Waging the war on cancer

Eleven months later, President Nixon formalised the war on cancer with the passage of the National Cancer Act, an act designed to promote the discovery of new treatments for cancer and to encourage early detection and prevention of disease. More than three decades later, cancer remains an important cause of mortality and morbidity in the US. In 2007, cancer represented the second leading cause of death, accounting for 23.2% of all deaths, or roughly 563,000 people (Centers for Disease Control and Prevention (CDC), 2010). Today, ∼1.5 million new cases of cancer are diagnosed in the US annually (American Cancer Society, 2010).

However, the last three decades have witnessed important strides in understanding, preventing and treating the disease. Scientists have identified some of the genes responsible for cancer, discovered new chemotherapeutic and biologic approaches to treating the disease and developed new imaging techniques to detect cancer earlier. Accompanying these scientific advances have been equally important public health campaigns and educational programmes, which have led to improved screening rates and lifestyle changes (e.g. reduced smoking, increased physical activity) that help prevent the occurrence and reoccurrence of disease. Consequently, median survival rates have improved for many cancers, particularly breast cancer, colon cancer, prostate cancer and non-Hodgkin's lymphoma (American Cancer Society, 2010).

Despite these advances, however, there is significant debate over whether the war is truly being ‘won’ and, if so, at what cost. Criticisms generally centre on two closely related issues. First, there are concerns that medical progress in cancer care have predominately focused on the development of expensive treatments that only marginally prolong life and may reduce its quality (Epstein, Reference Epstein2005; Faguet, Reference Faguet2005). As intimated previously, many commentators have questioned the cost-effectiveness of recent cancer treatments and whether the high cost of such therapies is therefore warranted (Berenson, Reference Berenson2005; Faguet, Reference Faguet2005; Kolata and Pollack, Reference Kolata and Pollack2008; Shih and Halpern, Reference Shih and Halpern2008; Malin, Reference Malin2010). Critics often point to the fact that the US spends twice as much as any other country for the same overall survival results (Meropol and Schulman, Reference Meropol and Schulman2009) as evidence that value in advanced cancer care is suboptimal. Second, some argue that too much effort has been placed on developing such costly treatments, at the expense of prevention and early detection (Faguet, Reference Faguet2005). The following sections critically explore the first issue in further detail.

The price tag on progress: costs of and access to new cancer drugs

Expenditures on cancer care have nearly doubled over the last 20 years, in part as a result of the prices and rapid uptake of new agents and other technologies, including advances in imaging and therapeutic radiology (Meropol et al., Reference Meropol, Schrag, Smith, Mulvey, Langdon, Blum, Ubel and Schnipper2009; Tangka et al., Reference Tangka, Trogdon, Richardson, Howard, Sabatino and Finkelstein2010). In particular, spending on cancer drugs has risen 15% annually in recent years, with many approved medications carrying costs of $5000–$10,000 or more per month of treatment (Gatyas and Longwell, Reference Gatyas and Longwell2008; Schickedanz, Reference Schickedanz2010). Sales of cancer drugs are now second only to those of drugs for heart disease and estimates suggest that more than 40% of Medicare drug spending is for oncology agents (McNeil, Reference McNeil2007; Smith and Hillner, Reference Smith and Hillner2011). To be sure, scientific advances in oncology have not come cheap.

The growing economic burden associated with such treatments begs the question of why costs are so high. First, there are factors related to their supply and demand that contribute to high prices. Most new cancer drugs are biologics, which are more costly to produce than traditional medications and administered via infusion or direct injection, both of which involve substantial costs. The high cost of production is partly due to the escalating expense and inefficiencies of clinical trials and the time and resources dedicated to meet regulatory requirements (Rawlins and Chalkidou, Reference Rawlins and Chalkidou2011). In addition, the majority of therapies are patented and under other market exclusivities that limit competition in order to preserve incentives for research and development (R&D), thereby reducing competitive (downward) pressures on prices (Danzon and Taylor, Reference Danzon and Taylor2010).

Furthermore, as many of these drugs treat advanced disease that respond to few therapeutic alternatives, they are able to command premium prices without significantly deterring provider and patient demand (Goldman et al., Reference Goldman, Jena, Lakdawalla, Malin, Malkin and Sun2010). Neumann et al. (Reference Neumann, Palmer, Nadler, Fang and Ubel2010) and Nadler et al. (Reference Nadler, Eckert and Neumann2006) found that 67% and 78% of oncologists, respectively, endorsed the notion that all patients should have access to available and effective care regardless of costs. For many terminally ill patients, even when prices are high, demand for cancer drugs has been found to be largely inelastic, as they understandably place a high value on any additional life and are therefore willing to pay more for treatment (Goldman et al., Reference Goldman, Jena, Lakdawalla, Malin, Malkin and Sun2010).

Second, existing health insurance distortions and payment policies contribute to high costs. In principle, insurance creates a gap between the out-of-pocket prices that consumers pay and those received by manufacturers, with the difference paid by insurance. In cases where insurance coverage is insufficient to cover costs, many patients have supplementary insurance to cover the remainder or they forgo treatment if they cannot afford the co-insurance, and manufacturers often offer discounted drugs or assistance programmes for some patients without the ability to pay (Danzon and Taylor, Reference Danzon and Taylor2010). Consequently, patients and physicians have had limited incentives to demand cheaper drugs, thereby giving manufacturers little reason to constrain prices, although this may change due to the rising cost of care and growing levels of patient cost-sharing, as discussed further below.

In the case of Medicare, current payment policies for cancer drugs used under both Parts B and D benefits further encourage rising costs. With regard to the former, existing policies effectively reduce physicians’ margins,Footnote ¹ which encourages manufacturers to charge high prices, especially at initial launch, to offer larger margins to physicians and indirectly influence use of costlier, more aggressive therapies (Jacobson et al., Reference Jacobson, O'Malley, Earle, Pakes, Gaccione and Newhouse2006; Danzon and Taylor, Reference Danzon and Taylor2010; Medpac, 2010). In contrast, reimbursement for cognitive services (e.g. end-of-life counselling) is dismally low (Smith and Hillner, Reference Smith and Hillner2010). While there is little to no evidence that physicians base treatment decisions solely on potential profit, when deciding between two equally efficacious treatments, oncologists tend to choose the more expensive therapy (Jacobson et al., Reference Jacobson, O'Malley, Earle, Pakes, Gaccione and Newhouse2006; Smith and Hillner, Reference Smith and Hillner2010). For drugs under Part D, pharmacy benefit managers contracted by Medicare have limited ability to negotiate price discounts for preferred formulary placement with certain classes of drugs, including those for cancer (Danzon and Taylor, Reference Danzon and Taylor2010). In fact, as of 2010, contracting private plans are required to include all drugs for conditions that are major or life-threatening (Bach, Reference Bach2009), regardless of cost or value for money. To that end, Medicare's overall standard for coverage is whether the drug is ‘reasonable and necessary’ for the diagnosis or treatment of disease; it is not permitted to refuse coverage of a drug based on grounds of either cost or cost-effectiveness.

A further source of upward pressure on costs is the extensive use of and payment for ‘off-label’ cancer drugs and treatments within Medicare.Footnote ² While off-label use in cancer care is commonplace, with estimates indicating that 50–75% of drug or biologic therapy used to treat cancer is off-label (Soares, Reference Soares2005), use is higher in advanced cases when patients are no longer experiencing benefit from standard approved treatments, and increasingly involves expensive therapies (Sullivan et al., Reference Sullivan, Peppercorn, Sikora, Zalcberg, Meropol, Amir, Khayat, Boyle, Autier, Tannock, Fojo, Sideroy, Williams, Camporesi, McVie, Purushotham, Naredi, Eggermont, Brennan, Steinberg, Ridder, McCloskey, Verellan, Roberts, Storme, Hicks, Ell, Hirsch, Carbone, Schulman, Catchpole, Taylor, Geissler, Brinker, Meltzer, Kerr and Aapro2011). Some off-label use is supported by clinical data, although in many cases there is little supporting evidence of benefit, with accumulating evidence of harm (Giezen et al., Reference Giezen, Mantel-Teeuwisse, Straus, Schellekens, Leufkens and Egberts2008; Hecht et al., Reference Hecht, Mitchell, Chidiac, Scroggin, Hagenstad, Spigel, Marshall, Cohn, Stella, Deeter, Shahlin and Amado2009; Tol et al., Reference Tol, Koopman, Cats, Rodenburg, Creemers, Schrama, Erdkamp, Vos, van Groeningen, Sinnige, Richel, Voest, Dijkstra, Vink-Borger, Antonini, Mol, van Krieken, Dalesio and Punt2009). Nevertheless, off-label use of cancer therapies has endorsement from the National Cancer Institute (NCI) and, in some circumstances, the Food and Drug Administration (FDA). Moreover, the Centers for Medicare and Medicaid (CMS) must pay for off-label use if available medical compendia – reference guides Medicare relies on to determine which off-label uses of cancer drugs to cover – recommend use, even if such recommendations are based on scant evidence of effectiveness and safety.Footnote ³

