Correspondence
Quality of Life after Myocardial Infarction: Canada versus the United States
N Engl J Med 1995; 332:469-472February 16, 1995
- Article
To the Editor:
. . . In the study by Mark et al. (Oct. 27 issue),1 I am struck by the large difference between Canada and the United States in the use of invasive diagnostic and therapeutic procedures in patients with acute myocardial infarction, despite small differences, if any, in survival or quality of life. The rate of angiography was almost 200 percent higher among U.S. patients during hospitalization for acute infarction (72 percent vs. 25 percent), and the rate of percutaneous transluminal coronary angioplasty or coronary-artery bypass surgery at one year was over 100 percent higher among the U.S. patients (53 percent vs. 24 percent). I would argue that these were the only substantial differences between the two groups. The U.S. patients had much more treatment and at considerable expense, with little if any benefit to show for it. The data presented by Mark et al., rather than provide the basis for a tribute to U.S. medicine, in fact cast further doubt on its cost effectiveness.
1 ReferencesMayer Bassan, M.D.
11 Windmill St., Jerusalem, Israel1
Mark DB ,Naylor CD ,Hlatky MA , et al. Use of medical resources and quality of life after acute myocardial infarction in Canada and the United States. N Engl J Med 1994;331:1130-1135
Full Text | Web of Science | Medline
To the Editor:
Mark et al. suggest that patients in the United States had a better quality of life after acute myocardial infarction than similar patients in Canada. The authors hypothesize that these “superior outcomes are due to differences in the process of care — specifically, the greater use of revascularization procedures and specialist services.” Although this hypothesis may be true, it does not address the essential economic question: Was the value of the improvement in the quality of life among the U.S. patients greater than the value of other improvements that might have been achieved by using the resources in other ways (e.g., treating more patients)? Conversely, was the loss in the quality of life among the Canadian patients outweighed by the improvements resulting from the alternative use of resources? Assuming that the more aggressive pattern of care in the United States required more resources, the real issue concerns the opportunity costs of those resources.
Neville Doherty, Ph.D.
University of Connecticut Health Center, Farmington, CT 06030To the Editor:
The article by Mark et al. suggests that the more aggressive U.S. pattern of care after myocardial infarction may lead to a better quality of life than the less aggressive Canadian pattern. In this era of health care reform, their article may fuel distrust of a Canadian-style system. We propose that the data do not support the authors' conclusions.
As they mention, an observational study may have biases and unmeasured confounding variables. For example, different base-line socioeconomic characteristics may explain any perceived differences in the quality of life. The only index of socioeconomic class that Mark et al. report is educational level, which was lower among the Canadian patients. Intuitively, one might expect noncomparability due to selection bias. Since Canada has universal coverage, all classes of Canadian patients were enrolled in the Global Utilization of Streptokinase and t-PA [tissue plasminogen activator] for Occluded Coronary Arteries (GUSTO) trial, whereas the fact that 72 percent of the American patients received angiography suggests a more selected population with adequate health care coverage and different expectations.
The investigators found no difference in unadjusted mortality rates. (The suggestion that adjusted mortality favors U.S. patients is difficult to explain, since there is no obvious imbalance in the base-line prognostic factors presented in Table 1 of the article, some of which — age, cigarette smoking, blood pressure on admission, and time to treatment — were actually more favorable among the U.S. patients.) Although many quality-of-life scales were used, none have been validated in a Canadian population, especially among French-speaking patients during telephone interviews.
Assuming that the authors actually measured what they claim, the clinical importance of the results is questionable. In large samples trivial differences can reach statistical significance. For example, the difference of 4 mm Hg in systolic blood pressure gives a P value of less than 0.001. Similarly, what is the clinical importance of P values of less than 0.001 for differences in scores on the restricted-activity scale and Carroll Depression Scale when the medians are identical? (Also, why were confidence limits not reported?) As the authors mention in the body of their article but not in the abstract, perceptions of general health and the rate of return to work were the same in the two groups of patients. Also, the more widely used and extensively validated Rose questionnaires showed no difference in the incidence of dyspnea of class 2 or higher.
