CHAPTER 112: ECONOMICS AND COST-EFFECTIVENESS IN CARDIOLOGY

In March 2010, the US Congress passed and President Obama signed a historic new health care legislation, the Patient Protection and Affordable Care Act (PPACA). The full impact of this complex bill on medical practice and US healthcare expenditures will likely not be clear for many years. The passage of this legislation was preceded by a vigorous public debate on the US healthcare system and options for its reform. Complex arguments about the best way to organize, deliver, and pay for healthcare became the stuff of the daily news reports and editorials. Major participants in the healthcare industry and other interest groups jostled to advance their perspectives before the US Congress and the court of public opinion. One major reason for all this attention and the justification given for both supporting and resisting passage of this legislation was a widespread concern about the current and future costs of healthcare and tremendous uncertainty about the best way to control these costs. According to the Centers for Medicare and Medicaid Services (CMS), the United States spent 17.5% of its gross domestic product (GDP) on healthcare in 2014, amounting to approximately $3 trillion.¹ Approximately 52% of this expenditure went for hospital care or physician/clinical services. Drug therapy accounted for another 10%. While the rate of growth in healthcare spending has varied in recent years, the trajectory for the national healthcare bill will continue to be upward for the foreseeable future. However, despite leading the world in healthcare spending, the United States has a system with which no one seems happy.²

Four major reasons can be offered to explain why healthcare spending is expected to keep increasing: (1) technological innovations that improve medical care³—the care of the cardiovascular patient in 2017 is vastly different (and largely better) than it was in 1980, but innovation in diagnosis and therapy generally increases costs relative to the prior patterns of care, as is reviewed in the second part of this chapter; (2) the aging of the population—an issue for all developed countries and one that has just begun to manifest itself (the vanguard baby boomers born in 1945 reached retirement age in 2010); (3) inflation, which refers to an increase in the cost of the same good or service (contrasted with an improved medical care service described in the first reason, above); and (4) raised expectations of the public regarding what medicine can and should offer them when they become ill, in part as a result of the global information revolution created by the Internet and search engines such as Google. Patients are no longer dependent on their doctor to tell them what is possible in a particular medical situation, and there appears to be no limit to the human desire to subdue morbidity and mortality technologically.

“Disruptive innovation” has been suggested as a guide to solving the ever rising healthcare cost dilemma. First proposed by Clayton Christensen in 1995, the concept was used to explain why certain complicated, expensive, but very mainstream products and services were driven to extinction by simpler, more affordable ones. Such innovations have clearly transformed modern life. It is unclear, however, that such innovations can be intentionally produced for the primary purpose of lowering healthcare costs as some have hoped.⁴ One key part of the disruptive innovation pattern was that leading companies and institutions are rarely the source of disruption; often, they are very late to recognize an impending disruption. With regard to healthcare, the continuing upward trajectory of costs is partly due to the “personal service” aspect of medicine, a feature that, together with certain regulatory barriers, resists attempts to make delivery of care more efficient.⁵

The cost of healthcare is not a unique concern for the United States. Other developed nations around the globe have all experienced unwelcome upward pressure on their medical care spending. Switzerland, France, Germany, Belgium, Austria, and Canada each spend approximately 10% to 11% of GDP on healthcare. Many of the remaining countries of the European Union spend between 8% and 10% of GDP on healthcare. The difference in spending between these countries and the United States appears primarily to be a result of a higher rate of increase in spending in the United States since 1980. Developing nations, although spending proportionately much less on healthcare at present, are projected to be facing an explosion of cardiovascular and other chronic diseases over the next half-century as they develop greater affluence and their populations adopt ever more atherogenic lifestyles.⁶ Clearly, healthcare costs will continue to be a major focus of attention at the policy level and a major pressure point in clinical practice for years to come.

This chapter has two overarching goals. The first is to explain briefly what the discipline of health or medical economics is (and is not) and what useful roles it might be able to play in debates about dealing with the ever-rising costs of the medical care system. To do this, we will need to examine the sorts of questions that health economics studies and the tools it uses. The second goal of the chapter is to review some of the recent and/or important health economics literature covering specific therapies and strategies used in the care of cardiovascular patients.

Economics is the discipline that uses formal models and statistical analysis to study how society manages and allocates its collective resources.^7,8 Medical economics, by extension, is the study of the production and allocation of healthcare resources.⁹ The standard model of (neoclassical) economics holds as axiomatic the proposition that society’s collective resources are finite and that there are many more possible uses for these resources than there are resources to be used. Consequently, choices need to be made about the best ways to use these limited (or scarce) resources to fulfill the needs and desires of society.¹⁰ The role of economics is thus conceptualized as serving to facilitate making these choices in an informed, rational manner. In the healthcare arena, for example, health economics offers tools to study the efficiency with which different healthcare goods and services can produce the product of interest, improved health. In the properly functioning free market envisioned in neoclassical economic models, buyers and sellers collectively determine the amount of a product that should be produced, with variations in price serving as the primary signal that connects these groups. Unfortunately, however, healthcare does not meet any of the standard criteria for a properly functioning free market, and thus prices cannot be relied upon to ensure that the “optimal” amount of healthcare services is being produced.

In a simplistic sense, economic analysis focuses on aspects of the efficiency of the healthcare system, whereas policy analysis is concerned with the directions it should take.¹¹ The two domains are connected by an ideological debate about the proper balance point between market-driven efficiencies and government management of healthcare. Those who believe that free markets, as conceived by neoclassical economic theory, are the best, most efficient way to allocate healthcare resources tend to favor a much more limited role for government in health policy. Deficiencies in the current healthcare system, according to this perspective, reflect too much, rather than too little, government intervention.¹² Economic analysis of the sort presented in this chapter serve (at best) a relatively minor role for advocates of the free market approach to healthcare. Conversely, those who feel that healthcare markets do not function as theory predicts (or who hold that the theory has major flaws) are much more supportive of government intervention at the policy level. In the context of “market failure” in healthcare, economists argue that the analysis methods (reviewed in this chapter) provide a guide, a “map” to inform rational policy-level decisions about how and where to allocate resources (spend money) and to what ends.¹³ The United Kingdom’s National Health Service and National Institute for Health and Care Excellence provide a well-developed example of how economic analysis methods can be used to inform and guide national health policy. In the United States, large payers, large provider groups, professional societies, the pharmaceutical and device industries, and the government interact in very complex ways to affect the levels of production and the patterns of distribution of healthcare. Thus, the role of economic analyses is less well defined in the United States, and its influence on policy is much more variable.

One other assumption that helps explain the worldview of health economic practitioners is that decision makers are “rational utility maximizers.” Utility is a concept borrowed from philosophy that refers to the amount of pleasure or satisfaction that is provided by something.¹⁴ In health economics, utilities provide a quantitative realization of the relative satisfaction felt by patients in various health states. In addition, health economics assumes that if one has two or more choices for treatment, one will always pick the one with the greater utility value. Because, in order to be satisfied, one must be alive, utilities and survival get combined to form the well-known but poorly understood metric of quality-adjusted life-years (QALYs). To give a concrete example, if treatments A and B both cost $1000, but treatment A yields on average 0.1 QALYs while treatment B yields 0.3 QALYs, health economists would always expect that B would be chosen over A. In order for this to be the case, however, decision makers would need to be indifferent to who gains from these choices and who loses, and such is rarely, if ever, the case. The “real world” is rarely as orderly, rational, and value neutral as it is portrayed in the economics literature.¹⁵ So health economics can perhaps be best thought of as providing simplified models of the healthcare economy, with the expectation that these models can provide some insights useful for decision-making, but are never the complete answer to any decision problem.

Relationship Between Health Economics and Comparative Effectiveness

Comparative effectiveness refers to research that seeks to answer important clinical questions related to real diagnostic and therapeutic options faced by significant numbers of patients.¹⁶ In short, it has the ambitious goal of developing the evidence needed to practice medicine. This is to be contrasted, perhaps, with evidence that is developed to seek regulatory approval for marketing of new products from the US Food and Drug Administration (FDA), where the comparator therapy in a pivotal trial might be “placebo.” It can also be contrasted with evidence developed in very focused, optimized clinical trials that seek to demonstrate a proof of concept.

Some have proposed that comparative effectiveness research will serve as a counterweight to healthcare policy decision-making that seeks primarily to maximize efficiency/minimize costs.¹⁷ The 2010 PPACA established the Patient-Centered Outcomes Research Institute (PCORI) as a nonprofit organization independent of the federal government. The institute is advisory in that none of its findings about comparative effectiveness bind the Secretary of Health and Human Services to cover or not cover specific types of medical care. Further, this new institute is specifically prohibited from using cost-effectiveness as a threshold to establish what types of care are recommended or reimbursed. The role of cost and economic considerations in comparative effectiveness research varies according to the source consulted. In the case of PCORI, high sensitivities about the possibility of the federal government rationing care as a way to save money resulted in a de-emphasis of the role of economics. However, there is not really any inherent tension between comparative effectiveness and high-quality health economics work, because both are built on the foundations of clinical comparative effectiveness questions. Thus, it may be useful to consider health economics research a comprehensive form of comparative effectiveness research that adds an extra dimension, that of efficiency.

The Three Big Health Economics Questions

Although it is natural to expect that the most important questions for a health economics analysis would deal with some aspect of cost, understanding the clinical effectiveness is actually almost always much more important.¹⁸ The first question of importance regarding a particular healthcare service or technology is: What long-term effects (both positive and negative) does it have on health outcomes, and what is the level of evidence supporting the current state of understanding about clinical effectiveness? Without a clear answer to this question, no economic analysis can be of much use. Sometimes the most important service an economic analysis can perform is to focus a bright light on the fact that the answer to this seemingly simple question is not at all clear. It is not uncommon to encounter a model-based economic analysis in the published literature where the authors have reached some conclusion about whether or not a given medical intervention is cost effective, yet the clinical literature is not at all settled even on what health effects that intervention produces.

The second critical health economics question is: Is it good value for the money? By good value, we basically mean that a given healthcare service or product can be used to produce a relatively large amount of “health” for relatively little money (Fig. 112–1). Another way to phrase this question is: How efficiently does this new therapy or strategy we are studying produce extra units of health (sometimes expressed as QALYs)? Answering this question requires that we combine our best understanding of the long-term clinical effectiveness with our best understanding of the long-term resource use/cost effects of the intervention in question. To do this, one of three possible economic efficiency analysis methods is often used (Fig. 112–2). These methods differ primarily in the way in which the health benefits are expressed in the analyses. Cost-benefit analysis requires all incremental health benefits to be converted to their monetary equivalent. As might be imagined, the methods for doing this are controversial because they require answers to some of society’s most difficult value questions.^11,19 Cost-effectiveness analysis avoids this hornet’s nest of difficulty by valuing all units of survival the same regardless of who benefits. Cost-utility analysis modifies this by using utility values (patient or societal preference weights) to adjust the survival data. Thus, survival in a severely disabled state or in severe pain would be valued significantly less (lower utilities and thus lower calculated QALYs) than survival in good or excellent health. Although the notion of utility might seem like a good solution to the difficult problem of valuing on a common scale all the types of health outcomes produced by the healthcare system, that solution hides considerable difficulties.²⁰ For example, who is best equipped to decide the utility weight of a year with a severe stroke or a year with severe heart failure? Should it be the member of the general public whose primary encounters with the medical system have been in the form of annual physicals? How much information about the health states being ranked do such individuals need in order to make informed judgments? Should it be healthcare professionals who are most familiar with the outcomes of these patients? Should it be the patients themselves? Two specific difficulties present themselves in the last option. First, some patients may be medically unable to participate in any assessment process. Second, no matter how severe or unpleasant the health state may seem to an outsider, humans have a remarkable ability to adapt and to accommodate to their disabilities and limitations (sometimes referred to as hedonic adaptation), and so their own valuation of their health state may be considerably higher than that given by a member of the general public.²¹

Cost-effectiveness analysis provides an estimate of the relationship between money spent and health benefits produced, but unlike cost-benefit analysis, the cost-effectiveness and cost-utility ratios have no natural interpretation. Thus, benchmarks are needed to help interpret the results of these analyses. Typically in the United States, a cost-effectiveness ratio of $50,000 or less per life-year (or QALY) saved has been considered “economically attractive,” whereas values of $100,000 or more per life-year (or QALY) saved have been considered “economically unattractive.” The $50,000 figure was apparently initially chosen because it was the cost to keep an end-stage renal failure patient alive on hemodialysis for a year relative to no dialysis. Dialysis is a form of medical care that Congress had guaranteed to fund for all affected renal failure patients and thus represents an explicit policy statement by the US federal government (representing “society”) of how much it is willing to pay to save a life-year. One criticism of this benchmark is that it has never been updated. A recent evaluation of 13 studies published between 1968 and 1998 on the cost of renal replacement therapy found that the cost of center dialysis was between $55,000 and $80,000 per life-year saved.²² A more recent analysis of this issue using a computer simulation model and data from 1996 to 2003 found that renal dialysis had a cost per QALY of $129,000.²³ Recent proposals to increase the benchmark to $150,000 per QALY have been variably adopted in published research.²⁴ Other countries use different, more conservative benchmarks. For example, the National Institute for Health and Care Excellence in the United Kingdom uses a benchmark of less than or equal to £20,000 per QALY to define economic attractiveness, although this is applied with some flexibility.^25,26

The third major question of health economics is: Can we afford it? This question really represents the intersection of policy and health economics. Public policy is concerned with the large decisions about how much of society’s resources to devote to producing more healthcare. In the United States, this is defined both by the amount of funding provided to Medicare (primary public funder of healthcare) by Congress and the healthcare investment and reimbursement policies in the private sector. Economics can then address the question of what the aggregate cost to society or to individual payers will be from funding different diagnostic and management strategies. In Canada and much of Europe, where healthcare spending levels are mostly controlled by central governments, the effect of new therapies and tests on the healthcare budget is always an issue, and reimbursement policies are carefully crafted to control dissemination. In the United States, Medicare is prohibited from explicitly considering cost in coverage decisions and the lack of a single central payer means that funding decisions get made in a more fragmentary and disjointed way. However, even with central decision making, funding decisions are not necessarily made to optimize health production. One major obstacle is that it is much easier in such a system to say no to some new innovation than to stop or reduce funds for some existing therapy because it is not evidence based or because it is now a less efficient method of producing more population health.

