Importance of Cost and Quality of Life in Decisions About Routine Angiography After Acute Myocardial Infarction
The Role of Cost-effectiveness Models
In 1992, health care was the fastest-growing sector of the US economy, accounting for 14% of the gross national product. Amid concern about whether such rapid growth could realistically be sustained, the editor of the Journal of the American Medical Association predicted that the US healthcare system would experience an apocalyptic “meltdown” by 1996 if the situation persisted.1 Four years later, the US healthcare system has not experienced “meltdown.” However, the manner in which healthcare services are provided, financed, and evaluated has changed dramatically. Healthcare insurers and policy makers have justified these changes by asserting that Americans receive much unnecessary health care and that even the necessary care is overpriced. In contrast to the US situation, Canada and the Western European countries spend less per capita on health care but have apparently similar levels of overall health as measured by key indicators.
Beginning with the use of diagnosis-related groups in the early 1980s, hospitals and physicians witnessed an accelerating shift of financial responsibility from insurers and other health services purchasers to healthcare providers. By 1992, many providers had found that reimbursements from Medicare and Medicaid were not adequate to cover their expenses for these services.2 3 Four years later, capitation and other forms of managed care have so limited reimbursements from private insurers that many providers are unable to cover their expenses for these patients as well. In this increasingly competitive healthcare environment, providers, purchasers, and policy makers are finding that they can no longer finance healthcare services at their pre-1992 levels, and they are faced with making hard decisions regarding the appropriateness and cost-effectiveness of competing medical therapies. Often, these decisions are made without access to relevant economic information and without the benefit of good long-term outcome information.
Use of Invasive Cardiac Procedures
Because of the high costs and the frequency with which they are performed, coronary revascularization procedures have been identified as a good target for cost-containment and waste-reduction efforts. Although some evidence in the form of randomized clinical trials now exists to help define the appropriate role of these procedures for selected groups of patients, the available trials clearly do not address all coronary revascularization entry points or all patient subgroups. For example, acute myocardial infarction patients frequently receive angiography and subsequent revascularization as part of their treatment regimen. Only two randomized trials (TIMI II and TAMI 5) have evaluated both the clinical efficacy and the economics of early angiography and percutaneous transluminal coronary angioplasty as adjuncts to thrombolytic therapy. The TIMI II economic substudy4 enrolled 376 patients and found an ≈$2000 difference in index hospitalization costs between the conservative and invasive strategies. The TAMI 5 economic substudy5 enrolled 575 patients and found that the aggressive interventional strategy was ≈$2500 more costly than the deferred strategy. Neither study was powered to look at differences in costs for important patient subgroups, and neither study was funded to assess the long-term implications of these strategies.
Although angiography is substantially less expensive than coronary revascularization, healthcare policy makers have a strong interest in the rates at which angiography is performed, because they see the procedure as a gatekeeper for other interventional procedures. There are wide variations in angiography utilization rates for acute myocardial infarction patients among the United States, Canada, and European countries and among regions within the United States.6 7 8 Within the United States, the use of angiography in patients with acute myocardial infarction appears to be a linear function of its availability in the admitting hospital.7
Once angiography is performed, referral for revascularization occurs at a surprisingly uniform rate in different regions of the United States and in different countries,7 despite the differences in angiography rates.9 Thus, the critical determinant of the use of revascularization procedures seems to be the use of angiography; healthcare policy makers may reasonably conclude they will be able to hold down the overall costs of care for acute myocardial infarction if they are able to limit access to angiography during the index hospitalization. Although the cost side of this issue is relatively easy to measure, the relationship with health benefits (ie, the value of the care provided) is much more difficult to define.
Decision Analysis and Cost-effectiveness Models
To address these health policy issues, we are in great need of access to large, high-quality national databases with carefully collected long-term outcome and cost data, but neither the federal government nor the major healthcare payers are currently willing to finance such efforts. Kuntz and colleagues,10 in this issue of Circulation, demonstrate a methodological alternative to such national databases. They used decision analysis and cost-effectiveness analysis to synthesize information from existing sources (published randomized clinical trials, other relevant literature, Medicare Part A patient-billing information, a survey of 1051 patients with recent acute myocardial infarction, and the Coronary Heart Disease Policy Model11 ) to assess the cost-effectiveness of routine angiography for patients with acute myocardial infarction. Although this work parallels and extends previous efforts that have sought to stratify patients with acute myocardial infarction according to risk, an important additional contribution is the incorporation of cost and quality-of-life considerations into the assessment.