Given that private plans generally follow Medicare's policies and coverage and reimbursement decisions, the aforementioned payment rules have a ripple effect across the whole of the health system, affecting physician treatment choices and distorting pricing on value and broader cost containment.

Third and finally, the high costs associated with advanced cancer care are also attributable to overutilisation of treatment or the utilisation of futile care. Several studies suggest that a substantial portion of the total cost of cancer care is for treatment delivered in the last months, weeks or days of life, and that much of this care is of little to no therapeutic benefit and potentially inconsistent with patients’ wishes (Lubitz and Riley, Reference Lubitz and Riley1993; Earle et al., Reference Earle, Landrum, Souza, Neville, Weeks and Ayanian2008; Zhang et al., Reference Zhang, Wright, Huskamp, Nilsson, Maciejewski, Earle, Block, Maciejewski and Prigerson2009; Mack et al., 2010). Existing evidence indicates that patient preferences for end-of-life cancer care vary widely (i.e. some patients want aggressive treatment until the very last day, others want home or hospice care with minimal medical intervention) and that existing national variations in careFootnote ⁴ largely do not reflect such preferences or differences in need (Barnato et al., Reference Barnato, Herdon, Anthony, Gallagher, Skinner, Bynum and Fisher2007; Earle et al., Reference Earle, Landrum, Souza, Neville, Weeks and Ayanian2008; Goodman, Reference Goodman2011; Goodman et al., Reference Goodman, Fisher, Chang, Morden, Jacobson, Murray and Miesfeldt2010, Reference Goodman, Fisher, Chang, Morden, Jacobson, Murray and Miesfeldt2011). Rather, regional differences in end-of-life care intensity are often guided by ‘supply-sensitive care’, resulting from uncertainty about how best to treat advanced cancer patients, practice styles of different health systems and physicians and the tendency of physicians to use the medical resources available to them, especially when it involves lengthening lifeFootnote ⁵ (Hollingsworth et al., Reference Hollingsworth, Zhang, Krein, Ye and Hollenbeck2010; Goodman et al., Reference Goodman, Fisher, Chang, Morden, Jacobson, Murray and Miesfeldt2011).

The growing price tag associated with these treatments is increasingly impacting patients and providers alike. Even Medicare beneficiaries and those with private health insurance can be faced with considerable out-of-pocket costs. Goldman et al. (Reference Goldman, Jena, Lakdawalla, Malin, Malkin and Sun2010) reported that although median patient out-of-pocket costs for new oncologic biologic therapies was only 1–2% of the total cost, the mean was 4–13% at the 90th percentile, and patient share of total cost ranged from 13% for imatinib to 51% for rituximab. For Medicare beneficiaries in particular, Faden et al. (Reference Faden, Chalkidou, Appleby, Waters and Leider2009) found that patient costs can total between $2900 and $6000 for three months of treatment to $7400 and $24,000 for an annual course of therapy. Such co-pays are particularly a concern for those with low or fixed incomes, who spend about a quarter of their annual income on such expenses (Langa et al., Reference Langa, Fedrick, Chernew, Kabeto, Paisley and Haymen2004; Berenson, Reference Berenson2005; Kim, Reference Kim2007). For uninsured and underinsured Americans, out-of-pocket costs can easily be four times higher; depending on the agent, annual therapy, for example, can soar to more than $100,000 (Faden et al., Reference Faden, Chalkidou, Appleby, Waters and Leider2009). Given that patients are likely to be taking multiple drugs, patient expenditures are presumably higher.

