This article has been excerpted from A New Deal for Cancer: Lessons from a 50 Year War edited by Abbe R. Gluck and Charles S Fuchs. Copyright © 2021. Available from PublicAffairs, an imprint of Perseus Books, LLC, a subsidiary of Hachette Book Group, Inc. 

Essential Underreimbursed Services

The fear of getting cancer, enduring cancer treatment, or dying of cancer maintains a grip on our national psyche that exceeds even the magnitude of cancer’s public health impact as the second- leading cause of death in the United States. Taken together, the many forms of human cancer are also among the most common illnesses encountered in Western society. One-quarter to one-third of women and one-third to one-half of men will receive a cancer diagnosis in their lifetime.1 It is thus not surprising that cancer care is a major driver of the overwhelming costs of health care. Indeed, the cost of cancer care has been growing considerably faster than the overall cost of health care. By some estimates, national expenditures for cancer care could climb from about $100 billion to nearly $150 billion in the next eighteen months. Any approaches taken to control overall health-care costs must address the best ways to contain cancer care costs while ensuring access to high-quality care at the cutting edge of rapidly evolving treatment options for all.

Several factors drive the accelerating costs of cancer care. Cancer has transitioned from an acute disease lasting a few months to a few years after diagnosis to a chronic disease, whereby a definitive cure is not often attained but life can be prolonged for many years or even decades. This has been a salutary accomplishment, but it comes with the continuing costs of ongoing monitoring and treatment. High-cost technologies, such as proton beam therapy, intensity-modulated radiation therapy (IMRT), and other contemporary precision radiation therapy techniques; complex imaging modalities, such as positron emission tomography-computed tomography (PET-CT) scanning; and the increasing indications for and use of stem cell transplantation, among others, have become the standard of care essential for both curative and palliative management. During the past twenty-five years, biotechnology has also enabled development of an array of more effective and less toxic but highly expensive drugs. These have become mainstays of treatment for many of the more common cancers. In addition, some providers, patients, and families feel the need to exhaust every means to “fight to the very end” even when the cancer has progressed to an untreatable state. This constellation of factors, complicated by perverse financial incentives and disincentives, tends to promote use of expensive drugs and procedures even when they are likely to be futile. In a very real sense, the partial successes of high technology– driven cancer care in prolonging survival without eliminating the cancer have created multiple structural drivers of rising cancer costs. The challenge for policy makers is finding the opportunities within this rapidly evolving ecosystem for true cost savings without compromising the best outcomes presently achievable or slowing progress toward better outcomes in the future.

While there is general agreement that the pace of growth in the costs of cancer care cannot be sustained, there is little consensus about what costs to target and how, or what strategies to pursue, to maximize quality and enhance positive outcomes at a sustainable cost. Among the most frequently mentioned and targeted objects for reduced costs are the high-priced cancer drugs and the overuse of high-technology surgery, radiation therapy, imaging, and diagnostic testing. Many view slashing these costs as a key component of reducing the overall cost of cancer care. If we are to sustain high-quality cancer care for all who need it, we must develop effective strategies and policies that will slow or even reverse the rising costs of these expensive technologies and therapies to the system, the nation, and the patients.

The premise of this chapter is that cost- and price-reduction strategies and policies must be developed holistically, accompanied by reimbursement reform that recognizes that many high-value and essential services for cancer patients are underreimbursed. The ability to sustain these services, as well as other essential contributions to the national ecosystem of health care, such as workforce development, public health outreach measures focused on prevention and early detection, and research, currently depends heavily on the phenomenon of “cross-subsidization,” whereby positive operating margins derived from highly reimbursed goods and services offset operating losses incurred by the delivery of underreimbursed goods and services. Unless policy makers consider the impact of targeted price reductions on this ability to cross-subsidize, critical services could be cut. Price cuts must thus be mitigated by compensatory increases in payment for these critical underreimbursed and thus cross-subsidized services. This correction would necessarily reduce the projected savings in the overall costs of cancer care compared to current projections that count only the reduced expenditures for drugs. This consideration does not relieve our collective obligation to achieve cost reductions for drugs, tests, and procedures. Rather, it delineates the need to achieve that goal in the context of what the current arrangement pays for in the actual settings where care is delivered. 