Data for Health Economics

As with any form of empirical research, the value of economic analysis is highly dependent on the quality and relevance of both the clinical and economic data available for the work. Ideally, data at the patient level can be obtained on both clinical outcomes and resource consumptions/costs from the same cohort. However, even this condition does not ensure the best quality. For example, there are many pivotal guideline-changing clinical trials that provide excellent clinical outcomes data, but have almost no useful resource use or cost information. Estimating costs from the few possibly relevant variables in such a data set may be the best that can be done, but the limitations of this should not be overlooked. Alternatively, a small clinical trial may provide detailed economic data, but its clinical efficacy data are not reliable as a result of the sample size.

When we refer to “economic data,” generally we mean resource use information and associated costs. Specifically, economists wish to assess the incremental resources needed to produce a good or service, by which they mean only the extra resources required. If we are comparing a new treatment with standard care, for example, and both of them will involve 5 days in the hospital at the same level of intensity of care from doctors and nurses, then there is no incremental hospital stay involved in the new treatment. Incremental costs then are the monetary value of the incremental resources consumed. The primary emphasis on resource consumption rather than money is deliberate. Money is a tool that facilitates the functioning of markets, so that barter is not required. The meaning of money is always traced back to the resources consumed. Economists refer to opportunity cost to designate what was given up or lost to society when a decision was made to produce something else: the value of (relevant) choices not made.⁸ If we decide to produce 10,000 more coronary bypass surgeries in the United States, for example, we will lose the resources consumed in that process to alternative uses. In the world of limited or scarce resources and rational utility-maximizing individuals referred to earlier, this notion of trade-off underlies the economists’ fundamental concept of cost.

Neoclassical economics further maintains that in a well-functioning free market, the equilibrium price of the good or service in question represents its opportunity cost. Because, as noted earlier, the medical marketplace violates a number of key conditions required to qualify as a well-functioning free market, the economic meaning of medical prices is problematic. Costs that reflect the production process of the medical care are a reasonable surrogate for economic value. However, the (uncorrected) charges that are contained on US medical bills are completely unreliable for economic analysis purposes,²⁷ and the reimbursements paid to hospitals and physicians are typically less suitable for this use outside of very specific questions relevant to payers or providers. Methods for estimating medical costs are reviewed elsewhere.^28,29

Induced costs/savings are the costs and cost savings that occur consequent to some treatment or strategy used in patient care. For example, use of thrombolytic therapy for ST-segment elevation acute myocardial infarction (MI) would entail the initial cost of the drug (for tenecteplase, > $2000). If the patient in question had an intracranial hemorrhage, the extra costs of caring for that complication would be induced costs, and clearly, those costs would most likely extend for the rest of that patient’s lifetime. If, instead, treating the patient during the first hour of her large anterior MI prevented her from developing debilitating heart failure later, then all the money that would not have to be spent over the ensuing years of her life caring for that complication would need to be counted as an induced cost savings. In an analysis that extends many years into the future, future costs and savings need to be discounted (typically at 3% per year) to calculate the present value of the lifetime cost stream. The reason is straightforward. We are not indifferent to the prospect of receiving $100 today or receiving the same amount in 10 years. We recognize that in 10 years that money will be worth less to us than it is today, because we could have had it in the bank earning interest for 10 years if we had received it today. Thus, discounting is a way of reducing the nominal value of future spending and savings to their present value equivalents.

A further complexity about the concept of cost is that a perspective (sometimes termed a viewpoint) must always be defined in order to measure and understand costs.⁹ For example, if a patient undergoes a successful percutaneous coronary intervention (PCI) with a drug-eluting stent (DES) and is discharged from the hospital without complications, but is readmitted 3 months later with a stent thrombosis, that readmission is a cost from the payer’s perspective and from the societal perspective, but not from the hospital’s perspective or the physician’s perspective unless they are operating in a capitated system. If a patient comes to clinic with chest pain and gets a clinical workup and a stress test, the cost of all that from the patient’s perspective (assuming they are insured) may simply be his co-pay. The clinic and society, however, each have different perspectives on that same episode of care. In most economic analyses, the primary analyses should be done using the societal perspective, which is the perspective in which the winner/loser problem described in the above examples is minimized.

The problem of measuring and comparing costs across international borders is a difficult one that offers no easy solution, but is increasingly important in the current era of international cardiology mega-trials.^30,31 To the extent that practice patterns are converging around common standards that cross international boarders, it may make sense to pool resource use data across countries and assign common costs. However, local variations in price data make it problematic to decide what prices to use in such an exercise. Further, there is still good evidence that local practice patterns differ substantially in different regions of the world. For example, length of stay after acute MI and the rate of referral for invasive angiography vary considerably across countries.³²

Models in Health Economics

Several different types of models are now commonly used in health economics. Descriptive and predictive statistical cost models are mathematical models that treat cost as a probabilistic quantity. Although that might sound quite technical and arcane, these models are quite analogous to descriptive and predictive regression models of clinical outcome variables. Such models might be developed to understand the major clinical correlates of cost in a particular cohort or to be able to predict the cost of future patients with particular clinical features treated in a specific way. To develop these sorts of models, one needs patient-level data—ideally for both clinical and cost descriptors of interest. More complex models based on regression methods are also now commonly used in health economics. For example, multilevel (sometimes called hierarchical) models can be used to explore patient-level, provider-level, institutional-level, and even country-level influences on cost and outcomes. High-quality patient-level data and regression modeling techniques can be used to estimate both lifetime costs and lifetime clinical outcomes in a cost-effectiveness analysis.

These patient-level models can be contrasted with deterministic models used in health economics, including decision models. These latter models have been most often used to estimate lifetime costs and outcomes for cost-effectiveness analysis. Typically, they do not use patient-level data, but instead employ group-level summary data obtained from various (unrelated) published sources. Their strength lies in providing a framework for combining such disparate data estimates into a summary model focused on a specific question. Because these data do not incorporate any measure of the variability in the estimates, the resulting models are often interpreted as showing what will happen (deterministic) rather than what might happen (probabilistic or stochastic). The simplest decision models employ basic probability calculations to estimate the (uncertain) outcomes from each strategy or therapy being studied. More complex methods of handling the uncertainty in such models includes the use of Markov models (also called state transition models), where the probability of moving from one health outcome state to another in each time cycle of the model (eg, 1 month or 1 year) is represented as an explicitly specified transition probability. The limitations to incorporating more of the nuances of clinical practice in such models are that the models become quite complex to manipulate and to explain, and also that the data sources available to populate the model may not be sufficient to support the complexity the analyst desires to examine.

One major reason that so much modeling is done in health economics is that clinical studies, even the best ones, rarely provide all the data that are needed for a comprehensive health economics analysis. The time frame relevant for a health economics analysis is the full time span during which some residual effect of the testing or treatment strategy being studied might be expected. For most clinical problems involving patients with a chronic disease such as the ones discussed in the second part of this chapter, the lifetime of the patient cohort is the appropriate time frame of interest. If the full stream of costs and outcomes over this time frame could be observed and measured, no modeling-based extrapolation would be required. Because for adult chronic diseases this is almost never possible, some amount of model-based extrapolation is required. The question of how best to do that has no simple answer, and different authorities in the field favor different approaches. Since the models are quite difficult to validate empirically, the matter of which approach is to be preferred has not been resolved.

Given this state of affairs, the consumers of these models can best arm themselves for forays into this literature by keeping in mind a few basic questions.³³ First, are the assumptions clear? All models require assumptions, and in some cases, the assumptions are quite important in deciding how much credence to give the results. Second, are the important cost and outcome inputs clearly specified? Third, are the major moving parts of the analysis sufficiently transparent for readers to understand the final results? The model should not be presented simply as a black box where data go in one end and an “answer” emerges from the other. Health economics is not discovery science. A cost-effectiveness model, therefore, is not a method of uncovering new medical insights, but rather should provide a clearer understanding between money spent and health outcomes produced for that money. Whenever possible, intermediate outputs from the model should be validated against empirical data. Finally, are the key uncertainties in the analysis fully explored via stochastic and sensitivity analysis methods?

Patterns of Relations Between Costs and Outcomes

The primary concern of economics is choosing among alternatives. Thus, economic efficiency metrics, such as a cost-effectiveness ratio, must be expressed in terms that are common to all the alternatives of the choices to be made. When analysts want to be able to compare medical investment choices with choices in other sectors of the economy, such as education, transportation, and defense, it is often necessary to use cost-benefit analysis because there is no other common method of expressing the outcome being produced by the investment being considered. However, when all choices reside in the healthcare arena, cost-effectiveness and cost-utility analyses are much more commonly used, with the most common effectiveness metric being the added life-year or the QALY. In such circumstances, from a cost-effectiveness analysis perspective, all medical outcomes must be translated into a composite metric reflecting their effect on length and quality of life. When dealing with an intermediate outcome, such as a change in a biomarker or a change in the degree of stenosis of a coronary artery, this requirement may seem overly burdensome. However, using therapies that have both risk and cost to alter biomarkers or coronary artery appearances without some reasonable expectation that the patient will benefit over the long term either in terms of length or quality of life is hard to justify on a number of different grounds. And if there is a belief that these outcomes are benefited, but the amount is uncertain because of a lack of appropriate clinical data, then that uncertainty is an important part of the story about the therapy or management strategy being studied.

The pattern by which quantity and quality of life are altered by therapies can vary in several different ways. For example, all the incremental outcomes differences may accrue early (eg, by 30 days), they may accrue steadily in a linear fashion over time, or they may even amplify over time. The pattern of cost differences over time also can vary in different circumstances and does not have to mirror the pattern of incremental outcome benefits. Cost differences may be large or small and may accrue early, steadily over time, or as a step function. Surgical therapies such as coronary artery bypass graft surgery (CABG) and heart transplant have very high upfront costs, and the initial clinical benefits may even be negative as a result of operative mortality. Outcome benefits may take years to be fully expressed in such cases so that the incremental outcomes calculated from short-term follow-up (eg, 1 year) are quite likely to underestimate such benefits. Secondary prevention, on the other hand, often involves a steady cost stream year after year and, in some cases, a small clinical benefit that tends to amplify over time. The key point that these examples serve to highlight is that a high-quality economic analysis must assess outcomes and costs over the long term, defined operationally as the period for which the effects of the therapy in question, both direct and induced, might reasonably be expected to persist.

Some Caveats About Health Economics

A few points that have been made earlier in this chapter are worth emphasizing because they are so central to understanding the perspective of health economics. First, the focus is typically on systems, countries, or society and not on individual doctors or patients. Health economics, then, does not usually focus on decision-making at the level of the individual patient, nor is it intended that individual physicians should be using these methods to make individual “funding decisions” about what care to provide to patients.

