Economic Issues in Pediatric Cancer
Susan K. Parsons
Pei-Jung Lin
Joshua T. Cohen
The rapid rise in health care costs in the United States has become an increasingly dominant news story in both the popular press and in academic medical journals. In part, this trend reflects the unprecedented growth in the number of older Americans. By 2030, nearly one in five Americans will be at least 65 years old.1 The disproportionate utilization of medical services by this group is expected to contribute substantially to further increases in health care spending, which is expected to balloon to more than 20% of gross domestic product (GDP) by 2018, from its already large share of approximately 16% in 2007.2 The projected increase in the number of individuals diagnosed with cancer is a case in point. Because cancer diagnoses are more common in older adults, the total cancer incidence is expected to increase by 45% from 1.6 million in 2010 to 2.3 million by 2030.3
The rapidly growing population of older Americans is not the only cause of increasing health care costs, however. New, more expensive technologies and medications are also critical factors. This is where childhood cancer fits into the picture. Although the incidence of cancer in children aged 14 years and younger has remained relatively steady at 15 per 100,000 over the past 5 years,4 the cost of therapy continues to increase. According to the National Cancer Institute, national expenditures for cancer totaled $124 billion in 2010, and this number is projected to increase more than 25% by 2020, reaching $157 billion. Most expensive are leukemia and lymphoma (accounting for 14% of total expenditures), breast cancer (13%), and colorectal cancer (11%). The cost increase is, in part, due to use of more new, higher-price oncolytic agents. As an example, a 2013 editorial signed by more than 100 chronic myeloid leukemia (CML) experts criticized high prices for cancer drugs, focusing in particular on tyrosine kinase inhibitors (TKIs)—one of the most successful targeted therapies for cancer.5 The introduction of the first TKI, imatinib, has improved 10-year survival rate in CML from 20% to 85%. The cost of imatinib, however, increased from roughly $30,000 per year in 2001 ($39,000 in 2012 dollar-equivalent) when it was first released, to $92,000 per year in 2012.
High costs of cancer care have imposed a substantial financial burden on patients and payers, leading to many discussions about the value of medical spending in cancer. One study reported that between 1988 and 2000, improvements in cancer survival saved 23 million life years at a cost of roughly $1.9 trillion, implying that the average life year costs $82,000.6 In an analysis comparing U.S. cancer costs and survival with those of 10 European countries, the growth in spending on cancer care has been found to be greater in the U.S. than in Europe, but that cancer survival gains in the U.S. have likewise been greater.7 In other words, the higher cancer costs in the U.S. appeared to be “worth it,” in terms of additional survival, with almost all gains coming from prostate and breast cancers. In one analysis of the value of medical spending in the U.S., Cutler et al.8 found that although health care costs have increased over time, the return on spending has been high. They reported that 90% of the life expectancy increases during 1960 to 2000 have resulted from reductions in infant mortality and in death from cardiovascular disease.
New medications offer promise to millions of people diagnosed with cancer each year. Yet the financial cost of cancer care for patients is substantial. For example, from 2008 to 2010, the annual excess economic burden of cancer survivorship among recently diagnosed adults was more than $16,000 per survivor. A survey showed that cost concerns were common among cancer patients, even though most of them had health insurance: 30% reported having concerns about paying for their cancer treatment; 22% reported that their family had made sacrifices to pay for their care; and only 8% stated that their insurance adequately covered their current health care costs.9
In a recent analysis, Bach10 estimated that average monthly cancer drug costs exceed $5000 per patient. As illustrated in Figure 51.1, these costs have increased substantially in recent years. Bach attributed this increase primarily to “a unique legislative and regulatory framework that shields cancer drugs … from the strategies that health care payers such as Medicare typically use to hold down the price and utilization of drugs and other health care goods.”
Resorting to the use of more and more expensive treatments to extend survival is consistent with what some observers have described as a uniquely American tendency to believe that the provision of care should not be influenced by cost considerations. This perspective is evident in the rejection in 1989 of a proposal that would have permitted Medicare to take into account care costs when making coverage decisions. Among the reasons for the failure of this proposal identified by Neumann11 (p. 1516) are “Americans’ affinity for new medical technology, a distaste for explicit limit setting, a sense of entitlement with regard to Medicare funds, [and] the perception that in a vast and wealthy country, health care resources are not really constrained….” Efforts by the state of Oregon in the 1980s to prioritize its coverage of Medicaid services based on health benefits per dollar spent were likewise rejected11 and no state since has attempted to introduce that kind of prioritization scheme.12 A 2009 survey conducted by the Kaiser Family Foundation, National Public Radio, and the Harvard School of Public Health13 revealed a similar sentiment among a majority of Americans older than 18 years. In that survey, 56% of respondents stated that insurance companies should be required to cover
expensive treatments even if those treatments have not been shown to be superior to less expensive options.
expensive treatments even if those treatments have not been shown to be superior to less expensive options.