The uncertainty about what is actually being measured is best illustrated by Figure 3 of the article. If improved access to revascularization accounted for differences in the quality of life, how can one explain the statistically significant difference in the quality of life between Canadian and U.S. patients who both received the desired intervention? The authors' explanation (that the Canadian patients who underwent revascularization were sicker than the U.S. patients and so could not benefit as much from the procedure) is a clumsy and convoluted attempt to account for a major discordance between the results and the authors' conclusions. The “improved” quality of life among U.S. patients, as compared with Canadians, at one year was observed for both patients undergoing revascularization and those not undergoing revascularization, suggesting that this intervention was not responsible for the changes.
Even if one ignores the probable noncomparability of the two groups of patients, the uncertainty about what was actually measured, and the potential lack of clinically meaningful differences, it is possible that any differences are due not to the different health care systems in the two countries but instead to the free, informed choices of Canadian patients and their physicians. Canadian physicians not only follow outcome-based pharmacologic research more closely than their U.S. counterparts, as stated in the article, but also seem to pay closer attention to the results of randomized trials showing no advantage of routine early catheterization after myocardial infarction.1,2
Canadian physicians have in their pattern of care obtained identical mortality rates and similar, if not identical, morbidity rates while using two thirds less resources than U.S. physicians. The correct interpretation of this study should help answer the question, “How does Canada do it?”3
3 ReferencesJ.M. Brophy, M.D.
L. Joseph, Ph.D.
McGill University, Montreal, QC H3G 1Y6, Canada1
The TIMI Study Group
Full Text | Web of Science | Medline2
SWIFT (Should We Intervene Following Thrombolysis?) Trial Study Group
CrossRef | Web of Science3
Fuchs VR ,Hahn JS . How does Canada do it? A comparison of expenditures for physicians' services in the United States and Canada. N Engl J Med 1990;323:884-890
Full Text | Web of Science | Medline
To the Editor:
Mark et al. appropriately define quality of life as a package of variables (functional status, angina, dyspnea, general health, employment status, and general psychological well-being) rather than a single variable. Their package, however, is missing some components. Sexual activity and performance are undeniably important factors in the quality of life, and they can become problematic after acute myocardial infarction. Cognitive functions, such as memory, concentration, perception, vigilance, and psychomotor performance, are also important factors that can be affected by cardiovascular disease. The same applies to the quality of sleep and the duration of its phases. In addition, potential differences between male and female patients in their subjective views of the importance of the various aspects of quality of life should be addressed.1
1 ReferencesZoltan Vajo, M.D.
Maricopa Medical Center, Phoenix, AZ 85008To the Editor:
Although it is true that employers pay the bills and therefore are customers of managed-care organizations, an employer's goal is not only to reduce medical expenses through the efficient use of resources but also to have employees return to work in a timely fashion, contributing to overall productivity. At times, managed-care organizations delay the return to work by delaying a referral to a specialist. The industrial athlete is left sitting on the bench, waiting for the generalist to decide whether and, if so, when to provide a referral to a specialist or order a definitive test. The disincentives to referral at times cost corporations much more in payments for short-term disability and lost productivity than a one-time visit to a specialist.
In the study by Mark et al. I found it interesting that in the United States, patients with acute myocardial infarction returned to work after a median of 58 days, whereas in Canada the median was 81 days. The total cost to employers in Canada was substantial when one considers that Canadian patients were out of work 23 days longer than U.S. patients.
When assessing the cost of medical care and the use of medical resources, one must consider the total cost. Short-term–disability payments and lost productivity add substantially to the cost of doing business and must be considered when estimating the cost efficiency and cost effectiveness of managed care. Managed-care organizations should be held accountable for the return-to-work outcome and evaluated accordingly. This may help remedy both underutilization and overutilization of resources.