Second, health economics is neutral on the issue of who benefits and who is deprived of healthcare services. If it were possible to generate more QALYs by treating an 80-year-old patient with advanced heart failure than a 30-year-old with hypertension or a premature infant with respiratory distress, and only one could be chosen, health economics would favor going the way of more efficiency. Clinicians usually do not face such overt choices, although triage situations do come close, as when coronary care unit beds or donor hearts must be allocated and there are more candidates than resources to satisfy all who could benefit. Health economics in this context is a tool that may clarify some important issues about the choices being considered. However, societal values and preferences may and often should override the maximal efficiency answer.

Finally, because clinicians are now accustomed to using fairly rigorous standards to decide whether new therapies and clinical strategies “work” and practice medicine using guidelines that rank recommendations according to the strength of the evidence, they may not appreciate that policy decisions are often made on the basis of “evidence” that would not pass rudimentary quality screens in medical/outcomes research. Thus, clinicians who are informed, discerning consumers of health economics research will be critical to help drive comparative effectiveness research toward higher quality rather than political expediency.

Heart Failure

(See also Chaps. 68, 69, 70, 71, 72, 73.) According to statistics compiled by the American Heart Association, approximately 5.7 million people in the United States have heart failure (the prevalence), and every year, 915,000 new cases are diagnosed (the incidence).³⁴ In the Medicare population (age ≥ 65 years), heart failure is the most common reason for hospitalization listed on medical claims data. In the total US population, the estimated number of annual heart failure hospitalizations is approximately 1 million a year. Further, patients who are hospitalized for this diagnosis have a substantially elevated risk of being rehospitalized (~25% within 30 days [approximately one-third for heart failure, the remainder for other medical problems] and ≥ 50% within the following 6 months [about 50% related to heart failure]).^35,36 The overall annual US medical spending attributed to heart failure in 2012 was approximately $30.7 billion, with 68% attributable to direct medical costs.³⁴ It must be kept in mind that given the difficulties in identifying heart failure on a population level, these estimates are extremely crude. However, the data are sufficient to conclude that heart failure is common and accounts for a large amount of healthcare spending. It follows to reason that if we could allocate resources to prevent the disease or, failing that, prevent it from progressing to decompensation, we might improve the health of these patients efficiently and without a massive increase in healthcare spending. Heart failure prevention means better control of the disorders most likely to progress to heart failure, primarily coronary artery disease (CAD) and hypertension (each confer a two- to three-fold increased risk of heart failure). However, because of the time spans involved (typically decades), the true population health benefits and efficiencies of primary prevention strategies are difficult to study empirically.³⁷ In the remainder of this section, we will examine three general types of care used to manage the established heart failure state: maintenance medical care to relieve symptoms, improve functional status, and improve prognosis; strategies to reverse the heart failure state and to prevent sudden cardiac death; and comprehensive management strategies to improve functioning and reduce acute decompensations.

Medical Therapy of Heart Failure With Reduced Ejection Fraction

Medical therapy of heart failure focuses on symptom relief and improvement of prognosis. As reviewed elsewhere in this text, the main agents shown to enhance survival are angiotensin-converting enzyme (ACE) inhibitors, β-blockers, angiotensin receptor blockers (ARBs), and sacubitril/valsartan, as well as isosorbide-hydralazine and aldosterone antagonists in selected populations. Unfortunately, much of the literature examining the economics of these therapies is now either dated or most pertinent to a non-US healthcare system. In addition, few of the pivotal trials that changed our clinical practice included careful prospective measurement of resource use and costs, so most of the literature is based on decision modeling frameworks.

In one such analysis, Glick and colleagues³⁸ examined the cost effectiveness of enalapril using data from the Studies of Left Ventricular Dysfunction (SOLVD) Treatment trial, which randomized 2569 symptomatic heart failure patients with left ventricular ejection fraction ≤ 0.35 to receive either enalapril or placebo. Over a projected 7-year life expectancy, these authors estimated that enalapril would add 0.3 discounted life-years (0.21 discounted QALYs) and would have a net lifetime incremental cost of $25, with a cost-effectiveness ratio of $80 per life-year added ($115 per QALY added) (Table 112–1). Given that generic enalapril (along with lisinopril and benazepril) can now be obtained in the United States and that the Glick analysis probably underestimated the long-term effects of therapy on reducing hospitalizations, it is very possible that enalapril as used in the SOLVD Treatment trial was cost saving over the long run.

Three β-blockers have been shown to improve survival in heart failure patients when added to ACE inhibitors: sustained release metoprolol, bisoprolol, and carvedilol. The clinical data show that these drugs also reduce all-cause hospitalization, although the magnitude of this effect varies among the trials. Carvedilol is now available in a generic form. The US Carvedilol Heart Failure Program, consisting of four trials, was used as the basis for a Markov decision model analysis, combining the trial data with varying assumptions about what would happen to benefits as the trial data were projected to a lifetime time horizon.³⁹ Life expectancy without carvedilol was estimated at 6.7 years, and carvedilol therapy was projected to add 0.31 to 0.95 life-years at an incremental lifetime cost (1999 costs) of $7600. Given the difference between proprietary drug costs used in that analysis and current generic costs, carvedilol would now be expected to have a cost-effectiveness ratio much more favorable than the $30,000 per life-year calculated in this older model. Bisoprolol is also available in generic form, but the trials supporting this drug were done in Europe, and no high-quality economic data are available for the United States. The Metoprolol Randomized Intervention Trial in Congestive Heart Failure (MERIT-HF) study group conducted an economic analysis of extended-release metoprolol succinate compared with standard therapy alone using a discrete event simulation, which assumes risks based on individual patient characteristics and then assigns health and economic consequences of associated events, management, and outcomes.⁴⁰ Costs were reported in 2001 US dollars, and the proprietary price of extended-release metoprolol was $912 over 2 years. The model predicted that metoprolol would prevent 7 deaths and 15 hospitalizations from heart failure per 100 patients over 2 years and result in a cost savings of $395 to $1112 per patient.

ARBs improve survival and reduce cardiovascular hospitalizations in heart failure patients unable to tolerate ACE inhibitors,^41,42 but their benefit as a first-line treatment in place of or as an adjunct to ACE inhibitors is less clear.^43,44,45 There are no prospective economic analyses from a US perspective of the major clinical trials comparing ARBs with placebo or with ACE inhibitors. An economic analysis from a European perspective of the Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity (CHARM) program found the use of candesartan to be economically efficient compared with placebo in heart failure patients with an ejection fraction ≤ 0.40, with candesartan achieving economic dominance (better clinical outcomes and lower costs) in some scenarios.⁴⁶ Several ARBs are now available generically.

The FDA approved the angiotensin-neprilysin inhibitor combination drug sacubitril/valsartan in 2015 for use in heart failure patients with reduced left ventricular ejection fraction. The pivotal trial supporting this approval, the Prospective Comparison of Angiotensin-Neprilysin Inhibitor With ACE Inhibitor to Determine Impact on Global Mortality and Morbidity in Heart Failure (PARADIGM-HF) trial, assessed 200 mg twice daily of sacubitril/valsartan versus 10 mg twice daily of enalapril in 8442 patients with New York Heart Association (NYHA) class II to IV heart failure and an ejection fraction of ≤ 40%.⁴⁷ The trial was stopped after a median follow-up of 27 months because of significant benefit with this combination therapy. The primary composite outcome of death from cardiovascular causes or hospitalization for heart failure occurred in 21.8% of the sacubitril/valsartan group versus 26.5% in the enalapril group (hazard ratio [HR], 0.80; 95% confidence interval [CI], 0.73-0.87; P < .001). The price of sacubitril/valsartan is approximately $12.00 to $13.00 per day (National Average Drug Acquisition Cost reference file, February 3, 2016). A Markov model–based analysis used published data from the PARADIGM-HF trial and other published data to assess the cost-effectiveness of sacubitril/valsartan compared with lisinopril.⁴⁸ This model predicted 6.78 years of survival and 5.56 QALYs together with total costs of $123,578 for the lisinopril arm. The corresponding projections for sacubitril/valsartan were 8.98 years of survival, 6.13 QALYs, and total costs of $152,716. The incremental cost-effectiveness ratio was $50,915 per QALY gained. A 5-year US budget model on the impact of this drug projected about 1.7 million patients taking sacubitril/valsartan for more than 1 year and annual costs per year of $3 billion.⁴⁸ Early evidence on prescriptions for the drug in the United States suggest this is a substantial overestimate.⁴⁹

The isosorbide-hydralazine combination studied in the African American Heart Failure Trial (A-HeFT) reduced heart failure hospitalizations during the study’s mean 13-month follow-up (from 0.47 per patient to 0.33) and reduced total medical costs during the study follow-up by approximately $4300, exclusive of drug costs.⁵⁰ Even with the assumed drug cost per day of $6, the isosorbide-hydralazine combination was economically dominant. Because the specific proprietary combination drug studied is no longer available, the cost of these two drugs would now be significantly less, reinforcing the likelihood that this treatment is very economically attractive for the population studied in A-HeFT.

Two aldosterone antagonists have been studied in selected populations with heart failure: spironolactone in the Randomized Aldactone Evaluation Study (RALES) and eplerenone in the Eplerenone Post-Acute Myocardial Infarction Heart Failure Efficacy and Survival Study (EPHESUS). A decision model–based analysis looking at the first 35 months of the RALES trial found that spironolactone added 0.13 QALYs and had a net lower cost than the placebo arm.⁵¹ In a lifetime analysis of the EPHESUS trial results, eplerenone added approximately 0.10 to 0.13 life-years and had a net cost of approximately $1400 as a result of the drug cost (in this analysis, ~$3.60 per day).⁵² The cost-effectiveness ratio, using several different data sets upon which to base the long-term extrapolations, was consistently under $20,000 per life-year added.

One interesting analysis addressed the economic impact of underuse of effective therapies by modeling the effect of increasing the use of ACE inhibitors, β-blockers, spironolactone, and digoxin each by 10% in a Canadian cohort of 86,000 patients with heart failure.⁵³ After accounting for the extra cost of the drug therapy ($4.3 million in Canadian prices), this strategy was estimated to save $2.3 million net in the first year as a result of reduced avoidable hospitalizations.

Heart Transplant and Left Ventricular Assist Device Therapy

(See also Chaps. 72 and 73.) For patients with advanced refractory heart failure, cardiac replacement therapy may offer the best option to restore quality of life and improve longevity. The economic impact of orthotopic heart transplant is strongly influenced by the fact that the supply of donor hearts for transplant has remained limited to approximately 2000 per year, with no signs this number is likely to increase. A review using data from the Nationwide Inpatient Sample between 2005 and 2009 found that the cost (calculated from charges using overall hospital cost-to-charge ratios) of the index procedure hospitalization for transplant increased 40% from $120,000 to $168,576.⁵⁴ Total annual costs to the US healthcare system for procedure admissions were over $375 million in 2009, representing the costs of 2227 heart transplants.⁵⁴ In one academic center, transplant patients averaged 2.1 hospitalizations and 11.9 outpatient encounters during the first posttransplant year.⁵⁵ Cumulative costs for the first year after transplant at another center averaged about $330,000 (2010 US dollars), with the index admission representing > 50% of that amount.⁵⁶ Contemporary annual posttransplant costs after the first year can total more than $70,000 per patient in part as a result of the need for expensive immunosuppressive drugs and the need for complex hospitalizations to treat rejection episodes and infections. Despite its high procedural and follow-up costs, the substantial increase in life expectancy (one of the largest of any cardiovascular therapy) and improvement in quality of life for transplant patients make it likely to be economically attractive compared with medical therapy alone in this population.

Because left ventricular assist devices (LVADs) are not limited in number the way donor hearts currently are, the potential exists for this form of cardiac replacement therapy to have a much larger clinical and economic impact. The technology and ancillary care have evolved significantly since the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) trial. In 2009, the unadjusted mean per-patient cost of the index hospitalization for LVAD transplant was $208,522.⁵⁴ Cumulative first-year costs in a small cohort at a different center averaged about $370,000.⁵⁶ With 2298 LVAD implants nationwide, the total cost to the US healthcare system was estimated at $479 million.⁵⁴ Although LVAD therapy is quite expensive initially, the therapy has the potential to be cost effective if, like heart transplant, it provides a significant enhancement of survival and/or quality of life. In 2038 patients receiving LVADs, hospital survival improved from 2008 (53%) to 2011 (87%).⁵⁷ Hospital costs increased during this period, reaching $235,000 between 2008 and 2011. Of those discharged alive with a device, during an average follow-up of 11 months, patients required an average of 2.2 readmissions.⁵⁸ Mean time to first readmission was 35 days, and 44% were readmitted within 30 days. Most common causes of readmission were gastrointestinal bleeding, driveline infection, and stroke. A model-based analysis examined the cost-effectiveness of continuous-flow LVADs for destination therapy versus optimal medical management in advance heart failure patients.⁵⁹ Continuous-flow LVAD patients had higher 5-year costs compared with medically managed patients ($360,407 vs $62,856), higher quality-adjusted life-years (1.87 vs 0.37), and more life-years (2.42 vs 0.64). The incremental cost-effectiveness ratio for the continuous-flow LVAD was $198,184 per QALY and $167,208 per life-year. A second model-based analysis came to a similar result: compared with inotrope-dependent medical therapy (life expectancy, 9.4 months), destination LVAD (life expectancy, 4.4 years) had a cost-effectiveness ratio of $202,000 per QALY.⁶⁰

Cardiac Resynchronization Therapy

Another quite different approach to the treatment of patients with advanced heart failure derives from observations that some patients have severe left ventricular dysfunction accompanied by a dyssynchronous pattern of depolarization and contraction usually as a result of left bundle branch block. In such patients, the use of biventricular pacing for cardiac resynchronization has been shown to improve heart failure symptoms (by approximately one NYHA class) and to improve survival. An economic analysis of the Cardiac Resynchronization in Heart Failure (CARE-HF) trial done using United Kingdom (UK) cost weights found a cost-effectiveness ratio for cardiac resynchronization therapy (CRT) plus medical therapy of €19,319 per QALY and €43,596 per life-year compared with medical therapy alone, indicating the important contribution of quality-of-life improvements to the estimated cost-effectiveness in this trial.⁶¹

In Europe, many CRT devices are implanted without implantable cardioverter-defibrillator (ICD) function (CRT-P), whereas in the United States, most of the CRT devices include an ICD function (CRT-D), a factor that influences the cost of the device. The incremental contribution of the ICD function versus the CRT function in these patients to the survival benefits seen remains unsettled. An economic model based on the US Comparison of Medical Therapy, Pacing, and Defibrillation in Chronic Heart Failure (COMPANION) trial data found that with extrapolation to 7 years, CRT-P had a cost-effectiveness ratio of $19,600 per QALY relative to medical therapy alone, and CRT-D had a cost-effectiveness ratio of $43,000 per QALY relative to medical therapy alone.⁶² What was left unaddressed by the COMPANION trial is the incremental benefits and cost-effectiveness of CRT-D relative to CRT-P.

The Multicenter Automatic Defibrillator Implantation Trial (MADIT)-CRT trial examined CRT-D versus ICD alone in 1820 NYHA class I or II heart failure patients (both ischemic and nonischemic) with an ejection fraction ≤ 0.30 and a QRS duration of 130 ms or more and found that the addition of CRT to ICD therapy reduced heart failure events by 41% (22.8% for ICD-only group vs 13.9% for ICD plus CRT; P < .001).⁶³ No effect was seen on all-cause mortality. A prospective cost study was conducted alongside the main MADIT-CRT trial in a total of 1271 patients.⁶⁴ Healthcare costs at 4 years after randomization were higher in the CRT-ICD patients, largely as a result of device and implantation costs. CRT-ICD had an incremental cost-effectiveness ratio of $58,330 per QALY compared to ICD alone. A Markov decision model of this same treatment choice using the published trial results estimated that CRT-D increased life expectancy (9.8 years vs 8.8 years for ICD alone), QALYs (8.6 years vs 7.6 years), and costs ($286,500 vs $228,600).⁶⁵ The base case cost-effectiveness ratio was $61,700 per QALY.

Prevention of Sudden Cardiac Death With Implantable Cardioverter-Defibrillator Therapy

(See also Chap. 91.) The MADIT-II trial and the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) established that ICD therapy improves survival over medical therapy alone in patients with reduced left ventricular ejection fraction.^66,67 A prospective economic analysis of the MADIT-II trial found ICD therapy in patients with reduced left ventricular ejection fraction after MI to have an estimated incremental cost-effectiveness ratio of $235,000 per life-year saved based on an average survival gain of 0.167 years after 3.5 years of follow-up.⁶⁸ However, this analysis made the overly conservative assumption for the base case analysis that no further survival benefits would accrue from ICD therapy beyond the empirical follow-up period of the trial. Empirical 10-year follow-up from MADIT-II subsequently showed that survival benefits from ICD therapy continued to amplify over time, indicating that the initial published economic analysis from this study was too conservative. A separate economic model based on MADIT-II eligibility criteria found that the estimated cost-effectiveness ratio in this patient population varied depending on follow-up time horizon, from $367,200 at 3 years to $67,800 at 15 years.⁶⁹

A prospective economic analysis of the SCD-HeFT trial using a lifetime time horizon found a more favorable cost-effectiveness ratio for ICD therapy compared with medical therapy alone in stable, moderately symptomatic heart failure patients with a left ventricular ejection fraction ≤ 0.35. In this economic analysis, the base case lifetime cost-effectiveness ratio was $38,389 per life-year saved and $41,530 per quality-adjusted life-year saved.⁷⁰ Similar to the aforementioned economic analyses, the SCD-HeFT investigators estimated the cost-effectiveness ratio would improve with extended follow-up, with $127,503 per life-year saved at 5 years and $58,510 per life-year saved at 12 years. However, in subgroup analyses, ICD therapy in patients with NYHA class III heart failure was found to increase costs without an incremental benefit and thus was dominated by medical therapy alone.

A Markov model using data from several clinical trials of ICD therapy was constructed to estimate the cost-effectiveness of ICD therapy under different assumptions.⁷¹ This analysis found that prophylactic ICD therapy added between 1.01 and 2.99 QALYs at a cost of $68,300 to $101,500 in patients with left ventricular systolic dysfunction. The incremental cost-effectiveness of ICD therapy in this analysis ranged from $34,000 to $70,200 per QALY, with ICD therapy remaining below the upper cost-effectiveness benchmark of $100,000 per QALY if the incremental survival benefit of ICD therapy lasted ≥ 7 years.⁷¹

Transcatheter Aortic Valve Replacement

The Placement of Aortic Transcatheter Valves (PARTNER) trials (see Chap. 47) examined the use of transcatheter aortic valve replacement (TAVR) in patients with and without operable aortic stenosis (AS). The PARTNER B cohort consisted of very symptomatic (93% NYHA class III or IV) elderly patients with inoperable AS. TAVR increased life expectancy from 1.2 to 3.1 years, and quality of life was substantially improved, with between 25% and 40% of patients benefitting dramatically from the procedure by gaining years of high-quality survival.⁷² The cost of the TAVR procedure was estimated at $43,000 and included the cost of the device, which was about $35,400.⁷³ Hospitalization for the procedure averaged $79,000. After accounting for resource use in the year following the procedure, total costs at 1 year for the TAVR patients averaged $52,000 more than the conservatively treated patients. The incremental cost-effectiveness ratio was $50,200 per life-year added and $61,889 per QALY. The PARTNER A cohort consisted of patients who were eligible for surgical aortic valve replacement.⁷⁴ Mortality in the TAVR and surgical aortic valve replacement groups was similar out to 3 years (~50%), but the TAVR group had 2.1 per 100 more major strokes and more early major vascular complications, and the surgical aortic valve replacement group experienced higher rates of early major bleeding. Index costs were $72,000 for the TAVR group and $74,400 for the surgical aortic valve replacement group.⁷⁵ Periprocedural complications accounted for approximately $12,475 per patient in the TAVR group and 2.4 days of hospitalization.⁷⁶ One-year cumulative costs for each group were around $100,000. These two treatment strategies were essentially comparable in terms of outcomes and costs.⁷⁷

Atrial Fibrillation

(See also Chap. 83.) Atrial fibrillation (AF) affects between 2.7 and 6.1 million Americans, and approximately 1.6 million new cases are diagnosed each year.³⁴ AF is associated with a shorter life expectancy, and it adversely impacts patient quality of life, increases the risk of cardiovascular events, complicates heart failure, and significantly increases the likelihood and severity of stroke, all of which are associated with substantial medical costs.

Because of its co-occurrence with common cardiovascular diseases such as CAD and heart failure, determining the unique economic burden of AF is difficult. In 2010, AF was the primary discharge diagnosis for approximately 479,000 hospitalizations in the United States. Kim et al⁷⁸ performed a retrospective, observational cohort study using administrative claims data and estimated the total annual incremental cost to be $8705 per AF patient which extrapolated to between $6.0 and $26.0 billion for the US healthcare system.

Analysis of a large claims database with charges converted to costs found the cost for a hospital admission with AF as a primary or secondary diagnosis was $15,000 to $16,800.⁷⁹ The costs associated with complications of AF have recently been reported from two registries. In Ontario, Canada, average costs, in Canadian dollars, were $19,937 for an ischemic stroke event, $22,347 for intracranial hemorrhage, and $12,313 for a minor stroke/transient ischemic attack (TIA).⁸⁰ In a US claims database, hospital reimbursements by insurance were $20,933 for an ischemic stroke, $59,034 for a hemorrhagic stroke, and $8616 for a TIA.⁸¹ The cost of a major bleeding episode requiring hospitalization in a fourth claims registry was $16,830.⁸²

Oral Anticoagulation

Warfarin therapy has been found to reduce the risk of stroke in AF patients by 62% (compared with a 22% risk reduction with aspirin) and is recommended in current practice guidelines. Unfortunately, the efficacy of warfarin therapy in preventing strokes is tempered in practice by its variable anticoagulation effects and associated risk of bleeding. For an optimal risk-benefit ratio with warfarin therapy, frequent monitoring must be performed to ensure that the international normalized ratio (INR) is maintained within the target therapeutic range. Such monitoring is particularly important in patients at high risk for bleeds (eg, elderly individuals). However, in clinical practice, regular monitoring poses logistical difficulties and increases the cost of care, thereby increasing the likelihood of poorly controlled anticoagulation. Despite these difficulties, warfarin is highly effective, and its generic price for maintenance therapy is less than $1 per day in the United States and pennies a day in European countries, which makes it very efficient from an economic perspective.

Novel oral anticoagulants (NOACs) that target either thrombin or factor Xa have emerged as an alternative to vitamin K antagonists (Table 112–2). Compared to warfarin, the NOACs provide a better pharmacologic profile including a more rapid onset of action with a more predictable response, which eliminates the need for regular monitoring. Dabigatran is a novel direct thrombin inhibitor that, at a 150-mg dose twice a day, has been shown to be superior to warfarin at reducing stroke or systemic embolism in AF patients (relative risk, 0.66; 95% CI, 0.53-0.82; P < .001), with both treatments showing similar rates of death and major hemorrhage after 2 years of follow-up.⁸³ Several model-based analyses have examined the cost-effectiveness of dabigatran at the 150-mg twice-daily dose compared with warfarin and found incremental cost-effectiveness ratios ranging from $25,000 per QALY to $86,000 per QALY under conservative base case assumptions.^84,85 The price of dabigatran and the rate of intracranial hemorrhage were the main cost drivers in these analyses.

Three direct factor Xa inhibitors have been approved by the FDA for the prevention of stroke in nonvalvular AF: rivaroxaban, apixaban, and edoxaban. The use of rivaroxaban 20 mg/d versus dose-adjusted warfarin in nonvalvular AF was tested in 14,264 patients in the ROCKET-AF randomized trial.⁸⁶ The investigators found that rivaroxaban reduced the primary end point of stroke or systemic embolism by 21% (HR, 0.79; 95% CI, 0.66-0.96). No economic analysis of ROCKET-AF by the original trial team has been done. Using published data from the ROCKET-AF trial and a variety of other sources, Lee et al⁸⁷ constructed a Markov model to assess the cost-effectiveness of rivaroxaban compared with warfarin from a US payer perspective. Rivaroxaban (assigned a monthly cost of $205) produced an average of 10.03 QALYs at a lifetime treatment cost of $94,456 (2011 US dollars), whereas warfarin produced an average of 9.81 QALYs at a lifetime treatment cost of $88,544. The incremental cost-effectiveness ratio for rivaroxaban was $27,498 per QALY for the base case analysis, with results most sensitive to risk of intracranial hemorrhage and stroke, drug cost, and time horizon. Current US pricing for rivaroxaban is approximately $10.75 per day (National Average Drug Acquisition Cost reference file, February 3, 2016) compared with the $6 per day this model assumed.

The cost-effectiveness of apixaban compared to warfarin has been assessed using patient-level data from the ARISTOTLE trial.⁸⁸ This prospective economic analysis was conducted from the US perspective using within-trial resource use paired with external cost weights from AF patients in a large registry of US hospitals. The analysis found similar costs in patients treated with apixaban and warfarin (excluding the cost of apixaban), but significantly longer life expectancy with apixaban (7.94 vs 7.54 QALYs), which resulted in an incremental cost-effectiveness ratio of $53,925 per QALY.⁸⁸ Edoxaban was compared with warfarin in 21,105 patients with moderate-to-high-risk AF in the ENGAGE AF-TIMI 48 trial who were followed for a median of 2.8 years.⁸⁹ Edoxaban was noninferior to warfarin for the prevention of stroke or systemic embolism and was associated with significantly lower rates of bleeding and cardiovascular death. Magnuson et al⁹⁰ performed a cost-effectiveness analysis using patient-level data from this trial together with a Markov model. Cost for edoxaban was $9.24 per day (2015 costs). Edoxaban had lifetime incremental costs of $16,384 and incremental QALYs of 0.444 (QALY: 7.16 years with edoxaban vs 6.72 years with warfarin), resulting in an incremental cost-effectiveness ratio of $36,862 per QALY.⁹⁰

Rate Control Versus Rhythm Control

The two primary approaches to managing AF are establishing and maintaining sinus rhythm (rhythm control) and controlling ventricular rate (rate control). Despite their conceptual appeal, rhythm control strategies have failed to demonstrate incremental reductions in death, stroke, worsening heart failure, or hospitalizations when compared with rate control.^91,92,93,94 In the Pharmacologic Intervention in Atrial Fibrillation (PIAF) pilot study, rhythm control therapy did not produce superior quality of life or symptom improvements over rate control, and it came with a higher rate of hospital admissions and adverse drug effects.⁹¹ The Comparison of Rate Control and Rhythm Control in Patients With Recurrent Persistent Atrial Fibrillation (RACE) trial, which randomly assigned 522 patients with persistent AF to either rate or rhythm control, found no significant differences in a composite primary end point of cardiovascular death, heart failure, thromboembolic complications, bleeding, pacemaker implantation, and severe adverse drug effects.⁹² The rhythm control arm experienced higher rates of thromboembolic events and adverse drug effects. In the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) trial, 4060 AF patients with a high risk of stroke or death were randomly assigned to rate or rhythm control and followed for up to 6 years. There was no significant difference in survival (the primary end point), and similar to earlier studies, the rhythm control strategy produced more hospitalizations and more adverse drug effects.⁹³ In a heart failure population with AF, in which the theoretical benefits of rhythm control over rate control should be easier to demonstrate, the Atrial Fibrillation and Congestive Heart Failure (AF-CHF) trial investigators found no significant differences between the two strategies in all-cause mortality, stroke, or worsening heart failure.⁹⁴ Consistent with earlier studies, the rhythm control arm experienced more frequent hospitalizations compared with the rate control arm (64% vs 59%; P = .06), with significantly more hospitalizations for AF (14% vs 9%; P = .001). Additionally, the rhythm control arm required more electrical cardioversion (59% vs 9%; P < .001).

Although there have been no prospective economic analyses in rate versus rhythm control trials, a retrospective economic evaluation of the AFFIRM trial found that rate control had a cost savings of $5077 per person over rhythm control and was more effective, thus dominating rhythm control from an economic perspective.⁹⁵

Atrial Fibrillation Ablation

Catheter ablation is a procedure that uses thermal energy or cryoenergy to selectively ablate electrical triggers and susceptible atrial substrate in an effort to eliminate AF.⁹⁶ As ablation techniques have evolved, pulmonary vein isolation (PVI) has emerged as the core ablation approach used in these procedures. Currently, PVI is primarily reserved for symptomatic AF patients who have not responded to at least one antiarrhythmic drug.

Several randomized trials have compared catheter ablation with antiarrhythmic drug therapy (AAD) and have focused largely on a reduction in AF recurrence as the primary end point.^{97,98,99,100,101,102} The current evidence base shows that ablation with or without AAD consistently outperforms AAD alone at preventing recurrence of AF.^{97,98,99,100,101,102} Two studies found significantly fewer hospitalizations in the ablation group compared with AAD alone,^99,101 whereas one study found no statistically significant difference in the median per-patient number of hospitalizations between the groups.¹⁰⁰ Three studies found significantly better improvements in quality-of-life measures for ablation compared with AAD.^97,101,102 One meta-analysis found radiofrequency ablation to have higher rates of efficacy with lower rates of complications than AAD therapy,¹⁰³ and another found higher rates of efficacy for PVI when compared with medical therapy, with the risk of complications associated with PVI comparable to that of other interventional procedures.¹⁰⁴

Although several studies have reported on the incremental economic effects of AF relative to no AF, there have been as yet no large-scale empirical cost studies of the contemporary AF ablation procedure.^105,106 An economic evaluation of the Radiofrequency Ablation for Atrial Fibrillation Trial (RAAFT) pilot study found the ablation strategy in that trial to be cost neutral at 2 years.¹⁰⁷ One literature-based decision model has explored some preliminary aspects of the cost-effectiveness of ablation therapy for AF and found it to be cost effective in patients with AF at moderate risk for stroke.¹⁰⁸ A Markov disease simulation model found ablation either in combination with AAD or alone to be cost effective for a hypothetical cohort of symptomatic, drug-refractory paroxysmal AF patients compared with AAD alone, with a cost-effectiveness ratio for ablation of $51,431 per QALY.¹⁰⁹

The National Heart, Lung, and Blood Institute–supported Catheter Ablation Versus Antiarrhythmic Drug Therapy for Atrial Fibrillation (CABANA; NCT00911508) trial randomized 2203 patients to receive either percutaneous left atrial catheter ablation as a first-line therapy or current state-of-the-art medical therapy using either rate control or rhythm control drugs. The primary end point is all-cause mortality, and the secondary end point is a composite end point of total mortality, disabling stroke, serious bleeding, and cardiac arrest. Prospective economic and quality-of-life assessments are part of the research program. Because the comparator strategy in CABANA is outpatient medical therapy, the ablation strategy will be significantly more expensive in the short-to-intermediate-term follow-up period because it requires both an invasive procedure and periprocedure inpatient therapy. The magnitude of this difference, based on current Medicare reimbursement rates, may exceed $10,000. If the expected reduction of AF is demonstrated, literature suggests that return to sinus rhythm will be accompanied by fewer physician and ED visits and a reduced rate of follow-up hospitalizations.^110,111,112 This may allow the ablation strategy to recoup at least a portion of the upfront procedural costs.^113,114,115

Coronary Artery Disease: Prevention of Coronary Artery Disease

CAD is still the number one cause of death in the United States. It is a major force of morbidity, disability, and human suffering and is a major driver in the growing cost of medical care. It is estimated that more than 15.5 million Americans ≥ 20 years of age have CAD and that one in every seven deaths is the result of CAD.³⁴ The combined direct and indirect burden of CAD was estimated to total more than $207.3 billion in expenditures in 2011 to 2012.³⁴ The conceptual appeal of prevention as an approach to the problem of CAD is clear. Although use of these terms in the literature is not always consistent, for this discussion, we shall take primary prevention to refer to therapies applied for the purpose of preventing overt disease manifestations, such as death, MI, angina, or heart failure in individuals without overt signs of atherosclerotic cardiovascular disease. The major therapeutic targets for primary prevention include smoking, lipid disorders, diabetes, hypertension, and physical inactivity. Secondary prevention refers to the application of therapeutics to prevent progression of disease in patients who already have manifest atherosclerotic cardiovascular disease. In addition to the therapies directed at the list of targets for primary prevention given above, secondary prevention adds pharmacologic agents according to the clinical presentation, which can include antiplatelet therapies, ACE inhibitors and ARBs, and β-blockers.

It is worth acknowledging that the seemingly sharp borders now present between primary and secondary prevention may become increasingly blurred as the growing use of coronary computed tomography angiography identifies more and more subjects with clear atherosclerotic arterial lesions that have not yet progressed to clinical disease. Whether and under what circumstances the more aggressive therapeutic strategies of secondary prevention should be applied to such subjects will need to be determined in large-scale clinical trials.

The economic analysis of CAD prevention illustrates well how cost-effectiveness analysis reflects the efficacy of production of health benefits. In both primary and secondary prevention, the general goal is the same: prevent death and morbidity. The risks (ie, the number of available events that need to be prevented) are of course much higher in secondary prevention patients. The incremental costs of prevention, however, tend to be relatively insensitive to the primary and secondary prevention distinction. In fact, a treatment as secondary prevention may have a lower net cost than the same treatment as primary prevention because of prevention of more expensive complications (eg, MIs) and expensive palliative procedures (eg, coronary revascularization). So secondary prevention costs the same or less and, for any relative reduction in major events, prevents a larger absolute number of complications. Primary prevention is often inefficient economically because many individuals with risk factors will never develop disease complications regardless of how they are treated. For this reason, many studies of primary prevention have attempted to identify higher-risk subsets of the at-risk population in which therapy may be more economically and clinically efficient. The use of a multicomponent inexpensive polypill in older adults is an alternate approach to efficient primary prevention targeting intervention at a population level based on simple demographic features.^116,117

Cholesterol Abnormalities

Although abnormalities of high-density lipoprotein (HDL) levels and other lipid subfractions appear to play a role in the risk some individuals have to develop and worsen atherosclerotic disease, our primary therapeutic evidence to date involves the use of 3-hydroxy-3-methyl-glutaryl–coenzyme A (HMG-CoA) reductase inhibitors (statins). For secondary prevention, a 2006 meta-analysis of four trials including 27,548 patients comparing intensive versus standard statin therapy in patients with stable CAD or acute coronary syndromes (ACS) concluded that intensive lipid lowering with high-dose statin therapy significantly reduced nonfatal cardiovascular events.¹¹⁸ A model-based analysis of hypothetical 60-year-olds found that high-dose statin therapy in patients with ACS added 0.35 QALYs, with the cost-utility ratio remaining below $44,000 per QALY and with a net daily price difference between high dose and conventional dose of $3.50.¹¹⁹ In patients with stable CAD, high-dose statin therapy added 0.10 QALYs, but for the cost-effectiveness ratio to remain below $50,000 per QALY, the net daily cost difference would need to remain below $1.70.

Recently, monoclonal antibodies that inhibit proprotein convertase subtilisin-kexin type 9 (PCSK9) (alirocumab, evolocumab) have been shown to lower low-density lipoprotein (LDL) cholesterol to levels beyond that achievable with statin therapy (by ~50%). Early clinical trial evidence suggests these drugs, which are administered by subcutaneous injection every 2 weeks, may also significantly reduce cardiovascular events (HR, ~0.50).^120,121 Use of these agents in one pilot study achieved LDL-cholesterol levels < 70 mg/dL in 75% of high-risk subjects versus 9% of patients on maximally tolerated statins alone.¹²² Pivotal clinical 5-year outcome trials are expected to report results in 2017. Following FDA approval of the first two agents in this class in 2015, pricing was set at around $14,000 to $15,000 per year (~$40/day). Preliminary economic analyses using the Coronary Heart Disease (CHD) Policy Model estimated cost-effectiveness ratios above $500,000 per QALY for secondary prevention and for treatment of familial hypercholesterolemia.¹²³ Five-year budget impact projections estimate 2.6 million persons receiving these therapies for 1 year or more at an annual cost of up to $74 billion. The drug cost estimated to achieve a cost-effectiveness threshold was $2412 per year, substantially below current prices. Updated economic analyses are expected once the long-term outcome trial data are published.

There are a number of important primary prevention trials of statin therapy, but none included prospective assessment of economics. A CHD Policy Model analysis of primary prevention strategies for lipid lowering assessed the cost-effectiveness and public health impact of Adult Treatment Panel III guidelines for targeting statin therapy in persons aged 35 to 85 years.^124,125 This analysis found that starting statin therapy in 9.7 million adults and intensifying statin therapy in an additional 1.4 million adults would reduce the number of MIs by 20,000 per year and the number of CAD deaths by 10,000 per year. Assuming a cost of $2.11 per low-intensity tablet and $2.81 per high-intensity tablet, over 30 years, the cost-effectiveness ratio for full adherence to Adult Treatment Panel III primary prevention guidelines would be $42,000 per QALY. In addition to drug costs, CAD risk level and age affected the economic attractiveness of statin therapy. Hypothetically, providing low-cost statin therapy to men over 44 years of age, regardless of lipid levels, might be an effective and economically attractive prevention strategy, although it seems unlikely to be viable on other grounds.¹²⁶

Oral Antiplatelet Agents

Aspirin

Long-term aspirin therapy is one of the primary pillars of secondary prevention. A recent patient-level meta-analysis estimated the benefit of aspirin as 10 fewer coronary events and 5 fewer strokes per 1000 patients treated.¹²⁷ Aspirin is now very inexpensive, with the generic drug costing approximately 1.5 cents per day. A cost-effectiveness analysis using the CHD Policy Model developed by Weinstein and Goldman estimated a cost per QALY with aspirin of approximately $11,000.¹²⁸ No large-scale empirical trials have studied the economics of aspirin for prevention, but it is quite likely that given the extremely low cost, the reduction in cardiac events would make it actually cost saving and therefore economically dominant.