Decision analysis and cost-effectiveness models have a long history in cardiology outcomes research. Several of the Agency for Health Care Policy and Research's patient outcomes research teams12 have used such models to combine outcomes and cost information from many sources and to summarize existing cost-effectiveness relationships for important patient subgroups. Although they generally do not generate new discoveries from existing data, these models can serve to focus attention on the assumptions underlying clinical practice patterns and can help to identify areas in which there is substantial disagreement about appropriate evaluation and treatment strategies. If used in this manner, they have the potential to guide research agendas and to inform clinical practice guidelines.
The implications of the article by Kuntz et al10 may be considered in two categories: practical, for physicians caring for patients and for physicians and administrators crafting guidelines for healthcare systems; and conceptual, for policy makers using cost-effectiveness to guide decision making in healthcare systems.
At first glance, the findings of this analysis may appear to be at variance with information from randomized clinical trials that individually have failed to show a benefit of routine postinfarction angiography. Adoption of the algorithm in Fig 3 of the article by Kuntz et al, for example, would lead to angiography in all patients with severe angina and in all men between the ages of 45 and 64 with prior infarction. Women <65 years old with prior myocardial infarction would generally undergo angiography only if they had an ejection fraction <50% or a positive exercise test for ischemia. Interestingly, patients >65 years old without prior AMI would undergo angiography only if their exercise test was positive or the ejection fraction was <50% in the absence of heart failure; in the presence of heart failure, the ejection fraction would need to be ≥50% to merit angiography.
The fact that the investigators needed to assemble such a broad and diverse set of data resources, including claims data, randomized trial results, and registry information, to address comprehensively the use of coronary angiography in acute myocardial infarction highlights the deficiencies of available clinical trials and provides a balance to the rigid advocacy of some people to changing practice only with definitive evidence from randomized trials. We simply do not have enough randomized trial information to conclude “no benefit” with any degree of certainty for most of the patients covered by this analysis. The conclusions of the analysis by Kuntz et al for the most part do not depart substantially from current mainstream clinical practice as reflected in large acute myocardial infarction registries such as the GUSTO I database, and they provide a rationale for performing angiography in a broad spectrum of patients after infarction.13
From a clinical perspective, however, it would be inappropriate to adhere rigidly to these guidelines. For example, a previously healthy 65-year-old man should not be treated differently from a previously healthy 64-year-old man. The different approaches recommended for men and women raise issues that will demand careful scrutiny of the model assumptions. Whether 65 years is the appropriate cutoff for recommending less aggressive treatment also requires detailed assessment of the strength of the information. Finally, the finding that in the presence of symptomatic heart failure, patients >65 years old who have left ventricular dysfunction do not gain enough life expectancy to merit the “up-front” costs depends on assumptions backed by very little data and may be contradicted by recent registry experience.13
The algorithm also does not directly address some important operational issues in clinical strategy. The detection of ischemia after myocardial infarction is critically dependent on the intensity of surveillance, and it is difficult to be certain about many ischemic episodes.14 15 An assumption appears to be made that a noninvasive assessment of left ventricular function will be a routine component of postinfarction care, which is consistent with emerging guidelines, although this strategy has not been adopted in many practices.16
We are hopeful that these findings will spark a healthy examination of variations in practice. If clinicians are using a different strategy, what assumptions are they making that differ from the present analysis? These findings also provide a valuable starting point for personnel in healthcare systems who want to develop guidelines for practice.