Barton Margoshes, M.D.
Bell Atlantic, Arlington, VA 22201To the Editor:
. . . In the article by Mark et al., the inclusion of large numbers of patients resulted in statistically significant differences between the U.S. and Canadian groups for many outcomes that were quantitatively similar. At one year, median values for days of restricted activity, the Duke Psychological Scale, and the Carroll Depression Scale were virtually indistinguishable despite P values of less than 0.01. Differences in dyspnea and chest pain diminished when only severe dyspnea or typical angina was considered. General perceptions of health were better in the U.S. cohort only when rated on a scale of 1 to 100, not when rated as poor, good, or excellent. The difference in the effect of health on functioning was predominantly due to the additional 6 percent of Canadian patients who reported having trouble doing some things, but at base line this cohort had a 6 percent higher frequency of previous angina. Instead of challenging “the contention that the greater use of cardiac procedures in the United States as compared with Canada has no effect on health outcomes,” the study reveals that despite a 120 percent higher rate of revascularization in the U.S. group, there was only a 17 percent increase in the Duke Activity Status Index, only a 6 percent increase in the current-health score, and no substantial difference in general health perceptions or work status.
Health care factors affecting the quality of life are extremely complex, and the authors should be congratulated on their analysis. The suggestion that more invasive procedures and more specialists result in a better quality of life may be true. Given the substantial implications for health care policy, however, one must be cautious in interpreting the results of this study as evidence in support of a more aggressive pattern of care.
Lawrence I. Deckelbaum, M.D.
West Haven Veterans Affairs Medical Center, West Haven, CT 06156To the Editor:
In our view, the results reported by Mark et al. are considerably weakened by inattention to the nonspecific nature of functional status or health-related quality of life. Functional status can be profoundly affected by a coexistent disease,1-3 sometimes as much as by the primary disease under investigation.
In the study by Mark et al., the Canadian patients had an average of two coexistent diseases, whereas the U.S. patients had an average of one (P = 0.03). Furthermore, the rate of smoking was higher, though not significantly so (P = 0.07), among the Canadian patients. Given that dyspnea — which along with angina was more frequent among the Canadian patients — is also a symptom of chronic lung disease, differences in coexistent diseases, especially chronic lung disease, could have accounted for part, if not all, of the significant but relatively small differences in quality of life (5 points on a 55-point scale) that favored the U.S. patients. Because of space considerations, Mark et al. may have omitted a multivariable model in which comorbidity was assessed in terms of its effect on functional status.
The similarity of the results in the U.S. and Canadian groups at 30 days does not rule out the possibility that coexisting conditions influenced the outcomes at 1 year, when the profound effect of the acute hospitalization had diminished.4
There are other reasons to think that the results may not be as convincing as Mark et al. suggest. For one thing, the general health perceptions were equivalent in the two groups, as were the days of restricted activity. Both measures have had considerably wider use than the Duke Activity Status Index.
In addition, as the authors note, cultural differences could have come into play, although they suggest that the similarity in function at 30 days makes that possibility unlikely. The similarity at 30 days, however, does not reduce the importance of social and cultural values, which have been demonstrated to account for large differences in quality-of-life measures among countries. In the initial period after hospitalization, the health measures may reflect the profound effect of the acute illness and hospitalization.5 The cultural differences in the way people perceive and report their quality of life are more likely to emerge over time, under the circumstances of everyday living.
5 ReferencesSheldon Greenfield, M.D.
Sherrie Kaplan, Ph.D.
New England Medical Center, Boston, MA 021111
Cleary PD ,Greenfield S ,Mulley AG , et al. Variations in length of stay and outcomes for six medical and surgical conditions in Massachusetts and California. JAMA 1991;266:73-79
CrossRef | Web of Science | Medline2
Greenfield S ,Apolone G ,McNeil BJ ,Cleary PD . The importance of co-existent disease in the occurrence of postoperative complications and one-year recovery in patients undergoing total hip replacement: comorbidity and outcomes after hip replacement. Med Care 1993;31:141-154
CrossRef | Web of Science | Medline3
Greenfield S, Sullivan L, Dukes KA, Silliman R, D'Agostino R, Kaplan SH. Development and testing of a new measure of case mix for use in office practice. Med Care (in press).
4
Keeler EB ,Kahn KL ,Draper D , et al. Changes in sickness at admission following the introduction of the prospective payment system. JAMA 1990;264:1962-1968
CrossRef | Web of Science | Medline5
Kaplan SH, Shea J, Dukes K, Kahn A, Schwartz S, Greenfield S. The effects of laparoscopic vs. traditional cholecystectomy on health related quality of life and recovery following surgery. Soc Gen Intern Med (in press).
Author/Editor Response
The authors reply:
To the Editor: Several correspondents question whether the better quality of life among U.S. patients in the GUSTO trial, as compared with Canadian patients, is in proportion to the extra use of resources and costs required to produce this benefit. We are currently comparing the incremental cost effectiveness of the U.S. and Canadian practice styles as reflected in the GUSTO trial. However, neither country currently considers opportunity costs or the efficiency of resource allocation (based on cost-effectiveness ratios) in setting health policy.
Drs. Brophy and Joseph raise many issues but miss the broader question: How do patients and policy makers view the differences we observed in symptoms, functional status, and time required to return to work? Are these differences important enough to warrant changes in the process of care or do we need more evidence? Rather than rationalize the frequency of morbidity, delay in return to work, and other problems identified in the Canadian group, we suggest that Drs. Brophy and Joseph determine how to enhance the return to a productive, symptom-free existence, employing an open mind and a variety of strategies that need further investigation, including elements of the more aggressive U.S. style of care after myocardial infarction.
Drs. Brophy and Joseph suggest that the high rate of angiography among U.S. patients indicates that enrollment in the GUSTO trial in the United States was dictated by insurance status. Actually, it is much more likely that the rate of angiography was a function of the hospitals and investigators involved. According to 1992 Medicare data for the entire United States, more than 50 percent of patients with acute myocardial infarction who were treated by cardiologists underwent angiography within 60 days (unpublished observations). The rates were significantly lower among patients receiving care from primary care physicians.
Dr. Vajo comments on the importance of sexual function, sleep, and cognitive function in determining the overall quality of life. A detailed assessment of these factors was beyond the scope of our study.
We agree with Dr. Margoshes that a full economic evaluation of the differences between Canada and the United States should take into consideration the indirect costs related to diminished productivity, as well as direct costs.
We also agree with Drs. Greenfield and Kaplan that functional status in patients with coronary disease may be affected by coexisting disease, and we examined this possibility to the extent that our data permitted. Smoking status at the time of myocardial infarction did not account for the observed difference in exertional dyspnea at one year. Greenfield and Kaplan are incorrect, however, that the number of days of restricted activity was the same in the two countries. Although the medians were the same, the distributions were not. The discrepancy between the ordinal measure of general health perception (rated poor to excellent) and other measures of general health, functioning, and symptoms suggests that this scale is insufficiently sensitive or that it reflects the remarkable ability of patients to accommodate to whatever level of health and functioning they believe they must accept.
Our findings could best be tested by a randomized trial of Canadian-style care versus U.S.-style care in each country, but it is unlikely that such a trial will be conducted. We believe that in the absence of data from such a randomized trial, our observational comparisons generate useful questions about the effectiveness and efficiency of care after myocardial infarction on both sides of the border.
Daniel B. Mark, M.D., M.P.H.
Duke University Medical Center, Durham, NC 27708-3485C. David Naylor, M.D., D.Phil.
University of Toronto, Toronto, ON M5S 1A8, CanadaPaul W. Armstrong, M.D.
University of Alberta, Edmonton, AB T6G 2R7, Canada