When confronted with high out-of-pocket expenses, some patients, particularly those without the ability to pay, may forgo or discontinue treatment (Szabo, Reference Szabo2008). Even those patients with some capacity to finance treatment may not undertake all recommended care in fear they will burden their families with unmanageable debt (Berenson, Reference Berenson2005; Kim, Reference Kim2007). Among those declaring bankruptcy for medical reasons, cancer is frequently the condition that precipitates the financial crisis (Himmelstein et al., Reference Himmelstein, Thorne, Warren and Woolhandler2009). Therefore, while increased patient cost-sharing may serve to control overall costs, there is evidence that it is associated with worse outcomes in the sickest and poorest patients, perhaps by causing lower use of necessary services (Hsu et al., Reference Hsu, Price, Huang, Brand and Fung2006). Growing patient cost burdens is also affecting physician treatment patterns. In a survey of oncologists (Neumann et al., Reference Neumann, Palmer, Nadler, Fang and Ubel2010), 56% and 84% reported that the cost of new cancer drugs and patient's out-of-pocket costs, respectively, currently influence which cancer treatment they recommend to their patients.

Evidence on the value for money of end-of-life cancer drug treatments

Given the substantial costs associated with advanced cancer drugs, their value for money is increasingly being questioned by patients, clinicians, payers and policymakers, with many observers arguing that these drugs, often used in the last months of life, are not cost-effective, in the sense of extending quality-adjusted life years (QALYs) at a reasonable price. Indeed, these treatments frequently have cost-effectiveness ratios that exceed (or well exceed) $100,000 per QALY, which is higher than the $50,000–100,000 per QALY threshold often suggested as denoting good value for money (Neumann et al., Reference Neumann, Sandberg, Bell, Stone and Chapman2000). Suggestions have been made for the threshold to be raised to take into account societal preferences and medical inflation (Ubel et al., Reference Ubel, Jepson, Baron, Hershey and Asch2003), and that cancer drugs (that increase life expectancy in the final month of life) in particular should be regarded differently than treatments that lower mortality risks only incrementally throughout an individual's entire adult life (Becker et al., Reference Becker, Philipson and Soares2007). As discussed in Chalkidou's article in this issue, the national health technology assessment body in the UK, the National Institute for Health and Clinical Excellence, recently extended its cost-effectiveness threshold for drugs aimed at end-of-life care under certain circumstances. The new policy was, in part, guided by societal preferences for the lives of terminally ill, mostly cancer, patients to be valued more than those suffering from other, potentially curable, chronic or acute conditions.

However, there is limited understanding of how these treatments are valued by different stakeholders in the US, although some evidence is starting to emerge. In a recent survey of oncologists on the value of cancer drugs, Neumann et al. (Reference Neumann, Palmer, Nadler, Fang and Ubel2010) found that they had an average implied cost-effectiveness threshold of roughly $300,000 per QALY. Interestingly, when later asked to generally define reasonable value for money, the majority (49%) indicated $50,000–100,000 per QALY, with ∼20% deeming a higher range of $100,000–150,000 per QALY as good value. A similar study (Nadler et al., Reference Nadler, Eckert and Neumann2006) demonstrated that 62% of oncologists believed that a life expectancy gain of at least two to four months justified the use of a hypothetical agent with a cost of $70,000 per year above the standard of care and another 20% believed four to six months would justify this cost.

Goldman et al. (Reference Goldman, Jena, Lakdawalla, Malin, Malkin and Sun2010) estimated patients’ willingness to pay for cancer drugsFootnote ⁶ to be equal to four times the total cost paid by the patient or his insurer. Annual net benefit ratios were slightly higher for those aged 65 years or older and female patients; no differences in willingness to pay were found by income, although there were acknowledged limitations in how income was imputed. Research by Becker et al. (Reference Becker, Philipson and Soares2007) and Slevin et al. (Reference Slevin, Stubbs, Plant, Wilson, Gregory, Armes and Downer1990) highlight that people who face imminent death, as in the case of advanced cancer patients, may place a much higher value on life-extending treatments than people for whom death is a remote risk. This may result in a willingness to spend any remaining wealth to live a few weeks longer, because the value of their wealth may lose much of its value at death.

