QALY concept for purposes of resource allocation policy.

The OHSC Method for Quality of

Life Measurement

The methodology adopted by the OHSC for estimating QALYs was based on a set of descriptions of poor


health or impaired quality of life borrowed (after slight modification) from a set of generic quality-of-life states developed by Kaplan and Bush. [14] These states consist of descriptions of physical or emotional symptoms and of different degrees of limitations or impairment in mobility, physical activity, and social activity. The OHSC assigned each quality-of-life problem a decrement value, based on the results of a telephone survey of 1,000 randomly selected Oregonians (see below), estimating the degree to which each problem would reduce overall quality of life. Figure 1 shows some of the quality-of-life states used by the OHSC, as well as their associated decrement values. For example, "moderate limitation in physical activity" and "seizure, fainting, or coma" were estimated to reduce overall quality of life by about 37 percent and 11 percent, respectively.

To derive a numerical estimate of the quality-of-life benefit from treatments, local physicians predicted which problems, if any, patients would experience five years after diagnosis, both with and without treatment. (The five-year window was selected as a reasonable yardstick against which to standardize the expected benefit of different treatments and procedures.) The numerical difference between the aggregate values of the treated compared to untreated health states was considered to be the net benefit from treatment.

An example of this process is shown in Figure 2; analogous calculations were performed for each service on the priority list. In this case, the evaluated treatment was surgery for bleeding into the brain. Consultant physicians estimated that without surgery 95 percent of patients with this condition would die within five years, and that the five-year mortality rate would drop to 20 percent if surgery was performed. Furthermore, physicians predicted that all untreated survivors would experience severe disability, including difficulty with cognition, whereas "only" 40 percent of treated patients would experience such a disability. On the other hand, 20 percent of the patients


who undergo surgery for this condition were predicted to suffer from residual side effects (loss of consciousness and difficulty speaking) that were not predicted to occur in untreated patients. Thus, the expected risks and side-effects of treatment were "deducted" from any expected benefits in estimating overall impacts on quality of life.

As shown in Figure 2, the predicted aggregate health status of untreated patients figured to a little less than zero--a little worse, that is, than being dead. Treated patients, on average, were expected to be restored to about 35 percent of normal quality of life (i.e., 0.348), which is a substantial improvement over a "worse-than-death" state. This example also demonstrates the lack of discrimination against handicapped people that occurs when only changes in quality of life with treatment are considered. Thus, even though most survivors of this condition are expected to be significantly handicapped, even with treatment, the large improvement in expected quality of life resulted in a relatively high priority ranking for this treatment (no. 86).

The impact of treatment on life-expectancy--the other component of the QALY measure--was integrated in two ways. First, life expectancy was factored into the equation by virtue of the "zero quality of life" status assigned to the percentage of patients estimated to die in each group. Second, the net-benefit component was multiplied by the expected duration of that benefit for computation of cost-effectiveness ratios, although the "net-benefit" component alone--without multiplication by expected duration of benefit--was used to prioritize services within categories for creating the final priority list.

Analysis of the Oregon Process

Several features of the described method for setting priorities are of interest. Most fundamentally, perhaps, the QALY approach to priority setting was a systematic alternative to more traditional methods of involving the public in health care resource allocation. More typically, citizens are asked which programs they prefer: neonatal intensive care units, transplant surgery programs, screening for high blood pressure, hospice care, and so on. This approach is limited by the fact that most people know very little about the expected health outcomes associated with these programs, or about associated costs and other important factors that should shape public policy. By contrast, the method adopted by the OHSC asked people about their preferences for the outcomes of care, including various states of ill health and disability. Everyone has been ill at one time or another and has experienced at least temporary periods of disability. Most people are in a position to recall what care can do, and therefore can offer opinions about various states of health that are better informed than their opinions about programs.

Several questions arise, however, regarding (1) the process of asking people to imagine a state of poor health and to apply a rating to that state and (2) the subsequent process of aggregating individual ratings. I take up these problems more fully elsewhere, [15] but a couple of issues deserve brief mention here. First, I believe that Oregon was correct in surveying generally healthy citizens, rather than actively ill patients, because the inevitable biases of the latter group would hopelessly confound the rating process. True, people may have difficulty imagining states of poor health or of impaired quality of life, but by surveying enough people (and 1,000 randomly selected individuals is more than enough) an adequate indication of public preferences can probably be obtained. Moreover, in related work we have found that the task of imagining and rating health states can be substantially facilitated by using visual representations of the various states, and by using paired comparisons in addition to simple ratings of individual health states. [16]

But what about the prototypical problem of the "happy quadriplegic"? How can we account for the fact that many patients with chronic illnesses or disabilities adjust emotionally to those states? This is a common enough finding. Might it be a mistake, therefore, to label a severely impaired or painful health state as less desirable than one without impairment or pain? Surely no, for although people in wheelchairs may adjust to their situation, it does not follow that they would not value a procedure that enabled them to walk again. This consideration relates again to the importance of focusing on changes in quality of life brought about by health services.

Another important observation is that OHSC's use of generic health states (those that can be applied to any illness or condition, rather than illness-or treatment-specific outcomes) enabled all types of health services to be denominated in terms of a "common currency" of benefits and harms. This is a critical consideration from the perspective of health care resource allocation, [17] since it is necessary to compare directly the outcomes of, say, coronary bypass surgery with those of chemotherapy to assess the relative value of these two procedures.

Thus, Oregon's process--explicitly estimating generic treatment outcomes, weighted according to empirically derived public values for those outcomes--was a sophisticated exercise in priority setting, congruent with contemporary thinking. Just this sort of approach is required if public values are to be meaningfully incorporated into the priority-setting process.