 Figure 51.1 Monthly costs for cancer treatment medications |
American oncologists also resist the idea that cost should be a factor in treatment decisions. In a survey of members of the American Society of Clinical Oncology (ASCO),14 67% of respondents stated that patient access to effective cancer treatment should not be influenced by cost. Nonetheless, oncologists recognize the inevitable impact of costs, with 56% saying that drug costs influence their treatment recommendations, and even more (84%) saying that patient out-of-pocket (OOP) expenses influence their decisions.
There can be little doubt that for cancer interventions in general, and in the case of pediatric oncology in particular, the role of cost in the provision of care will continue to be controversial. On one hand, cancer survival rates for both adults and children continue to improve as new, effective therapies emerge4 and significant progress is made in early detection and treatment. On the other hand, as costs increase, trade-offs must be confronted even when new therapies improve outcomes. Hillner and Smith15 pointed out that it is easy to support payment for the so-called “home run” therapies that provide substantial health benefits. They add, however, that “The much stickier issue will be drugs that add a small benefit at a high cost … [A]s oncologists we have to discuss these issues with our patients.” Recent publications suggest approaches for discussing monetary concerns with patients with cancer and mark an emerging interest among oncologists to understand the impact of economic issues on patient care.16,17
PATIENT-PHYSICIAN COMMUNICATION ABOUT COST OF CANCER CARE
Honest, evidence-based, and transparent patient-physician communication is the cornerstone of shared decision-making and informed consent, and cost of cancer care is a crucial part of the communication. The 2009 ASCO guidance on the cost of cancer care further highlights the importance of cost discussion. This statement argued that patient-physician discussion regarding the cost of care should be recognized as an important component of high-quality care. Clinicians are advised to communicate with their patients about their OOP costs and help patients take into account these considerations in treatment decisions. Additionally, resources should be developed to educate patients about costs of treatment and guide them in choosing treatment options.
Importantly, the ASCO statement also noted that oncologists should integrate cost implications for individual patients into treatment considerations. According to a 2010 ASCO survey, most doctors stated that cancer drug costs were influencing their clinical practice decisions. About 73% of these oncologists said that, going forward, the costs of new cancer drugs “will play a more significant role in my decisions regarding which cancer treatments to recommend for my patients.” However, only 42% reported that they always or frequently discussed costs with patients. These results may reflect the sensitivity of this topic but also point to the need for better patient-physician communication about cancer costs.14
To fully address this situation, two issues must be confronted. First, how do we determine whether an intervention’s health benefits are sufficiently meaningful to justify its costs? The first part of this chapter reviews the tools developed by health economists to assess the value of an intervention. Second, if the benefits are sufficiently large, how do we finance the provision of care? The second part of this chapter describes the available financing mechanisms in the United States and explores their limitations.
METHODS OF ECONOMIC EVALUATION
The need to make health care more economically sustainable has stimulated demand for evaluations of intervention costs. This demand gave rise to cost analysis, which formally reviews what is being spent on the intervention being evaluated, where, and by whom. Cost analysis aids identification of gaps or the problems with current expenditures and helps delineate future trends. Analysis of costs is particularly salient in the context of cancer treatment due to the use of new and expensive technologies and the aging of the population.
The value of conducting economic evaluations of health care expenditures is based on the assumption that although demand for services is essentially infinite, resources for health care are scarce.18 This assumption implies that decreasing expenditures on interventions that deliver limited value can be beneficial because the saved resources can be reallocated to more beneficial applications. In particular, the saved resources can be spent on health care interventions that deliver better value, thus improving overall population health. Cost-effectiveness and cost-utility analysis are methodologies that assess the extent to which reallocation of resources from one health care intervention to another can improve population health. Alternatively, resources appropriated from an inefficient health care intervention can be spent on goods and services people value outside the health care sector. Cost-benefit analysis addresses the question of whether such a reallocation is worthwhile.