The remainder of this discussion contextualizes cancer care in the United States in light of our overall health-care costs and outcomes. In particular, we explore how the changes we outline above might affect the ability of comprehensive cancer centers to contribute to the national cancer effort and suggest how their current contributions could be preserved.

The Nation’s Cancer Care in the Context of the Nation’s Health-Care Dilemma

Medical care expenses in the United States far outpace those in other countries in the developed world. Yet, major health outcomes are generally worse despite the higher cost borne by American society.2 The role of cancer care in this cost escalation is complex. On the one hand, as noted earlier, the rate of rise in cancer care costs outpaces the overall trends; on the other hand, outcomes for cancer care in the US actually compare favorably with other developed countries.3 This stands in contrast to many other areas of medicine. Cancer care in the US needs a cost-reduction strategy that preserves performance in areas that contribute to our relatively good outcomes and recognizes the resources needed for improvements in access and performance in other areas, while fostering a sustainable investment in the nation’s efforts to eradicate the fear and suffering caused by cancer.

Many remedies have been proposed to decelerate or even reverse the worrisome growth of overall health-care costs, both absolutely and as a percentage of the annual gross national product. These remedies include replacing the traditional fee-for-service forms of reimbursement with “value-based” algorithms that tie reimbursements to certain pre-agreed outcomes deemed to improve the quality of the care delivered. Within health policy, value is often defined in terms of the degree to which the processes and outcomes are favorable to patients, society, and payers; innovations that improve cost efficiency while maintaining quality are rewarded as well. Alternatively, some health-care stakeholders recommend massive shifts in the reimbursement formulae away from current algorithms that favor procedures, high-cost diagnostics, and pharmacy charges, and toward important underreimbursed services, such as “evaluation and management” (that is, the charge for a visit with a provider and the provider’s time), preventive health measures, palliative and end-of-life care, hospice care, and mental health services. Finally, still other stakeholders propose blanket, prepaid health-care formulae whereby the providers and provider institutions receive flat monthly or annual fees per patient for the management of certain populations, for example, hypertension or diabetes patients. Various pilot programs sponsored by the government, health insurance companies, and the Veterans Administration have been tried with varying, but usually limited, degrees of success. An imperative to move the financial risk of care onto providers, be they institutions or individual practitioners, underlies many of these proposals, while supplying safeguards intended to prevent compromises in the quality or quantity of services delivered for financial reasons.
The challenge for many of these “population medicine” strategies with regard to cancer care is that cancer is a not a single disease within which one can organize the very large populations needed to define the cost and services calculations that yield predetermined per-patient or per-episode reimbursements. Cancers are highly heterogeneous, comprising over five hundred distinct types,4 with diverse physician, nursing, hospital, and outpatient needs for optimal care. Moreover, recent advances in dissecting the molecular roots of cancers in individual patients have revealed even greater heterogeneity. For example, molecular analysis has uncovered over a dozen forms of breast cancer and several forms of non-small-cell lung cancer (NSCLC),5 each dictating the use of distinct therapeutic modalities and agents. This “precision oncology” paradigm is spreading rapidly to many other forms of cancer. It has created an impediment to establishing population-based reimbursement paradigms that must balance the rapid shift toward highly individualized treatment regimens with the desire to define coherent populations upon which to calculate appropriate reimbursements and allocation of risk. The costs of individual goods and services in a fee-for-service setting may thus continue to be targeted in the cancer domain, even as shifts to population medicine take hold in the overall system.