Similarly, efforts such as this should focus the analytic community on the issue of how to integrate the findings of cost-effectiveness analysis into medical practice. Given the extreme pressure placed on practitioners by constricting budgets, these analyses are no longer “ivory tower” academic exercises. Models based on diverse data sources and containing many pivotal assumptions will be used by health system administrators to develop practice policies. Clinicians need to be involved in and informed about these analytical solutions to medical controversies. The primacy of clinical judgment can no longer be counted on to prevail in every conflict between clinicians and those who would constrain clinical decision making in some way.
Because of rapid evolutions of medical technology and improvements in medical knowledge, many individual cost-effectiveness models will have limited lives and will need to be updated periodically. However, even if these models need to be altered radically to incorporate an emerging therapy, such as a new form of reperfusion therapy that significantly changes many of the probabilities in the model, this should be seen not as a deficiency of the techniques used but rather as an opportunity to examine carefully what is known about the effectiveness of the new therapy and how it should affect clinical and policy decisions. Creating a model, such as the one described by Kuntz and colleagues, forces us to be brutally honest about what we know versus what we believe but for which we have little evidence. Such critical self-examination is a vital starting point for improving the quality of medical care.
The opinions expressed in this editorial are not necessarily those of the editors or of the American Heart Association.
- Copyright © 1996 by American Heart Association
Charles ED, Rogers WJ, Reeder GS, Chesebro JH, Papapietro SE, Maske L, Bartolucci AA. Economic advantages of a conservative strategy for AMI management: rt-PA without obligatory PTCA. J Am Coll Cardiol. 1989;13:152A. Abstract.
Mark DB, Lam LC, Hlatky MA, Melton JR, Davidson-Ray L, Woodlief L, Lipscomb J, Sasahara A, Califf RM. Effects of three thrombolytic regimens and two interventional strategies on acute MI costs: results from a prospective randomized trial. Circulation. 1991;84:221. Abstract.
Mark DB, Naylor CD, Hlatky MA, Califf RM, Topol EJ, Granger CB, Knight JD, Nelson CL, Lee KL, Clapp-Channing NE, Sutherland W, Pilote L, Armstrong PW. Use of medical resources and quality of life after acute myocardial infarction in Canada versus the United States. N Engl J Med. 1994;331:1130-1135.
Peterson ED, Mark DB, Armstrong PW, Knight D, Granger CB. Does a conservative post-MI catheterization strategy miss severe coronary disease? Comparisons of use and diagnostic yield in Canada and the US. Circulation. 1995;92(suppl I):I-120. Abstract.
Kuntz KM, Tsevat J, Goldman L, Weinstein MC. Cost-effectiveness of routine coronary angiography after acute myocardial infarction. Circulation. 1996;94:957-965.
Sonnenberg FA, Roberts MS, Tsevat J, Wong JB, Barry M, Kent DL. Toward a peer review process for medical decision analysis models. Med Care. 1994;32:52-64.
Pilote L, Miller DP, Califf RM, Topol EJ, for the GUSTO I investigators. Classification and regression trees (CART) models to delineate use of angiography and revascularization after thrombolysis for acute myocardial infarction. J Am Coll Cardiol.. 1996;27:245A. Abstract.
O'Connor CM, Puma JA, Gardner LH, Califf RM, Jones RH. A 25-year experience in patients with coronary artery disease and chronic heart failure: outcomes with medical therapy and bypass surgery. J Am Coll Cardiol. 1996;27:142A. Abstract.
Langer A, Krucoff MW, Klootwijk P, Veldkamp R, Simoons ML, Granger CB, Califf RM, Armstrong PW, for the GUSTO Investigators. Noninvasive assessment of speed and stability of infarct-related artery reperfusion: results of the GUSTO ST-segment monitoring study. J Am Coll Cardiol. 1995;25:1552-1557.
Barbagelata A, Granger CB, Topol EJ, Worley SJ, Kereiakes DJ, George BS, Ohman EM, Leimberger JD, Mark DB, Califf RM, for the TAMI Study Group. Frequency, significance, and cost of recurrent ischemia after thrombolytic therapy for acute myocardial infarction. Am J Cardiol. 1995;76:1007-1013.
Peterson ED, Shaw LK, Califf RM. Risk stratification after myocardial infarction. Ann Intern Med. In press.