The assumption of scarce health care resources is most evident in countries with a single-payer health care system because (excluding private purchases of health care outside the public system) the payer’s budget represents the national cap on health care spending. In the context of the decentralized reimbursement system in the United States, where health care is covered by a mix of public programs (e.g., Medicare and Medicaid), a huge number of private employers, and individuals, the finite nature of the resources available for health care is less obvious. Economic analysis in health care results in the unusual wedding of principles of economics and clinical decision-making—linking planning decisions on behalf of the community with clinical decisions made on behalf of the individual patient.
Components of Cost and Analysis Perspective
Before describing the different analysis methods just mentioned, it is useful to consider the different components of costs and why those components may be included or omitted from a particular economic analysis. “Costs” refer to impacts that deny resources for other uses.19 Three common categories considered are those costs incurred by the patient and/or family, costs incurred by the health care system (health care delivery institutions and payers), and costs incurred by any sector in society.
The perspective of an analysis—that is, the identity of the decision maker for whom the analysis is conducted—determines which cost categories are included. Consider a program that decreases inpatient care by having more of the care conducted at home. An insurer will be interested in the impact on health care sector costs, including inpatient hospital care costs, and the cost of professional in-home care providers, but may not be concerned with patient/family costs, including the burden on the family of providing patient care. A societal perspective analysis would take account of all of these costs. It would also consider the cost of parent absenteeism from work incurred by employers. It can be useful to conduct an analysis from different perspectives. For example, the intervention in this example could have a favorable impact on costs from the healthcare system perspective, but an unfavorable impact from the patient/family perspective. Other cost categories (and perspectives) may be appropriate, depending on what question the analysis is addressing and for what audience. For example, an analysis could limit consideration to those costs incurred by the employer in order to help a company identify preventive measures that will best promote profitability.
Patient/Family Costs
These costs include copayments and expenditures on health care goods and services not covered by health insurance. Copayments
can be substantial in the context of high-cost oncology medications. They also include time and OOP expenses associated with receiving treatment. Repeated visits to hospital facilities for chemotherapy treatments can require a substantial amount of parental time. Visits to distant institutions for treatments not available at a local facility can involve substantial travel expenses. Finally, patient/family costs include reduced salary associated with time lost at work, for example, due to the demands of caring for a child with cancer.
can be substantial in the context of high-cost oncology medications. They also include time and OOP expenses associated with receiving treatment. Repeated visits to hospital facilities for chemotherapy treatments can require a substantial amount of parental time. Visits to distant institutions for treatments not available at a local facility can involve substantial travel expenses. Finally, patient/family costs include reduced salary associated with time lost at work, for example, due to the demands of caring for a child with cancer.
Health Care Sector
These costs include those expenses borne by public and private payers, including Medicare, Medicaid, and private health care insurance companies.
Society
The “societal perspective” accounts for costs incurred by the patient/family and the health care sector (earlier). It also accounts for costs incurred by other sectors (e.g., time spent by volunteers caring for a patient). Finally, it includes economic losses suffered by employers when employee productivity is adversely affected by health conditions (net of decreased wages). Productivity losses are sometimes considered a distinct, fourth cost category.19
Other Cost Categorization Schemes
Some writers classify costs into three major categories: direct, indirect, and intangibles. In recent years, however, use of these terms has fallen out of favor, in part because they have been used inconsistently and tend to lead to confusion.19 Direct costs refer to costs affecting the health care sector, but are sometimes also used to refer to costs incurred by patients/families and by other sectors for health care sector goods and services. The term “Indirect costs” has often been used to refer to time spent by patients and their families to receive treatments, and to productivity losses. The term has led to confusion because accountants use it to refer to “overhead costs.” Finally, the term intangibles has been used to refer to impacts whose monetary value is difficult to measure (e.g., changes in health status or pain and suffering associated with treatment). These impacts, however, are not costs as defined earlier because they do not deny the use of resources for another purpose.
Types of Analyses
Whereas the analysis perspectives (and resulting determination of which costs are to be included) are driven by who is the decision maker, the options available drive what type of analysis is most appropriate (see Table 51.1).
TABLE 51.1 Types of Economic Analysis | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
Cost Analysis
A cost analysis is appropriate in cases where a decision maker is interested in understanding the resource consumption impacts of alternative interventions. Such analyses are not designed to address the appropriateness of interventions per se, but rather the cost implications of those interventions.