The Issue of Drug Costs and Drug Prices

There is no question that the reimbursement system in the United States is broken or that it creates perverse incentives for overutilization by both patients and providers. It is heavily skewed toward procedures (e.g., surgery or interventional radiology) and testing (e.g., magnetic resonance imaging [MRI] or PET imaging), and away from some of the key parts of medical intervention, such as evaluation and management of the patient’s condition and supportive care of patients and their loved ones.6 Within the health-care industry, among patient advocacy groups, and in the political arena, attention has focused on the rising prices of highly reimbursed pharmaceuticals. Both policy makers and public opinion prominently implicated high drug prices as a major cause of the American health-care cost disaster and a prime target for correction of the overall problem. Drug costs for general health care make up 17 percent of the total health-care “bill” and are rapidly rising.7 For cancer care, the percentage is somewhat higher and rising faster.8

There are several reasons the prices of new drugs are rising at the alarming rate that has attracted the attention and ire of the public and policy makers. One is that the search for more effective drugs now depends heavily on applying complex scientific experimentation and technology to identify more precisely targeted, efficacious, and safe agents. A consequence of this is that the “market size” for any single drug is much smaller, useful only for those in whom the target drives the disease. For example, non-small-cell lung cancer, an extremely common cancer, and the major cause of cancer deaths, can no longer be regarded as a single diagnosis. Rather, the availability of molecular technologies to characterize the tumors in individual patients has revealed that there are many subtypes of NSCLC driven by myriad genomic alterations.9 These create a variety of targets for therapy, but each treatment is useful only in those patients having that target. This population can comprise as little as 1 to 2 percent of the general lung cancer population.10 Thus, these drugs are useful only to much smaller groups of patients and development costs must be amortized over these smaller groups. A second driver is the methodology for financing new drug development. A potential drug candidate often starts with a discovery in an academic laboratory that is then patented. Start-up companies are founded on the basis of this intellectual property and the promise that the candidate will ultimately achieve sufficient utilization in clinical practice to generate a return on investment. These start-ups are supported at the earliest high-risk stages by multiple rounds of financing from angel investors, venture funds, etc. Often, such start-ups enter the public market, sometimes long before a single agent has received FDA approval. The total private and public market investments in this process frequently amount to hundreds of millions or even billions of dollars before the product reaches patients and generates revenue.11 Those investing in these endeavors demand extremely high rates of return, on the order of six- to tenfold, on individual drug successes, so as to offset the failure rate for new drug development, which can be close to 90 percent of all drugs entering Phase 1 clinical trials.12 Drug development is also a prolonged process. Patents issued at the time of the discovery of a target, compound, or application sometimes have a limited post-approval life span.13 Investment costs must, in these cases, be recouped over a shorter time frame. 

Despite these acknowledged drivers of the high cost of drugs, public outrage over the perception that pharmaceutical companies charge overinflated prices for these agents persists, driven in part by the companies’ annual profit margins and stock performance. Moreover, the price of drugs in the United States is aggravated by the fact that many other countries regulate drug prices or permit government-run health systems to negotiate drug prices with the industry. Public and political opinion holds that both drugs and drug development are provided to the public much less expensively outside the US. Regardless of the reasons for the rapidly rising costs of drugs, there is an unquestionably strong public and political momentum for reducing the costs of pharmaceuticals, as an at least partial solution to the overall excessive costs of American health care. While a few practical approaches have been proposed (for example, permitting Medicare and Medicaid to negotiate directly with drugmakers, allowing importation of drugs from Canada at their price points, etc.), it seems increasingly likely that concrete proposals will come either in the form of legislation or Medicare/Medicaid policy and regulations.