The situation with primary prevention is more complex as a result of the ambiguity of the clinical data supporting this form of therapy.¹²⁹ A recent patient-level meta-analysis calculated that primary prevention with aspirin produced a 5 per 10,000 reduction in nonfatal MI, but no effect on death and a 3 per 10,000 increase in major bleeding.⁹⁴ A randomized trial of 100 mg of aspirin versus placebo in 29,980 subjects without known cardiovascular disease who were aged 50 to 75 years at study onset and who were judged to be at increased risk as a result of a low ankle-brachial index failed to detect a difference in vascular events or in mortality over a mean of 8.2 years.¹³⁰ However, a meta-analysis of primary prevention in men versus women showed that aspirin reduced MI in men by approximately one-third and reduced stroke in women by 17%.¹³¹ This meta-analysis was used as the basis for a model-based cost-effectiveness analysis that calculated that primary prevention aspirin had a cost per QALY of more than 100,000 for low-risk 55-year-old men, but was economically attractive for higher risk 55-year-old men and all men aged 65 years or older.¹³² An economically attractive cost-effectiveness ratio was obtained only in the women who were 75 years of age and in the highest risk women 65 years of age. A separate model-based analysis calculated that aspirin therapy was less costly and more effective than no therapy in 45-year-old men with a cardiovascular risk of 7.5% or higher over 10 years.¹³³ That same level of cardiovascular risk in 65-year-old women produced a cost per QALY of $13,300 with primary prevention aspirin therapy.¹³⁴ Although these economic analyses are interesting and potentially useful, the ambiguities of the underlying data should be kept firmly in mind, because a therapy that is not clearly clinically effective cannot be cost effective.¹³⁵ For both primary and secondary prevention, proper dosing and frequency remain unsettled.¹³⁶ PCORI has just funded a 20,000-patient secondary prevention trial comparing 81 mg versus 325 mg of aspirin that will hopefully clarify some aspects of this issue (theaspirinstudy.org).

Clopidogrel

The use of clopidogrel in addition to aspirin is considered here as part of the treatment of ACS. The additional question to be considered in this section is the value of extending therapy beyond the first year and possibly indefinitely. One model-based analysis that examined this issue found that, although therapy was very economically attractive for the first year, as it was extended beyond 2 years, economic attractiveness dropped substantially, particularly in lower risk patients.¹³⁷ A second model-based analysis that examined the cost-effectiveness of lifetime clopidogrel therapy added to aspirin for secondary prevention yielded a cost per QALY of greater than $130,000.¹²⁸ These results are dependent on two important variables: first, the use of a proprietary cost for clopidogrel, which is now generic and much less expensive, and second, the assumption that the treatment benefits of clopidogrel attenuate over time. Regarding this second point, the Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance (CHARISMA) trial randomized 15,603 patients with clinical CAD or multiple risk factors to clopidogrel or placebo plus low-dose aspirin and followed them for a median of 28 months.¹³⁸ Overall, the clopidogrel arm was not superior in the prevention of cardiovascular death, MI, or stroke, the primary composite end point. However, in the subgroup with clinically evident atherosclerosis, clopidogrel reduced the composite end point by 1 per 100 patients treated. A cost-effectiveness analysis of the secondary prevention subgroup of the CHARISMA trial projected a $2607 higher cost for the clopidogrel arm (using proprietary drug costs) and a 0.072 increase in life expectancy as a result of events prevented during the active treatment phase of the trial.¹³⁹ The resulting incremental cost-effectiveness ratio was $36,343/life-year saved. Although no updated analyses of this treatment comparison have been done for the United States since the advent of generic clopidogrel, a substantially lower drug cost would make the results described above more economically attractive. The results are primarily dependent on the assumptions of the analysis about the life expectancy losses attributable to a nonfatal MI (3.3-7 years) and to a nonfatal stroke (4-13 years dependent on severity). In this analysis, the assumption was made that clopidogrel would be stopped at the end of the CHARISMA treatment period, so these data do not address the value of longer-term prevention therapy with this drug. The Dual Antiplatelet Therapy (DAPT) study found that at least part of the incremental clinical benefit of 30 months of clopidogrel and aspirin over 12 months came in the form of prevention of cardiovascular events unrelated to the patient’s coronary stent.¹⁴⁰ However, a meta-analysis including the DAPT study failed to find a benefit for extended-duration dual therapy.¹⁴¹

Angiotensin-Converting Enzyme Inhibitors and Angiotensin Receptor Blockers

ACE inhibitors have been tested in several different secondary prevention patient subsets, including early after acute MI, with or without a long-term chronic phase of therapy, and in patients with chronic stable CAD. The Survival and Ventricular Enlargement (SAVE) trial was pivotal in establishing the benefit of captopril started within 3 to 16 days of an acute MI in patients with an ejection fraction of 40% or less.¹⁴² After a mean of 42 months of therapy, captopril produced an increase in survival of 5 per 100 patients, as well as significant reductions in severe heart failure, follow-up hospitalizations for heart failure, and recurrent MIs. A model-based cost-effectiveness analysis of SAVE found that captopril therapy was very economically attractive (cost-effectiveness ratio of $10,400 per QALY or less) as long as the benefits of therapy persisted at least for 4 years.¹⁴³ Similar clinical results were observed in the Trandolapril Cardiac Evaluation (TRACE) trial, which randomized 1749 patients 3 to 7 days after acute MI who had an ejection fraction ≤ 35% to trandolapril or placebo.¹⁴⁴ During a follow-up period of between 2 and 4 years, trandolapril produced 7 extra survivors per 100 patients treated relative to placebo. The beneficial effects of trandolapril were still evident at the 10- to 12-year follow-up of the study cohort, even though all patients were recommended for ACE inhibitor therapy at the end of the study’s active treatment period.¹⁴⁵ TRACE and SAVE together show a strong and persuasive mortality benefit for ACE inhibitor use in the post-MI population with left ventricular dysfunction. Given the number of serious and expensive complications prevented by the ACE inhibitors in these trials, had a detailed cost analysis been done assuming the use of a generic ACE inhibitor and all other things being equal, the ACE arms would almost certainly have been cost saving, thus making this form of therapy economically dominant.

In the Valsartan in Acute Myocardial Infarction (VALIANT) trial, acute MI patients with left ventricular dysfunction and/or heart failure had equivalent outcomes from the ARB valsartan or the ACE inhibitor captopril.⁴⁵ An economic analysis of the trial showed that the availability of generic captopril made this option the better choice from an efficiency perspective.¹⁴⁶

Use of ACE inhibitors as secondary prevention in patients with stable CAD without left ventricular dysfunction or clinical heart failure has been tested in three large trials: Heart Outcomes Prevention Evaluation (HOPE), European Trial on Reduction of Cardiac Events with Perindopril in Stable Coronary Artery Disease (EUROPA), and Prevention of Events with Angiotensin-Converting Enzyme Inhibition (PEACE). A combined analysis of these three trials involving 29,805 patients found that ACE inhibitors significantly reduced both all-cause mortality and nonfatal MI each by 1.1 patient per 100 patients treated.¹⁴⁷ A post hoc economic analysis of ramipril use in the HOPE trial showed that estimated cumulative 4.5-year costs did not differ between the ramipril and placebo arms.¹⁴⁸ In the EUROPA trial, the incremental cost-effectiveness for perindopril based on assigning costs to events prevented was £9700 per QALY.¹⁴⁹ The issue that remains unsettled regarding this form of secondary prevention is whether the ACE inhibitors are affecting outcome through their antihypertensive effects or through some other antiatherosclerotic mechanism.¹⁵⁰

Exercise

Physical inactivity is an established risk factor for cardiovascular disease,¹⁵¹ and the American Heart Association recommends that healthy adults aged 18 to 65 years engage in 30 minutes of moderate-intensity exercise 5 days per week or 20 minutes of vigorous exercise 3 days per week to improve physical health.¹⁵² Although these exercise recommendations appear modest, 23.7% of adults report being physically inactive, and more than 50% do not engage in minimum physical activity recommendations.¹⁵³ A population-based analysis by Wang et al¹⁵⁴ using a nationally representative sample estimated the burden of cardiovascular disease as a result of physical inactivity at $23.7 billion in direct medical expenditures in the United States during the year 2001. Although clinicians tend to favor pharmacology over lifestyle interventions for secondary prevention, the available evidence suggests they offer comparable benefits.¹⁵⁵

A recent analysis used a Markov model to estimate the cost-effectiveness of seven different public health programs aimed at improving participation in physical activity in healthy adults aged 25 to 64 years when compared with no intervention.¹⁵⁶ Data from clinical trials, population-based surveys, and published literature were used to estimate resulting reductions in disease burden. Cost-utility ratios ranged from $14,286 per QALY for an intensive 8-week paid media campaign to promote walking among sedentary adults to $68,557 per QALY for a community-wide campaign involving media messaging to promote physical activity, community walking events, and work and school programs.

A 2005 meta-analysis of 63 trials involving 21,295 patients with CAD found that exercise-based rehabilitation was associated with significant reductions in all-cause mortality (at 24 months but not at 12 months) and recurrent MI.¹⁵⁷ Although prospective economic analyses of exercise intervention in US patients with CAD are limited, a 2009 systematic literature review found various exercise interventions to be cost effective compared with usual care in trials from several countries.¹⁵⁸

Smoking Cessation

Eliminating tobacco use is an essential component of interventions to reduce the burden of cardiovascular disease. According to the US Centers for Disease Control and Prevention, smokers are at greater risk than nonsmokers of debilitating and costly chronic conditions, cardiovascular and respiratory diseases, and cancer, resulting in an average loss of 14 years of life per smoker as well as substantial productivity losses.¹⁵⁹ The costs of tobacco use in the United States were estimated at $193 billion annually between the years 2000 and 2004.¹⁶⁰ Understanding the complex issues surrounding tobacco, however, also involves recognition of its economic benefits. The tobacco industry is a source of substantial employment in some regions and huge tax revenues for most high-income countries.¹⁶¹

A 2003 comprehensive review of tobacco smoking and its direct medical costs in cardiovascular diseases found clinically-based smoking cessation therapies in several studies between 1977 and 1997 to have cost-effectiveness ratios between $1216 and $15,777 per life-year saved.¹⁶² Most of these studies reported on transdermal nicotine patches or nicotine gum therapy plus physician counseling compared with no treatment. A 2008 review of various tobacco control strategies from public policy interventions to quit lines to pharmacotherapy found the vast majority to be cost saving or highly cost effective.¹⁶³ One study conducted a model-based economic analysis of cognitive behavioral therapy added to nicotine replacement therapy for up to 40 weeks versus a 12-week smoking cessation program (using 2010 US dollars) and found the extended smoking cessation approach to be cost effective at $6324/QALY.¹⁶⁴

Blood Pressure Control

Hypertension is an established risk factor for MI, congestive heart failure, and stroke, all of which incur substantial economic and personal costs. The American Heart Association estimated the annual direct and indirect costs of high blood pressure to be $48.6 billion in 2011.³⁴ A 2009 meta-analysis of 147 randomized trials examining the use of blood pressure–lowering pharmacotherapy for the prevention of cardiovascular disease concluded that ACE inhibitors, ARBs, β-blockers, calcium channel blockers, and thiazides were all similarly effective at reducing cardiovascular events and stroke for a given reduction in blood pressure in both primary and secondary prevention cohorts.¹⁶⁵ β-Blockers were found to be significantly more effective in preventing recurrent cardiovascular events in post-MI patients, and calcium channel blockers were more effective at preventing stroke, but less effective at reducing the incidence of heart failure.¹⁶⁵

The literature on the economics of hypertension therapy is not as mature as might be expected as a result of wide variations in methods, data sources, and analytical approaches.¹⁶⁶ In general, the costs related to therapy can be broken down into the direct costs of drug therapy, the costs associated with treatment of side effects of therapy, the cost savings as a result of reduced cardiovascular events, and the costs related to increased or decreased productivity, particularly related to employment. An economic analysis of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) compared lisinopril and amlodipine with chlorthalidone as first-line treatment in patients with hypertension and one additional cardiovascular risk factor.¹⁶⁷ In this analysis, lisinopril had higher incremental cost and lower incremental life-years gained compared with chlorthalidone and thus was dominated from an economic perspective. Amlodipine had higher incremental cost, but provided an incremental gain of 0.016 life-years, with a cost-effectiveness ratio of $160,000 per life-year saved compared with chlorthalidone at 6 years of follow-up. Lifetime projections were a gain of 0.099 life-years and a cost-effectiveness ratio of $48,400 per life-year saved. The drug cost of amlodipine in this analysis was calculated at $2.47 per 10-mg tablet using 2004 wholesale prices for a proprietary formulation, whereas chlorthalidone was valued at $0.19 per 25-mg tablet. The patent for amlodipine expired in 2007, and generic versions are now available at a retail cost of $0.17 per 10-mg tablet versus $0.43 per 25-mg tablet of chlorthalidone, suggesting that amlodipine could economically dominate chlorthalidone given today’s drug prices.¹⁶⁸ These price fluctuations underscore the temporal limitations of cost-effectiveness analyses.

