2427 E D I T O R I A L Economic Analysis of Expensive Technologies The Case of Erythropoietin Jennifer J. Griggs, M.D.1,2 Alvin I. Mushlin, M.D., Sc.M.1 1 Departments of Medicine and Community and Preventive Medicine, University of Rochester School of Medicine and Dentistry, Rochester, New York. 2 Department of Hematology/Oncology, University of Rochester School of Medicine and Dentistry, Rochester, New York. T See referenced original article on pages 2588 –96, this issue. Address for reprints: Jennifer J. Griggs, M.D., University of Rochester School of Medicine and Dentistry, 601 Elmwood Avenue, Box 704, Rochester, NY 14642. Received August 31, 1998; accepted September 18, 1998. © 1998 American Cancer Society he assumption underlying the use of economic analyses of health care interventions is that health care resources are limited and should be directed toward health strategies that provide the most benefit per dollar spent. Even in the absence of an explicit budgetary constraint, there is an implicit need to control excess spending. As outlined in the American College of Physicians Ethics Manual (Fourth Edition), physicians have a “responsibility to use all health-related resources in a technically appropriate and efficient manner” as they “consider the best interest of all patients and of each patient.”1 It is not enough to ask whether a new drug or device is effective—we must also ask at what cost, or “What are we getting for our health care dollar?” The Panel on Cost-Effectiveness in Health and Medicine, convened by the U.S. Public Health Service in 1993, views costs as “opportunity costs,” that is, dollars spent on one health care intervention are not available for others. In a cost-effectiveness analysis, the consequences of an intervention are expressed as “natural units,” for example, years of life saved, procedures avoided, or improvements in quality of life. The costeffectiveness ratio is therefore expressed as cost per year of life saved, etc. For erythropoietin used in the treatment of chemotherapy-related anemia, the natural unit might be “cost per transfusion avoided.” Cost-utility analyses (CUA) (one type of CEA) employs in the denominator the quality-adjusted life year (QALY). QALYs are calculated by multiplying the number of years of life by a quality adjuster (also called utility)—a number from 0 (“dead”) to 1 (often “perfect health”), thereby combining quality of life and life expectancy. Measuring utility attempts to capture the values, or preferences, that patients have for different health states. Little agreement exists as to which method for elicitation of utility is the “gold standard.”2 Furthermore, the criterion used to determine whether an intervention is cost-effective, and therefore to indicate that it should be adopted, is controversial. Cost-benefit analysis attempts to avoid these problems by measuring both costs and consequences in monetary terms. If the results indicate savings, the decision is fairly straightforward. For many 2428 CANCER December 15, 1998 / Volume 83 / Number 12 years, the “human capital approach” was used as a means of valuing benefits in monetary terms. Improved health should increase productivity, and wages could be used as a proxy for productivity and as a way of valuing improvements in outcome. The major limitations of this approach are as follows: 1) it assumes that the goal of a society is to promote health in order to maximize the gross national product, and 2) it is biased against some groups of people, including retired workers, children, volunteers, and women who do not work outside the home or who make less than their male peers. Furthermore, assigning a monetary value to a life saved or to improvements in health makes CBA less acceptable to physicians and other members of the medical community. Assessment of willingness to pay for an intervention is an alternative approach to assigning a monetary value to an intervention and its expected benefits.3 Respondents are asked how much money they would be willing to contribute to have a medical intervention made available to them or, alternatively, how much compensation they would require to forego such a benefit. As with the other methods, debate exists as to whose preferences should be measured: those people with the condition, or members of the society at large.4 – 6 The preference for a particular health state is generally higher when the participants are, in fact, patients.4 Erythropoietin in the prevention or treatment of anemia due to cancer chemotherapy is an example of a medical technology that offers benefits to some patients at relatively high cost. The consequences of transfusions versus erythropoietin may be measured in terms of increase in hemoglobin—for example, the ability of 1 unit of blood or 1 month of erythropoietin to increase the hemoglobin by 1 gram. Another approach would be to assess the improvements in quality of life that can be achieved with the two treatments or, alternatively, preferences for treatment with the two strategies. In this case, the incremental benefits of erythropoietin compared with transfusion are difficult to separate from the benefits of treating the anemia. Patients with both ESRD and cancer who have an increase in hemoglobin while on erythropoietin experience a concomitant improvement in functional status and quality of life.7–10 It is noteworthy that, among patients with ESRD, the Canadian Erythropoietin Study Group found no differences in the psychosocial aspects of health or in utility as measured by the time trade-off method, one commonly used approach to measuring preferences.11 That is, despite demonstrated improvements in exercise tolerance, the patients’ utility based on their “health status” or “how they feel” was no different in the transfusiontreated group compared with the erythropoietintreated group in this placebo-controlled, randomized trial. Until now, patient preferences for receiving erythropoietin versus transfusions have not been available, so cost-utility and cost-benefit analyses comparing these two treatment options have not yet been performed. In this issue of Cancer, Ortega et al. present the results of a willingness-to-pay approach to determine the cost-benefit of erythropoietin in the prevention of chemotherapy-induced anemia.12 Study subjects included both patients undergoing treatment with chemotherapy and members of the general public contacted via telephone. By including both patients and members of the general population in their study, the authors have addressed the cost-benefit of erythropoietin from both perspectives. Patients were asked how much they would pay for the drug itself. To make the question realistic to nonpatients, the payment for the drug was represented as an increase in health insurance premiums that would make the drug available should they require treatment in the future. Validation of this method of assessing preferences, in this study and others, is difficult because people rarely face the hypothetical decision being studied. One therefore cannot assess whether or not the responses to a questionnaire correspond to actual choices that people would make. The use of a scale by Ortega et al. to rate the importance of the new drug reveals what might be a disparity. Although participants considered it important to have the drug available for the prevention of cisplatin-induced anemia, with a mean value of 9 out of 10 for patients and 8 out of 10 for nonpatients, their willingness to pay for the drug did not render the drug cost-beneficial. A large part of this may reflect participants’ lack of knowledge regarding the actual cost of medical treatments in general. As one might expect, participants’ willingness to pay increases with individuals’ incomes, or ability to pay. This correlation has been demonstrated in other studies.13 Even at the highest level of willingness to pay, however, the drug was not cost-beneficial. An important caution in the use of health state scenarios is that all aspects of the health state and the technology under study must be included. In this case, for example, the state to be avoided should be anemia rather than merely transfusion therapy. Practical application of this study in determining health policy is also limited by the estimations of transfusion needs based on the Abels study of 55 patients. Differences among patient groups will lead to differences in the cost-benefit ratio. For a patient who Editorial/Griggs and Mushlin requires frequent transfusions, for example, continuous erythropoietin will be a more cost-effective treatment strategy than for a patient who requires transfusion only rarely. By varying the need for transfusion in additional sensitivity analyses, the authors may have revealed a subset of patients for whom erythropoietin would be cost-beneficial. Nonetheless, the use of the willingness-to-pay approach in evaluating the cost-benefit of prophylactic erythropoietin gives us important information about the value these subjects place on avoidance of transfusion. This method of valuation is understandable to participants and explores preferences for novel and expensive technologies. The use of an increase in insurance premiums as a model for cost makes the technique applicable to different types of health care systems. By enumerating clinical values in terms of cost-benefit analyses, we can find a way to include the preferences of our patients and future patients in negotiations with policymakers and payers. REFERENCES 1. 2. 3. 4. American College of Physicians. Ethics manual. 4th edition. Ann Intern Med 1998;128:576 –94. Froberg DG, Kane RL. Methodology for measuring healthstate preferences. II: Scaling methods. J Clin Epidemiol 1989; 42:459 –71. O’Brien B, Viramontes JL. Willingness to pay: a valid and reliable measure of health state preference? Med Decis Making 1994;14:289 –97. Boyd NF, Sutherland HJ, Heasman KZ, Tritchler DL, Cummings BJ. Whose utilities for decision analysis? Med Decis Making 1990;10:58 – 67. 5. 6. 7. 8. 9. 10. 11. 12. 13. 2429 Richardson J, Nord E. The importance of perspective in the measurement of quality-adjusted life years. Med Decis Making 1997;17:33– 41. Llewellyn-Thomas H, Sutherland HJ, Tibshirani R, Ciampi A, Till JE, Boyd NF. Describing health states: methodologic issues in obtaining values for health states. Med Care 1984; 22:543–52. Moreno F, Aracil FJ, Perez R, Valderrabano F. Controlled study on the improvement of quality of life in elderly hemodialysis patients after correcting end-stage renal diseaserelated anemia with erythropoietin. Am J Kidney Dis 1996; 27:548 –56. Canadian Erythropoeitin Study Group. Association between recombinant human erythropoietin and quality of life and exercise capacity of patients receiving haemodialysis. BMJ 1990;300:573– 8. Glaspy J, Bukowski R, Steinberg D, Taylor C, Tchekmedyian S, Vadhan-Raj S. Impact of therapy with epoitin alfa on clinical outcomes in patients with nonmyeloid malignancies during cancer chemotherapy in community oncology practice. J Clin Oncol 1997;15:1218 –34. Henry DH, Abels RI. Recombinant human erythropoietin in the treatment of cancer and chemotherapy-induced anemia: results of double-blind and open-label follow-up studies. Semin Oncol 1994;21:21– 8. McNamee P, van Doorslaer E, Segaar R. Benefits and costs of recombinant human erythropoietin for end-stage renal failure: a review. Benefits and costs of erythropoietin. Int J Technol Assess Health Care 1993;9:490 –504. Ortega A, Dranitsaris G, Puodziunas ALV. What are cancer patients willing to pay for prophylactic epoetin alfa?: a costbenefit analysis. Cancer 1998;83:2588 –96. Appel LJ, Steinberg EP, Powe NR, Anderson GF, Dwyer SA, Faden RR. Risk reduction from low osmolality contrast media: what do patients think it is worth? Med Care 1990;28: 324 –37.