Drug Costs, Drug Reimbursements, and Comprehensive Cancer Centers

Cancer centers, particularly the National Cancer Institute’s fifty- one designated Comprehensive Cancer Centers (CCCs), remain wary of a sudden reduction in drug prices.14 These centers are the nation’s hubs for highly specialized and subspecialized cancer care; education and training of the national cancer care workforce; basic, translational, clinical, and population-based research; and community outreach for prevention and early detection. For these single-specialty hospitals, the margins on cancer drugs are high, particularly for some of the newer agents that are based on precision medicine and therefore given to select smaller groups of patients. The margins (i.e., net revenues) for drug charges are among the highest that CCCs receive for any of their activities, because of purchase-price markups that are permitted under existing formulae. This is especially important to the sustainability of the CCCs, even those with robust inpatient and surgical activity, because cancer care has shifted significantly to the ambulatory setting, where reimbursements for most health-care services tend to be lower.15 The arc of cancer care is now associated, fortunately, with longer survival, but often in the setting of continued presence of the cancers, making cancer more of a chronic disease akin to diabetes or hypertension, requiring intensive but less well reimbursed ambulatory care.16 The responsibility for palliative care and end-of-life care falls disproportionately on CCCs, even in general health-care systems.17 While imaging, surgery, and certain testing do provide positive reimbursement margins, cancer care drugs have become an important component of net revenues available to cross-subsidize these and other critical activities of the CCCs.18

The true costs and value of cancer care are thus mismatched with the reimbursements for services, tests, and procedures in such a way that slashing reimbursements for drugs, procedures, and infusions might reduce expenditures, but only at the expense of the provision of a broad array of high-value, moderate-cost services. The latter are currently cross-subsidized because they are underreimbursed or nonreimbursed. Abrupt reductions in drug prices without compensating improvements in reimbursement for money- losing services could have a substantial negative impact on CCCs and the national cancer care effort. These mismatches have profound effects on the operations and sustainability strategies of both large cancer centers with substantial investments in research and training, and smaller cancer centers that are often embedded in communities with lesser access to economies of scale and other opportunities for cross-subsidization. These challenges inevitably influence the utilization of services, the competition for privately insured patients, and the outcomes of cancer care in the United States. 

The foregoing concerns are amplified by the fact that, in addition to delivery of highly sophisticated subspecialized care for all cancers, including rare and particularly complex cancers, the CCCs are the nation’s primary generator of the cancer care workforce, including oncology-trained physicians, nurses, pharmacists, social workers, research scientists, and so on.19 They are the major national hubs for the performance of clinical research, particularly the critical and highly complex early-stage “first in human” clinical studies that determine the viability of potential new drug candidates.20 They engage, frequently with little or no reimbursement, in major community outreach activities that are critical to improving national cancer health outcomes, such as mammography screening, human papillomavirus (HPV) vaccination, colorectal cancer screening, and malignant melanoma cancer prevention efforts, among others. Finally, CCCs comprise by far the largest and most robust hubs for basic, early, and late-stage translational, applied clinical, and population research in the nation. Efforts to recoup the costs of these essential services are necessarily fragmented, expensive, and invariably incomplete: in other words, the financial resources sufficient to offset the structural fixed costs of these highly expensive activities at CCCs are already precarious.

The Need to Cross-Subsidize Might Limit Savings from Reduced Drug Costs

It is widely accepted throughout the American health-care system that provider institutions must cross-subsidize to survive under the current reimbursement formulae.21 For example, emergency services, mental health services, primary care clinics, addiction medicine services, and so on are reimbursed at or below actual cost and sustained by cross-subsidization from highly reimbursed services such as pharmaceuticals, surgery, and imaging.22 Reliable statistics on the extent and intensity of cross-subsidization as a survival mechanism are hard to come by, but at least one study shows the negative impact that loss of highly reimbursed activity can have on the provision of less-reimbursed services by an institution.23 In this study, a hospital providing a comprehensive array of services was operating in a health-care market into which an entity focusing solely on highly reimbursed cardiovascular services entered. This entry resulted in the loss of this activity by the generalist institution. In response, the institution was forced to reduce its provision of underreimbursed services, such as mental health services, that had been made possible by the margins from cardiovascular medicine.24 The complexities of cross-subsidization are also at work in CCCs.

Second opinion consultations are a vital service performed by CCCs. One major cancer center, MD Anderson Breast Center, felt that it could no longer meet the growing need for these consultations and temporarily curtailed them. MD Anderson stated, among other reasons, that the cost of providing second opinions was too high, compared to the reimbursement recovered.25 This policy was later rescinded, but not because second opinions became more lucrative; rather, the CCC simply accepted higher operating losses from this service line. In CCCs, pharmaceutical prices provide significant amounts of the margin available for this form of cross-subsidization.

The preceding discussion makes apparent one aspect of the discussions around pharmaceutical reimbursement that extant conversations often do not recognize. That is, what patients and society are paying for when reimbursing a CCC for a cancer drug is not simply the all-in costs of the cancer drug itself and the indirect costs associated with storing, handling, and dispensing the drug, and so on, but also a significant contribution to the broad array of comprehensive clinical services, public health activities, education, training, and research that make CCCs important national resources. It follows that abrupt reductions in the reimbursements for cancer drugs will need to be balanced by recognition, either by reimbursement reform or subsidies, of the essential services that could be negatively impacted by a sudden loss of substantial revenue from pharmaceuticals. The need to account for those no longer cross-subsidized costs implies that the savings anticipated by reduced drug prices might thus be lower than anticipated. The alternative could otherwise be a reduction in vital services.

Reducing Drug Prices Must Be Accompanied by Other Payment Reforms

There is no doubt that drug prices are escalating far too rapidly. The trend needs to be significantly slowed and reversed. This chapter argues that the reduction in drug prices needs to be accomplished holistically, with anticipation of unintended consequences. At least in the case of cancer drugs, one should not expect to recoup the total reduction in drug costs in the form of health-care savings, unless one is prepared to forgo in part or completely the availability of other critical services both to patients and to society that CCCs perform. The goal of any solution should be to rationalize the reimbursement system so that the prices charged for drugs are reduced to levels appropriate for the costs of their development, production, utilization, and monitoring. The portion that results in net revenue loss to provider institutions should be mitigated by implementing appropriate levels of reimbursement for services that now operate at a loss. In other words, the price of goods and services should directly reflect their cost and value, both to patients and society.

Another approach to reducing the cost of cancer drugs to consumers and society is to reduce the cost of development. The chances of an early drug candidate that is matched to a promising physiologic target, such as a particular cancer mutation, progressing through the long and complex process of drug development, testing, and approval are about 5 to 10 percent.26 Much of the expense associated with that high failure rate occurs because the knowledge and technology are not always available to advance those agents with a higher chance of success or eliminate those unlikely to succeed at earlier stages of development.

Given that situation, a national priority should be to de-risk 

some of the earliest stages of pharmaceutical research, thus reducing a portion of the enormous but necessary early-stage investments in companies or development programs. This would involve creating funding mechanisms, whether governmental or private, that would pool funds and drug development candidates, subsidizing rigorous interrogation of many therapeutic targets and many approaches to manipulating promising targets at the earliest, least expensive stages of drug development. In that manner, drug candidates destined for failure would be eliminated early. Only the most promising candidates would need to be supported by investment in the later, more expensive stages.27 The return on agents that do succeed would offset the costs and risks of evaluating the broader array of early candidates. Mechanisms to achieve this, such as those proposed by Dr. Andrew Lo and colleagues, are under investigation.28


Cancer care and research in the United States continue to improve both the quality of care and the quantity of positive outcomes. At least some of these gains are due to services that are underreimbursed but supported by cross-subsidization from more highly reimbursed goods and services, including cancer drugs. However, the rate of rise of overall cancer care costs is unsustainable. Innovative mechanisms to reduce costs are essential if cancer care is to remain anything close to affordable for patients, the government, insurers, and employers. Managing the rate of rise of cancer drug prices must constitute a significant part of policy discussions involving such mechanisms. Such discussions must take into consideration the consequences, intended and unintended, of abrupt changes in drug-related reimbursements to providers. Indeed, the savings from reducing drug prices will ultimately be less than one might predict if one takes into account the indispensable services delivered by cancer care centers that pharmacy revenues cross-subsidize. Compensatory increases in cost recovery for currently underreimbursed services and underfunded activities must accompany reduced drug prices. In taking such a holistic approach, the United States can still anticipate a salutary reduction in overall health-care costs, as providers will be able to increase the availability of services, such as preventive cancer medicine and palliative care, which are known to improve patient outcomes and satisfaction, reduce the cost of care, and prolong high-quality survival.