A cost-effectiveness analysis of the implementation of the 2014 guidelines for hypertension therapy relative to the status quo (reflected in National Health and Nutrition Examination Survey data from 2007 to 2010) found that full implementation of the guideline recommendations would reduce cardiovascular deaths and nonfatal events in the United States while also being cost saving.¹⁶⁹ Treatment of stage 2 hypertension (systolic blood pressure ≥ 160 mm Hg) was cost saving even if drug therapy costs were doubled over the base case. Treatment of stage 1 hypertension (systolic blood pressure 140-154 mm Hg) had a cost-effectiveness ratio < $50,000 per QALY for all men and for women aged 45 to 74 years.

CORONARY ARTERY DISEASE: Acute Coronary Syndromes

(See also Chaps. 39 and 40.) ACS is commonly divided into ST-segment elevation MI or non–ST-segment elevation ACS, because of the different initial therapeutic approaches required. The most important initial decision in the ST-segment elevation acute MI patient with important economic implications is the assessment for, and selection of, acute reperfusion therapy, either primary PCI or fibrinolytic therapy. For the non–ST-segment elevation ACS population, the major management decision with important economic ramifications relates to the risk stratification strategy to be used, either early invasive or early conservative. All patients also receive secondary prevention, which is discussed elsewhere in this chapter.

Acute Reperfusion Therapy

As noted at the beginning of this chapter, economic analysis is centrally concerned with trade-offs: What are the (long-term) consequences of doing X rather than Y? This means that the relevant economic questions about acute reperfusion therapy depend on what alternatives are available. At some centers, for example, primary PCI may not be available, and the options may be treat with fibrinolytic therapy or transfer for primary PCI. At primary PCI centers, the fibrinolytic option may not be relevant unless all the catheterization laboratories are full, with almost all reperfusion-eligible patients going for emergent catheterization. But even in such referral centers, current guidelines recommend different choices depending on how quickly patients can reach the catheterization laboratory. When both primary PCI and fibrinolytic therapy are available, the weight of evidence favors primary PCI as a result of improved survival (~2 extra survivors per 100 patients treated) and lower stroke risk.^170,171

A recent cross-sectional study from Europe of acute ST-segment elevation MI care in 2010 to 2011 found a substantial number of patients still not receiving any reperfusion therapy and large variations in the amount and type of therapies used in different countries.¹⁷² The largest modern trial, The Danish Acute Myocardial Infarction 2 (DANAMI-2) study, recently reported that, at a median follow-up of 7.8 years, the primary end point of death or MI had occurred in 35% of the primary PCI group and 41% of the fibrinolysis group (HR, 0.78).¹⁷³ Importantly, the relative benefits of primary PCI appeared to be the same in older and younger patients.¹⁷⁴ In addition, DANAMI-2 reported that at 1 month, several major dimensions of quality of life were better in the primary PCI arm.¹⁷⁵ Whether these benefits persisted for the longer term is unknown.

The only relatively modern empirical cost comparison of these two strategies comes from the 123-patient Canadian Stenting Versus Thrombolysis in Acute Myocardial Infarction Trial (STAT).¹⁷⁶ The cost of alteplase in Canada when this study was conducted in 1999 was $1809, whereas the cost of the primary PCI procedure itself was estimated at $2009. In this unblinded trial, the primary PCI arm had a 2-day shorter hospital stay and fewer unscheduled invasive coronary angiograms than the alteplase arm during the index hospitalization and fewer follow-up readmissions. By 6 months, the primary PCI arm had approximately $2500 lower medical costs than the fibrinolytic arm. Taking the clinical data showing superior survival effects in the short term that persist long term and the suggestion of lower costs, the evidence available suggests that in situations when primary PCI has clinical superiority over fibrinolysis, it is economically dominant (lower costs as well). One important caveat about these comparisons is that changes in the price of fibrinolytics or stents over time can affect the incremental differences. For example, in the early stent era, the Primary Angioplasty in Myocardial Infarction (STENT PAMI) trial estimated that the PCI arm had higher hospital costs as a result of the higher costs of the stents, although these costs were partially recouped by the lower need for follow-up invasive procedures.¹⁷⁷ Over the intervening years, the costs of stents, particularly bare metal stents, have come down sharply, whereas the costs of t-PA and tenecteplase have remained stable (primarily because of the lack of generic competition).

Several European model-based analyses have examined the cost-effectiveness of primary PCI versus fibrinolysis and found either economic dominance¹⁷⁸ or a cost-effectiveness ratio of less than £20,000.¹⁷⁹

Early Invasive Versus Early Conservative Management Strategies

In the patient with non–ST-segment elevation ACS, two major management options have been used and tested in clinical trials. The early invasive strategy involves routine referral to invasive angiography for all patients without contraindications, typically within the first 18 to 48 hours after presentation. The early conservative strategy reserves invasive study for a subgroup of patients known to be at particularly high risk because of specific features, such as ongoing ischemic symptoms that cannot be controlled medically, and treats all others with initial medical therapy. In evaluating the comparisons of these two strategies, it is important to recognize that there have been many variations of the early conservative strategy tested, some of which used early angiography in only approximately 10% of patients, whereas others referred more than 50% of such patients. In addition, the options for adjunctive medical therapy have been evolving over time, first with the use of intravenous glycoprotein IIb/IIIa inhibitors and more recently with adenosine diphosphate (ADP) receptor antagonist antiplatelet agents.

A meta-analysis from the Cochrane Library involving five trials and 7818 patients found a trend toward harm with the early invasive strategy for in-hospital mortality (relative risk, 1.6), with no evidence for a longer-term mortality benefit.¹⁸⁰ Both MI rates and refractory angina were reduced in the early invasive strategy, and early and intermediate rehospitalization rates were also reduced. The invasive strategy was associated with a two-fold increased risk of a periprocedural MI. The complexity of attempting to get an “answer” about the best strategy using meta-analysis can be illustrated by comparing trials with different versions of the early conservative strategy. In the Fast Revascularization During Instability in Coronary Artery Disease (FRISC) II trial, 10% of the conservative arm received catheterization during the index hospitalization.¹⁸¹ In this trial, the early invasive strategy was associated with a 1.7 per 100 absolute reduction in mortality and a 2 per 100 absolute reduction in MIs. In the Randomized Intervention Treatment of Angina (RITA) 3 trial, only 16% of the early conservative group was referred for angiography during the index admission.¹⁸² No difference in death or MI was observed early in the trial, but at 5 years, the early invasive arm had a 22% lower mortality.¹⁸³ By contrast, in the Treat Angina with Aggrastat and Determine Cost of Therapy with an Invasive or Conservative Strategy–Thrombolysis in Myocardial Infarction (TACTICS-TIMI 18) trial, 51% of patients in the early conservative arm were referred for cardiac catheterization during the index hospitalization.¹⁸⁴ No difference in mortality between the two arms was evident out to 6 months, but there was a 2 per 100 patients treated absolute reduction in the rate of MI. Economic analysis of this trial showed that the index hospitalization was approximately $1600 more expensive for the early invasive arm, but follow-up costs out to 6 months were approximately $1000 lower in this group.¹⁸⁵ Thus, the net 6-month incremental costs of the early invasive strategy were $670. Costs per life-year added for the invasive arm relative to the conservative arm were approximately $13,000. This result rests on the pivotal assumption that the MIs prevented during the early phase of the trial would translate into added survival with much longer follow-up of the study cohort.

In the Timing of Intervention in Acute Coronary Syndromes (TIMACS) study, Medicare diagnosis-related group reimbursements were used to compare the costs of an early versus delayed (> 36 hours) invasive strategy in ACS patients.¹⁸⁶ The early invasive group averaged $1170 lower costs overall and $3720 lower costs in high-risk patients (Global Registry of Acute Coronary Events score ≥ 141). The primary source of the difference in costs was a shorter length of stay for the early invasive patients.¹⁸⁷

Anticoagulant Therapy

Intravenous unfractionated heparin has been a standard part of medical therapy for ACS since before the initial American College of Cardiology/American Heart Association Guidelines were published in 1994, but it is quite inexpensive, and the economics have not been studied empirically relative to medical therapy without anticoagulation. Instead, heparin has formed the “usual care” group in comparisons with newer anticoagulation therapies. Clinical outcomes for the low-molecular-weight heparin enoxaparin have been compared with unfractionated heparin in several meta-analyses.^188,189 These analyses show a 9% to 16% relative reduction in death or MI for enoxaparin. However, these benefits were not evident in the Superior Yield of the New Strategy of Enoxaparin, Revascularization and Glycoprotein IIb/IIIa Inhibitors (SYNERGY) trial, in which 92% of patients underwent diagnostic catheterization and the majority received clopidogrel and/or a glycoprotein IIb/IIIa inhibitor.¹⁹⁰ These benefits were also not seen in the A-to-Z Trial in patients who received tirofiban and were intended for an early invasive strategy.¹⁹¹ However, there was a large benefit evident in the patients treated with tirofiban and an intended early conservative strategy.

The Organization for the Assessment of Strategies for Ischemic Syndromes (OASIS)-5 trial, which compared enoxaparin with fondaparinux in 20,078 patients with non–ST-segment elevation ACS, found a reduction in death, MI, or refractory ischemia with fondaparinux that was primarily a result of an unexpected mortality reduction.¹⁹² A model-based analysis of the OASIS-5 trial from a US perspective found that over the first 6 months, the fondaparinux regimen was associated with a cost savings of approximately $550.¹⁹³ A long-term extrapolation indicated that the fondaparinux strategy would be economically dominant in most scenarios considered.

Antiplatelet Regimens

Three oral ADP receptor antagonist antiplatelet agents are of relevance to management of ACS patients during the first year after hospitalization: clopidogrel, prasugrel, and ticagrelor (Table 112–3). The issue of longer term antiplatelet therapy is addressed in the portion of this chapter devoted to prevention.

The clinical benefits of clopidogrel therapy in this clinical context were firmly established by the Clopidogrel in Unstable Angina to Prevent Recurrent Ischemic Events (CURE) trial, which randomized 12,562 patients with non–ST-segment elevation ACS to a load (300 mg of clopidogrel or placebo) followed by daily maintenance therapy (75 mg of clopidogrel or placebo).¹⁹⁴ Over the mean study follow-up of 9 months, clopidogrel significantly reduced the primary composite end point (death, MI, or stroke) from 11.4% to 9.3% and nonfatal MI from 6.7% to 5.2%. Cardiovascular death alone was also reduced from 5.5% to 5.1%, but the reduction was not statistically significant. An economic analysis from the CURE trial reported that the clopidogrel arm generated $325 per patient in hospital cost savings and had $766 of extra costs for the clopidogrel, resulting in a net cost of $442 per clopidogrel patient.¹⁹⁵ Using Framingham Study data to project life expectancy for the trial patients (0.069 life-years per clopidogrel patient), these investigators calculated a cost per life-year saved of $6318. A second model-based analysis using the published results of CURE and assuming a year of clopidogrel therapy calculated a similar incremental cost-effectiveness ratio.¹³⁷

Prasugrel has a more rapid onset of effect and more platelet-inhibiting potency than clopidogrel. This enhanced potency translated into a significant reduction in the composite of cardiovascular death, MI, or stroke in the pivotal phase III trial used for FDA approval, the Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition With Prasugrel Thrombolysis In Myocardial Infarction (TRITON-TIMI)-38.¹⁹⁶ Over a median follow-up of 15 months, 9.9% of patients in the prasugrel arm and 12.1% in the clopidogrel arm had a primary event. The cost-effectiveness analysis used prospective data from eight countries participating in the trial.¹⁹⁷ Hospitalization costs were estimated using US diagnosis-related group weights. The cost of the drug regimens was assigned using net wholesale prices: $4.62 per day for clopidogrel and $5.45 per day for prasugrel. Life expectancy for the study cohort was estimated using long-term survival data from the Saskatchewan Health Database. Over the median study follow-up, the prasugrel arm saved $221 per patient, primarily as a result of reduced need for follow-up PCI. Incremental life expectancy with prasugrel was calculated as 0.102 life-years per patient, primarily as a result of the early reduction in nonfatal MIs. Given better life expectancy and lower cost, prasugrel was economically dominant over clopidogrel in this analysis. Substituting generic clopidogrel at a cost of $1 per day yielded a cost-effectiveness ratio for prasugrel of approximately $10,000 per life-year saved. As with the PURSUIT economic analysis discussed earlier, this result is dependent on the assumption that prevention of end point MIs as occurred in this trial is of continuing long-term prognostic benefit. The population of TRITON-TIMI-38 comprised ACS patients referred for PCI. The Targeted Platelet Inhibition to Clarify the Optimal Strategy to Medically Manage Acute Coronary Syndromes (TRILOGY-ACS) trial evaluated the use of prasugrel (10 mg daily) versus clopidogrel (75 mg daily) for up to 30 months in 7243 patients < 75 years of age with unstable angina or non–ST-segment elevation MI who did not undergo revascularization but were receiving aspirin.¹⁹⁸ Among this group of patients, prasugrel did not significantly reduce the frequency of a composite of death from cardiovascular causes, MI, or stroke, and similar risks of bleeding were observed.

Ticagrelor, an oral, reversible ADP receptor antagonist, has recently been shown to be superior to clopidogrel in the Platelet Inhibition and Patient Outcomes (PLATO) trial involving 18,624 ACS patients with or without ST-segment elevation.¹⁹⁹ At 12 months, the composite of vascular death, MI, or stroke occurred in 9.8% of the ticagrelor patients and 11.7% of the clopidogrel patients. All-cause death was reduced from 5.9% to 4.5% (P < .001) by ticagrelor. No difference in major bleeding was observed. An economic analysis of this study assessed the cost-effectiveness of ticagrelor versus clopidogrel from the perspective of the US healthcare system.²⁰⁰ Within-trial resource use was estimated using the portion of the US cohort on low-dose aspirin therapy to match the FDA-approved indication. Quality-adjusted life expectancy was assessed using the total PLATO population combined with long-term survival data from the PURSUIT cohort. Ticagrelor therapy in this population had a cost-effectiveness ratio of $29,665 per QALY relative to clopidogrel.

A decision model–based analysis using the perspective of the Ontario Ministry of Health found ticagrelor had a cost-effectiveness ratio of $12,205 (Canadian dollars) per QALY relative to clopidogrel, whereas prasugrel had a cost-effectiveness ratio of $57,630 per QALY.²⁰¹ A second analysis found that genotyping might improve the cost-effectiveness of prasugrel and ticagrelor, although the strategies require clinical trial validation.²⁰²

Coronary Artery Disease: Chronic Stable Coronary Artery Disease

Atherosclerosis is a chronic progressive disorder that commonly involves the epicardial coronary arteries. In the absence of effective therapy, it tends to progress to symptomatic limitation of function from angina pectoris, acute MI, progressive left ventricular dysfunction, and death. The course of individual patients, however, can be quite variable. Two main therapeutic strategies have evolved to address the symptomatic and prognostic consequences of this disease, preventive medical therapy (see the prevention section of this chapter) and coronary revascularization.

Coronary Revascularization Overview

(See also Chaps. 42 and 44.) The number of revascularization procedures in the United States peaked in 2010 and has been declining since then.²⁰³ In 2010, 397,000 CABG procedures and 954,000 PCI procedures were performed.³⁴ Unfortunately, the evidence pertaining to the clinical questions of most importance (which patients derive significant clinical benefit from CABG and which from PCI) is largely out of date and, therefore, of unclear relevance to contemporary practice. The trials of CABG versus medicine that demonstrated a survival benefit for CABG in high-risk anatomic CAD subsets were all done before the elements of current optimal medical therapy were in use. In addition, the operations that were done during that era did not routinely employ internal mammary arteries for the left anterior descending graft and did not employ aggressive secondary prevention after surgery.²⁰⁴ In a 1994 meta-analysis of the CABG versus medicine trials, Yusuf et al²⁰⁵ estimated that CABG extended the life expectancy of left main disease patients by 19 months and that of three-vessel disease patients by 5 months over a projected 10-year follow-up. Although there are no modern empirical cost studies relevant to this comparison, model-based analyses demonstrate that with such a large survival benefit, even an incremental cost of $30,000 or more per patient treated with CABG rather than medicine yields an economically attractive cost-effectiveness ratio under most reasonable assumptions.²⁰⁶ It should also be noted that model-based analyses suggest that CABG is also cost effective in highly symptomatic patients in whom durable and effective symptomatic control cannot be achieved with medicine alone.²⁰⁶

The advent of PCI, initially as percutaneous transluminal coronary angioplasty followed by bare metal stents and then first- and second-generation DESs, has altered both the clinical playing field and the way economic questions about this area are framed. As with the CABG and medicine comparisons, evaluation of the PCI clinical trials is made challenging by the evolving technology, changes in the population considered suitable for the procedure, and changes in the adjunctive medical therapy used during and after the procedure. Two sorts of clinical trials have been performed with PCI, asking very different, but complementary, questions. The first set of trials, comparisons with medicine and with CABG, attempt to define the role of PCI in the larger management picture of CAD. The second set of trials compares various improvements in PCI technology and adjunctive medical management in order to determine how to make the procedure as effective as possible. None of these latter trials by themselves, however, establish which patients should receive a PCI rather than an alternative therapy.

Percutaneous Coronary Intervention Versus Medicine

In a meta-analysis using data from more than 25,000 patients with stable CAD, no form of PCI (including use of DESs) reduced the risk of death or MI relative to medical therapy.²⁰⁷ However, a larger meta-analysis found a mortality benefit for second-generation DES versus medicine.²⁰⁸ As with medical therapy and CABG, the clinical and economic effects of PCI have evolved substantially over the past two decades.²⁰⁹ The most recent large-scale trial of PCI versus medicine is Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE), which randomized 2287 stable CAD patients at 50 North American sites between 1999 and 2004 to medical therapy or medical therapy plus PCI with bare metal stents.²¹⁰ As with the meta-analysis just discussed, the primary end point of death or nonfatal MI was not different in the two arms after a median follow-up of 4.6 years. One important feature of the COURAGE trial was a careful prospective analysis of angina relief and quality of life.²¹¹ Six months after randomization, the PCI group had a small, but statistically significant, advantage in angina relief and quality of life, but this treatment effect attenuated over time and was gone by 3 years. Using Medicare reimbursement rates to assign cost weights, the COURAGE investigators estimated an incremental cost of PCI of $11,410 initially, falling slightly to $9451 when projected over the study cohort’s lifetime (~12 years).²¹² Using only the empirical trial follow-up, the COURAGE investigators estimated a 0.03 life-year survival advantage for PCI and a 0.05 QALY advantage. In the lifetime extrapolation, the QALY advantage for PCI was 0.06 (95% CI, –0.21 to 0.43). In the lifetime extrapolation, the cost per QALY produced with PCI was $168,000, and only 10% of 5000 bootstrap replications fell below $50,000 per QALY. The message from COURAGE clearly is that routine PCI in this population is a very inefficient (and societally unattractive) way to produce extra health benefits. This result, as with most cost-effectiveness analyses, is primarily driven by the very small (and statistically insignificant) treatment effect of PCI on hard clinical outcomes. What remains unclear is whether a different result might have been obtained if COURAGE had required all patients to have a significant inducible ischemia on functional testing. This question is the subject of the ongoing approximately 5000-patient International Study of Comparative Health Effectiveness with Medical and Invasive Approaches (ISCHEMIA) trial (ISCHEMIA; NCT01471522).

The most recent comparison of medicine versus revascularization is the Bypass Angioplasty Revascularization Investigation (BARI) 2D trial, which randomized 2368 patients with type 2 diabetes and stable CAD.²¹³ Overall, total mortality (the primary end point) did not differ in the two arms. One interesting feature of the BARI 2D was a prospective definition of a PCI and a CABG subset before randomization. In the CABG subset, major cardiovascular events were reduced by CABG relative to medicine alone. Economic analysis results differed substantially in these two strata.²¹⁴ In the PCI stratum, at the end of 4 years, PCI costs were higher by $5600 and clinical outcomes were better in the medical arm. Extrapolating to a lifetime perspective, total costs for the medical arm were approximately $200 higher than PCI, but life expectancy was longer and the cost-effectiveness ratio for medicine versus PCI was $600 per added life-year. In the CABG stratum, 4-year costs were $20,300 higher for the CABG patients. In the lifetime projections, CABG increased survival by 0.52 life-years per patient at an incremental cost that remained at approximately $20,000. The cost-effectiveness ratio for CABG versus medicine was $47,000 per added life-year.

Percutaneous Coronary Intervention Versus Coronary Artery Bypass Graft

Trials performed to evaluate, among patients requiring revascularization, which procedure should be performed have gone through several stages, largely reflecting evolution in the technology of PCI. One interesting observation from a comparison of the BARI randomized and registry cohorts is that compared with the registry patients, patients randomized into this trial tended to be higher risk than the typical percutaneous transluminal coronary angioplasty patients and lower risk than the typical CABG patients.²¹⁵ Thus, when interpreting the clinical trials comparing these two revascularization strategies, it is worth keeping in mind that the patients are not likely representative of unguided clinical practice. A 2014 meta-analysis of CABG with one or more arterial grafts versus PCI with stents included six trials involving 6055 patients and found that CABG reduced total mortality at 4 years as well as MI and repeat revascularization.²¹⁶

Economic analyses of the trials from the bare metal stent era of PCI versus CABG comparisons (the Arterial Revascularization Therapy Study [ARTS] and the Stent or Surgery [SOS] trial) typically showed the expected initial advantage of PCI that narrows over the first year as a result of repeat procedures. Longer-term economic follow-up was not reported for these trials.

The Synergy between PCI with Taxus and Cardiac Surgery (SYNTAX) trial compared PCI with DES versus CABG in three-vessel or left main disease patients.²¹⁷ After 5 years of follow-up, there were statistically significant differences favoring CABG in all-cause mortality (9.2% vs 14.6% for PCI) and the composite of death/stroke/MI (14% vs 22% for PCI) as well as MI alone (4% vs 9.2% for PCI).²¹⁸ Cohen et al²¹⁹ used a model-based analysis to assess the costs and cost-effectiveness of CABG relative to PCI using within-trial data from SYNTAX. PCI had higher initial procedural costs ($3415 per patient), whereas CABG had higher total hospitalization costs ($10,036 per patient). PCI had higher 5-year costs (excluding the index revascularization procedure) as a result of more frequent hospitalizations, revascularizations, and medication costs. CABG had higher projected lifetime costs ($92,509 vs $87,428 for PCI; difference of $5081), higher QALYs (10.46 vs 10.12 for PCI; difference of 0.34), and a base case cost-effectiveness ratio of $16,537 per QALY and $12,329 per life-year.

The Future Revascularization Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multivessel Disease (FREEDOM) trial randomized 1900 patients with diabetes mellitus and multivessel CAD to PCI with DES or CABG. The primary end point of death, MI, or stroke was significantly lower in the CABG-treated group (18.7%) versus PCI (26.6%) at 5 years of follow-up.²²⁰ A prospective economic evaluation was conducted alongside the trial, with costs assessed from the perspective of the US healthcare system. The CABG group had $8622 per-patient higher index hospitalization costs and $3641 higher per-patient 5-year cumulative costs.²²¹ Life expectancy was higher in the CABG group (4.665 vs 4.613 years) and remained higher when quality adjusted (3.719 QALYs vs 3.688 QALYs). With within-trial survival extended to a lifetime horizon, the incremental cost-effectiveness ratio for CABG versus PCI was $6791 per life-year and $8132 per QALY gained.

Although the full impact is presently far from clear, the passage of the PPACA has initiated a reshaping of the medical economy that will play out for decades to come and is equaled in potential impact only by the creation of Medicare and Medicaid in 1965. What the PPACA has done most clearly is expand some level of health insurance coverage to millions of uninsured Americans. Its effects on the cost of US medical care are much less clear, and some analysts have predicted that the bill will actually increase costs by 2019 over projections made before the bill was passed. Many in the policy world may look to the bill’s new support for comparative effectiveness research as a tool for controlling the growth of medical spending. The perspective advanced in this chapter is that an understanding of the concepts, tools, and approaches of medical economics can help make cardiovascular clinicians more effective advocates in the debates sure to come about what forms of care are worth supporting with societal healthcare dollars. However, neither medical economics nor comparative effectiveness provides the whole story. Questions of value and justice and how we see ourselves as a society will always be critical elements in deciding how much and what types of healthcare get supported in the United States. The only thing that is certain is that the American healthcare system is very much a work in progress.