Two examples from large cooperative group analyses illustrate the use of cost analysis and exemplify useful methodologies. First, Green et al.,20 from the National Wilms Tumor Study Group, compared two treatment regimens for children with newly diagnosed Wilms tumor that varied by treatment duration (short, 6 months; long, 15 months). Because 4-year relapse-free survival did not significantly differ by treatment duration, the difference in costs for the two approaches became the most salient distinguishing factor. Annual total cost (medical costs, estimated from relative value units and Medicare charges) for the short duration was 50% of that of the long duration, with an estimated aggregate savings of $730,000 per annum. In a second example, Bennett et al.21 reported a cost analysis of filgrastim in children with T-cell leukemia and advanced lymphoblastic leukemia. This study was conducted retrospectively using data on resource use from participants enrolled in a Pediatric Oncology Group randomized clinical trial. The Bennett et al. study is exemplary because it estimated costs based on resource use as tabulated on study case report forms, an approach that minimized work from the cooperative group personnel. The study also identified cost drivers, a process that could be replicated in prospective analyses to focus efforts on collection of the most important cost information. Key limitations acknowledged by the authors included the lack of statistical power in the economic analysis due to the extent of variance in cost estimates.
As outlined by Drummond et al.,19 cost estimation involves a number of considerations, some of which are particularly relevant to the evaluation of interventions to address pediatric cancer. First, because parents’ care of a pediatric patient with cancer can consume a substantial amount of time, how the analysis assigns a value to parent time can influence the results. For example, time can be assigned a value based on market wages (how much a parent could earn), or the value people assign to leisure time.
Second, the amount charged for medical services may not reflect actual resources consumed as a result of the provision of care to an additional patient. Historically, institutional charges have included costs directly associated with provision of the service, costs to cover overhead (e.g., building maintenance), and a “cushion” to cover “nonrevenue” units within the facility or cover “bad debt” (charity care or unpaid claims). The Medicare and Medicaid programs traditionally set reimbursement at some fraction (e.g., 80%) of the “reasonable and customary” fee structure. Which costs are included in charges is related in part to how risk is shared between health care payers and health care providers. In a traditional “fee
for service” system, the health care payer is responsible for paying for all costs associated with the care of a patient, even when costs rise dramatically due to unforeseen events. Under a “capitated” system (e.g., prepaid health care), charges are more closely tied to a prespecified level that depends on the condition being treated. Unforeseen costs are absorbed by the health care provider (although some plans have offered oncology services as a “carve out,” based on negotiated discount). Importantly, in a capitated system, charges do not necessarily represent costs for a particular patient. Instead, charges represent average costs for a population of patients, assuming that it is possible to properly estimate the mix of patients, costs associated with typical “base case” patients, and costs associated with “outlier” patients.
for service” system, the health care payer is responsible for paying for all costs associated with the care of a patient, even when costs rise dramatically due to unforeseen events. Under a “capitated” system (e.g., prepaid health care), charges are more closely tied to a prespecified level that depends on the condition being treated. Unforeseen costs are absorbed by the health care provider (although some plans have offered oncology services as a “carve out,” based on negotiated discount). Importantly, in a capitated system, charges do not necessarily represent costs for a particular patient. Instead, charges represent average costs for a population of patients, assuming that it is possible to properly estimate the mix of patients, costs associated with typical “base case” patients, and costs associated with “outlier” patients.
Third, results depend on the analysis “time horizon,” which determines how far into the future costs are tracked. Long-term costs depend on morbidity and the proportion of childhood cancer survivors who will enjoy extended survival. Long-term care costs relate not only to cancer-related morbidity but also to noncancer illnesses for which the survivor is at risk due to extended survival. The use of a shorter time horizon (e.g., 1-year, 5-year) for which outcomes are better elucidated, although methodologically easier, ignores the long-term survival cost impacts.