Inevitably, however, the Oregon effort to implement a preference-for-outcomes approach was limited by imperfections in available methods. A couple of examples will suffice. The described health states are very broad, often encompassing a wide variety of problems and conditions. For example, the "trouble talking" state incorporates everything from a slight lisp to total mutism. Moreover, the decrement values assigned to the states often do not make sense; "unable to use transportation outside the home" was rated the same as "bed-bound," while "burn over large areas of the body" was rated the same as an "upset stomach."

The rating problem may be due in part to the simplified telephone rating task used to obtain decrement values. Raters were asked to assign a numerical rating between 0 and 100 (0=death, 100=good health) to each of 10 health states. A typical scenario was: "You can go anywhere and have no limitations on physical or other activity but have stomach aches, vomiting, or diarrhea." (Ratings were then rescaled to a decrement value between 0 and 1, with higher numbers associated with more serious, lower-rated problems.) Thus, only one problem at a time was stipulated to be present in each scenario; everything else was assumed to be normal. Obviously, this assumption is problematic, since illness and disability tend to occur together. It is difficult to imagine a patient with a large burn, for example, who is not limited in mobility and social activity. In related work, we have observed that the use of a similar one-dimensional, direct rating approach failed to produce valid preference ratings. [17]

A related potential problem with OHSC's quality-of-life measurement method is that the decrement values of the various problems and conditions are assumed to be cumulative in their effect on overall quality of life. There is evidence, however, that the interactions among quality of life problems is more complicated than simple addition would suggest. [18]

These methodological limitations notwithstanding, the Oregon priority-setting exercise was a remarkably far-sighted and sophisticated effort to incorporate quality of life measurement into resource allocation policy. The OSHC correctly focused on the change in quality of life afforded by treatments, rather than on point-in-time quality of life assessments of patients or on people's feelings about health care programs. The method selected by the OHSC properly included the explicit estimation of (generic) outcomes of treatment and empirically derived preferences for those outcomes.

Only the details of this process--among them, better specification of health states--need change in subsequent iterations of the Oregon process. In the meantime, can Oregon's final list serve as a legitimate basis for public policy? The answer to this question is unclear, and may depend on whether and how Oregon better specifies treatments and indications for treatment. [19] Much more will be known in a few months, after the Oregon legislature has considered the list and federal action is taken on a Medicaid waiver. The events in Oregon are critically important for American health and social policy; they will bear close watching over the coming months and years.

David C. Hadorn is senior research associate at the Center for Health Ethics and Policy, University of Colorado, Denver, Col., and a postdoctoral fellow in health services research at The RAND Corporation, Santa Monica, Calif.


[1] Oregon Senate Bill 27.

[2] Milton Weinstein and William Stason, "Foundations of Cost-effectiveness Analysis for Health and Medical Practices," NEJM 296 (1976):716-21.

[3] T. Egan, "Problems Could Delay Proposal by Oregon to Ration Health Care," New York Times, 30 July 1990.

[4] David Hadorn, "Setting Health Care Priorities in Oregon: Cost-effectiveness Meets the Rule of Rescue," JAMA 265 (1991):2218-25.

[5] Hadorn, "Setting Health Care Priorities."

[6] Hadorn, "Setting Health Care Priorities."

[7] Dean Mayer and Merit Kimball, "Oregon Commission OKs Medicaid Pecking Order," Healthweek, 25 February 1991, pp. 1, 36.

[8] Brian Hines, "Health Policy on the Town Meeting Agenda," Hastings Center Report 16, no. 2 (1986): 5-7; Bruce Jennings, "Community Health Decisions: A Grassroots Movement in Bioethics," Hastings Center Report 18, no. 5, Special Supplement (1988).

[9] "Quality of Life in Allocating Health Care Resources," adopted by the Citizens Health Care Parliament, 23-24 September 1988, Portland, Oregon.

[10] Jackob Najman and Sol Levine, "Evaluating the Impact of Medical Care and Technologies on the Quality of Life: A Review and Critique," Social Science and Medicine 15 (1981): 107-15.

[11] Warren Reich, "Life: Quality of Life," Encyclopedia of Bioethics (New York: The Free Press, 1978), pp. 829-40, at 837.

[12] John Harris, "QALYifying the Value of Life," Journal of Medical Ethics 13 (1987): 117-23, at 121.

[13] Graham Loomes and Lynda McKenzie, "The Use of QALYs in Health Care Decision Making," Social Science and Medicine 28 (1989): 299-308.

[14] Robert Kaplan and James Bush, "Health-Related Quality of Life Measurement for Evaluation Research and Policy Analysis," Health Psychology I (1982): 61-80.

[15] David Hadorn, "The Role of Public Values in Setting Health Care Priorities," Social Science and Medicine 52 (1991): 773-81.

[16] David Hadorn, T. Hauber, and Ron Hays, "Improving Task Comprehension in the Measurement of Health State Preferences: A Trial of Information Cartoon Figures and a Paired-Comparison Task." Unpublished.

[17] Hadorn, "The Role of Public Values."

[18] David Hadorn and Ron Hays, "Multitrait-multimethod Analysis of Health-related Quality of Life Measures," Medical Care 1991, forthcoming.

[19] G. Torrance, M. Boyle, and S. Horwood, "Application of Multi-Attribute Utility Theory to Measure Social Preferences for Health States," Operations Research 30 (1982): 1043-69.

[20] Hadorn, "Setting Health Care Priorities."
COPYRIGHT 1991 Hastings Center
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1991 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:cost effectiveness and quality of life as considerations of medical care in the Oregon Health Services Commission's implementation of the Oregon Basic Health Services Act of 1989; supplement
Author:Hadorn, David C.
Publication:The Hastings Center Report
Date:May 1, 1991
Previous Article:Setting health care priorities; Oregon's next steps.
Next Article:IVF International.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters