FREE NEJM E-TOC    HOME   |   SUBSCRIBE   |   CURRENT ISSUE   |   PAST ISSUES   |   COLLECTIONS   |   Search Term  Advanced Search

Sign in | Get NEJM's E-Mail Table of Contents — Free | Subscribe

Previous Volume 339:1943-1944 December 24, 1998 Number 26 Next

Add to Personal Archive Add to Citation Manager Notify a Friend E-mail When Cited Related Article by Wright, J. C. Related Article by Wright, J. C. PubMed Citation

To the Editor: Wright and Weinstein's article (Aug. 6 issue)¹ on the gains in life expectancy from medical interventions allows physicians to compare the value of different medical interventions in various at-risk populations. We have used similar data to motivate patients to exchange bad health habits for good ones. Much as Friedman found that investors discounted the gain from investments to zero when the gain was more than three years away,² we have found that patients discount the value of good health choices for a gain in life expectancy in the seemingly distant future. When we translated medical interventions, through a mechanism similar to Wright and Weinstein's, into a more immediate time frame, we were successful in motivating patients to reconsider preventive health strategies. Our unit of measurement, called RealAge, is the measure of physiologic improvement that results from an intervention affecting health. Analogous to net present value, which makes investment decisions seem rational, RealAge is the translation of increased life expectancy into net present value. The addition of five years to one's life may not be a sufficiently immediate gain to provide the motivation to reduce one's blood pressure from 150/93 mm Hg to 120/80 mm Hg; however, some patients respond enthusiastically to the idea of retarding the speed of aging or becoming physiologically younger. Thus, a 49-year-old man who controls his blood pressure, quits smoking cigarettes, or exercises may have a RealAge of 37 years.

Like Wright and Weinstein, we found that the value of immunization was just six days, but control of blood pressure, cessation of cigarette smoking, exercise, or stress control makes one more than five years "younger." Various theoretical models have been developed to explain health-related behavior, including the health belief model, the theory of reasoned action, the social-learning theory, and the transtheoretical model and stages of change.³ The model proposed by Wright and Weinstein, or a version of it transformed into RealAge measurement of physiologic changes, borrows from the Chicago school of economics to motivate patients to choose healthy behavior. We believe Wright and Weinstein's model should be expanded to include either the best data or the results of a meta-analysis on each topic. The model could further include interaction and covariance terms derived from the literature or data base to account for interdependence among predictor variables. We believe that such models will allow patients and doctors to make more rational choices to enhance health.

Michael F. Roizen, M.D.
Keith Roach, M.D.
University of Chicago
Chicago, IL 60637

Axel Goetz, M.D., Ph.D.
RealAge, Inc.
San Diego, CA 92121

Editor's note: Drs. Roizen and Roach are consultants to RealAge, Inc., and have an equity interest in the firm.

References

1. Wright JC, Weinstein MC. Gains in life expectancy from medical interventions -- standardizing data on outcomes. N Engl J Med 1998;339:380-386. [Free Full Text]
2. Friedman M. Windfalls, the "horizon" and related concepts in the permanent income hypothesis. In: Christ CF, Friedman M, Goodman LA, et al., eds. Measurement in economics: studies in mathematical economics and econometrics in memory of Yehuda Gurnfeld. Stanford, Calif.: Stanford University Press, 1963:3-28.
3. Allegrante JP, Roizen MF. Can net-present value economic theory be used to explain and change health-related behaviors? Health Educ Res Theory Pract (in press).

We have reported the values for discounted gains in survival for a number of anticancer treatments (together with information on gains in survival for thrombolysis in patients with acute myocardial infarction).² A reanalysis of these oncologic data with and without discounting (annual discount rate, 3 percent or 5 percent) gives the following results. The discounted gain (with the undiscounted gain in parentheses) is 1.34 (3.08) years per patient for interferon therapy as compared with no adjuvant therapy in patients with high-risk resected cutaneous melanoma,³ 0.43 (0.45) year per patient for paclitaxel plus cisplatin as compared with standard chemotherapy in patients with advanced ovarian cancer, 1.06 (3.57) years per patient for cyclophosphamide, methotrexate, and fluorouracil as compared with no adjuvant chemotherapy in patients with node-positive breast cancer (not 3.6 undiscounted months per patient, as erroneously reported by Wright and Weinstein in Table 3 of their article), 0.89 (2.18) year per patient for intraportal chemotherapy as compared with no adjuvant chemotherapy in patients with colorectal cancer, 0.37 to 0.93 (0.30 to 1.22) year per patient for interferon therapy as compared with cytotoxic therapy in patients with chronic myelogenous leukemia,⁴ and 5.99 (3.01) years per patient for autologous transplantation as compared with chemotherapy in patients with chemosensitive non-Hodgkin's lymphoma in relapse. This information supplements the undiscounted data reported by Wright and Weinstein and confirms the well-known concept that discounting introduces marked changes in survival gains.

In conclusion, we believe that gains in survival are better described by reporting both the discounted and the undiscounted values. Although the approach that considers only undiscounted data is acceptable when survival data are not censored and there is no extrapolation, the presentation of both discounted and undiscounted values is more appropriate when the analysis has a lifetime perspective and the survival data are in part experimental and in part extrapolated.

Andrea Messori, Pharm.D.
Sabrina Trippoli, Pharm.D.
Enrico Tendi, Pharm.D.
Azienda Careggi
50134 Florence, Italy

References

1. Weinstein MC, Siegel JE, Gold MR, Kamlet MS, Russell LB. Recommendations of the Panel on Cost-Effectiveness in Health and Medicine. JAMA 1996;276:1253-1258. [Free Full Text]
2. Messori A, Trippoli S. Lifetime survival gain using streptokinase or TPA in acute myocardial infarction. (See: http://www.bmj.com/cgi/content/abstract/316/7141/1337#responses.) (Also available from NAPS, c/o Microfiche Publications, P.O. Box 3513, Grand Central Station, New York, NY 10163-3513.)
3. Messori A, Becagli P, Trippoli S, Tendi E. A retrospective cost-effectiveness analysis of interferon as adjuvant therapy in high-risk resected cutaneous melanoma. Eur J Cancer 1997;33:1373-1379.
4. Messori A. Cost-effectiveness of interferon in chronic myeloid leukaemia: analysis of four clinical studies. Ann Oncol 1998;9:389-396. [Free Full Text]

To the Editor: We thank Messori et al. for pointing out our incorrect transcription of their estimate of the gain in life expectancy from adjuvant chemotherapy for node-positive breast cancer, which is indeed 43 months, not 3.6 months, as we reported. This correction further strengthens our conclusion that gains in life expectancy tend to be higher for effective treatments than for primary or secondary prevention. We apologize for the error.

Messori et al. favor reporting discounted as well as undiscounted gains in life expectancy, citing the recommendations of the Panel on Cost-Effectiveness in Health and Medicine,¹ which one of us cochaired. These recommendations apply to cost-effectiveness analyses intended to inform resource allocations at the population level, not to decisions for individual patients. Exponential discounting may or may not be an appropriate way to measure personal preferences for future extensions of life at the individual level.² In any case, it is incorrect to discount because of the uncertainty surrounding extrapolation of survival data. Ideally, the underlying survival curves should be made available by the investigators, who have reported the source data, so that individual utility functions for survival gains can be applied.³

The question of the most effective way to communicate information on survival gains to patients and physicians is addressed by Roizen et al., who favor a measure based on the equivalent reduction in chronologic age. Although actuarially equivalent to the gain in life expectancy, their RealAge measure places a different "spin" on the data, in an attempt to offset the often incorrect perception that gains in life expectancy accrue at the end of life. Their efforts, along with those of others who study the communication of information about risk, should be applauded. Much more needs to be learned about communicating health risks and benefits effectively, both in terms of enhancing an understanding of them and in terms of promoting informed health decisions by patients and physicians.

Milton C. Weinstein, Ph.D.
Harvard University
Boston, MA 02115

Janice C. Wright, Ph.D.
63 Houghton Bay Rd.
Wellington, New Zealand

References

1. Weinstein MC, Siegel JE, Gold MR, Kamlet MS, Russell LB. Recommendations of the Panel on Cost-Effectiveness in Health and Medicine. JAMA 1996;276:1253-1258.
2. Redelmeier DA, Heller DN. Time preference in medical decision making and cost-effectiveness analysis. Med Decis Making 1993;13:212-217.
3. McNeil BJ, Weichselbaum R, Pauker SG. Fallacy of the five-year survival in lung cancer. N Engl J Med 1978;299:1397-1401. [Abstract]

Add to Personal Archive Add to Citation Manager Notify a Friend E-mail When Cited Related Article by Wright, J. C. Related Article by Wright, J. C. PubMed Citation

This article has been cited by other articles:

• Charlson, M. E., Allegrante, J. P., McKinley, P. S., Peterson, J. C., Boutin-Foster, C., Ogedegbe, G., Young, C. R. (2002). Improving health behaviors and outcomes after angioplasty: using economic theory to inform intervention. Health Educ Res 17: 606-618 [Abstract] [Full Text]