A related issue is the use of “discounting” to make costs occurring at very different points in time comparable. Health economists consider the use of discounting critical in order to properly account for the general preference of people to defer costs (and to consume goods and services as soon as possible). Generally, discounting decreases the importance of costs that occur at more distant points in the future. In practice, discounting converts “future costs” into “present costs” by scaling the future costs downward by a discount factor that grows exponentially with time. For example, if the annual discount rate is 3% (a common, recommended value), a cost incurred 1 year in the future is scaled by 0.97 = 1/(1.03)1, whereas a cost incurred 20 years in the future is scaled by 0.55 = 1/(1.03)20. Use of modestly larger discount rates (e.g., 5% annually) can have a dramatic impact on cost estimates, and this impact is greater for costs that are further in the future. The use of discounting is particularly relevant for evaluations of pediatric cancer interventions due to the potential life expectancy of pediatric patients. For example, at even a 3% annual discount rate, the present value of costs incurred 70 years in the future is reduced by approximately a factor of 8.
Fourth, the cost of technology can change over time. The learning curve with new technologies or drugs may be steep, resulting in an exaggeration of cost early on. Drug costs can change substantially when they come off patent. However, full accounting of start-up costs is not routinely examined. In a recent review of 181 articles of cost analysis, only 14 (14.4%) of studies including actual cost data (97 of 181) accounted for start-up costs.
Finally, analysis results can depend on the inclusion of nonhealth costs influenced by health care interventions. For example, curing a child of fatal cancer can result in a lifetime of productivity gains. Whether nonhealth costs should be included in cost analyses (or other types of economic analysis) can depend in part on perspective (e.g., productivity impacts may be omitted from analyses conducted from the health care payer perspective, whereas they should be included for analyses conducted from a societal perspective).
A major limitation of cost analysis is that it does not consider an intervention’s benefits. For example, a more costly intervention can still be desirable so long as its incremental benefits are sufficiently large to justify its incremental costs. Cost analysis has no way to address that possibility. The remaining types of analyses all consider benefits, as well as costs.
Cost-Effectiveness Analysis
Cost-effectiveness analysis compares an intervention to a “comparator” by computing the cost-effectiveness ratio, defined to be the incremental costs of the intervention divided by its incremental benefits. For example, Kievit et al.22 evaluated a new strategy for identifying patients with a gene mutation placing them at elevated risk of colon cancer and reported that the new strategy cost 141 euros more per patient than contemporary practice. On the other hand, the new strategy boosted detection of carriers with the genetic mutation by 1.9%. The cost-effectiveness ratio for the new strategy amounted to an incremental cost of 7330 euros per additional mutation carrier detected.
Cost-effectiveness analysis can be used to compare programs whose benefits can all be measured using the same units. For that reason, it is most useful when considering alternatives for allocating a fixed budget to address a specific health care goal. For example, the analysis described by Kievit et al. could be used to compare alternative approaches for identifying individuals with important gene mutations. The intervention with the smallest cost-effectiveness ratio—that is, the lowest “price” per individual detected—is the most efficient.
Comparing cost-effectiveness ratios to identify the most efficient option makes sense only if the ratios compare interventions with the same comparator. For example, it makes sense to compare the cost-effectiveness of targeted screening for cancer based on family history and the cost-effectiveness of targeted screening based on physical symptoms only if both strategies are compared with the same thing (e.g., universal screening). For example, it would not make sense to compare cost-effectiveness ratios for these two targeted strategies if one cost-effectiveness ratio compared its strategy with universal screening and the other compared its strategy with no screening.
Cost-Utility Analysis
Because cost-effectiveness ratios can be compared only if the benefits are expressed in the same units, highly specialized measures (e.g., cost per identified individual with a specific gene mutation) can limit comparisons to a narrow collection of interventions. More general benefit measures, such as lives saved, make it possible to compare a broader set of interventions that all reduce fatalities. It may be desirable to account for other aspects of benefit, however. For example, quantifying benefits in terms of life years saved can help distinguish programs that prevent childhood fatalities from programs that prevent adult fatalities. Although the life years saved measure accounts for mortality impacts, it does not account for impacts on morbidity. For example, two medications that have the same impact on cancer survival may have very different side effect profiles.
The quality-adjusted life year (QALY) is a common measure that accounts for the impact of interventions on both length and quality of life (e.g., freedom from pain and ability to take part in normal activities). A year in perfect health has a value of 1 QALY, whereas a year with an adverse condition generally has a value between 0 and 1 QALY. The more severe the condition, the less 1 year with that condition is worth. Being dead has a value of 0 QALY. A program’s incremental benefit is the amount by which it increases the number of QALYs an individual gains by increasing life expectancy, improving average quality of life, or some combination of the two.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