Despite this initial evidence suggesting high valuation for advanced cancer drugs, there are qualifications to making deductions about their value for money. Importantly, oncologists and terminally ill patients may have different perceptions regarding economic value and benefit than other physicians, patients and broader society. For example, oncologists often practice in a context where grave situations in which treatments offer modest benefits are the norm and this arguably influences their perceptions of value (Nadler et al., Reference Nadler, Eckert and Neumann2006). Smith and Hillner (Reference Smith and Hillner2010) suggest this may indirectly cause oncologists to overvalue (and overutilise) costlier drugs, as they expect sufficient financial reward in return for routinely dealing with complicated treatments and the “fractured dreams and extreme suffering” of patients.

Furthermore, Nadler et al. (Reference Nadler, Eckert and Neumann2006) found that physician perceptions of economic value do not correlate highly with their own practice patterns. The majority (78%) of oncologists surveyed believed that patients should receive ‘effective’ treatment regardless of cost, but when queried whether a therapy that has been shown effective offered good value, most were unsure or disagreed, which suggests that they offer drug therapy to patients regardless of perceptions of value for money as long as the treatment is considered effective. In addition, physicians often either under- or over-estimate survival benefits and costs, indicating that some oncologists’ knowledge on the matter is lacking or that they have exceedingly high cost-effectiveness thresholds. How these findings comport with evidence that oncologists are increasingly considering costs in treatment decisions is worth further exploration.

Similar issues hold for advanced cancer patients, but are likely heightened given their lack of medical expertise, variable understanding of the risks, benefits and costs involved,Footnote ⁷ and the distress of facing death, which as stated earlier, may result in higher valuations of these therapies than would be elicited by healthy people or patients with acute or chronic conditions.

In addition, given increasingly limited resources to fund all health care, the value of cancer drugs should be considered from a broader societal perspective, which has yet to be explored in the American context, to ascertain how public valuations of these treatments might differ from cancer patients and their physicians. Determinations of their social value should be gauged against that of other interventions or policy priorities (inside and outside of health care), many of which may provide better value for money.

The road to value-based funding for and access to cancer drugs: technical, political and social challenges

It is clear that use of these drugs at the end of life often entails a complex cost-benefit trade-off for physicians, patients, payers and society alike, which will only get more challenging in the future with greater R&D activity in this area, increasingly expensive therapiesFootnote ⁸ (Schickedanz, Reference Schickedanz2010) and tighter health care budgets. In attempts to address the “value crisis in oncology” (Ramsey and Schickedanz, Reference Ramsey and Schickedanz2010), there have been recent developments towards better understanding and measuring the value of cancer drugs and advanced cancer care more generally.

In particular, in the last few years considerable attention and investment has been put towards CER. The Patient Protection and Affordable Care Act (PPACA) dedicated $1.1 billion to fund CERFootnote ⁹ and establish the public–private Patient-Centered Outcomes Research Institute (PCORI) to oversee the research. The aim of CER is to provide evidence of which interventions work in practice to allow physicians and patients, among others, to make better health care decisions. Cancer was one of the key priorities identified in early CER priority-setting processes (Institute of Medicine, 2009) and recent research suggests that there is demand for CER on cancer drugs. In particular, 79% of oncologists surveyed wanted more government research on their comparative effectiveness and 80% favoured greater use of cost-effectiveness data in coverage and payment decisions for these treatments (Neumann et al., Reference Neumann, Palmer, Nadler, Fang and Ubel2010).

While CER is arguably needed in the context of cancer drugs and other available interventions – less than half of medical care in the US is based on or supported by evidence about its effectiveness (IOM, 2007) – there are aspects of cancer drugs that introduce technical challenges to measuring their value (Mullins et al., Reference Mullins, Montgomery and Tunis2010). Due to the incentives inherent in the clinical research enterprise, randomised controlled trials (RCTs) are designed for specific purposes of regulatory approval and maximum market penetration, not to reflect the complexities of real-world patient care, where patients may have significant co-morbidities and other clinically mitigating factors, and they often do not collect economic information. Moreover, trials are rarely designed to compare a new drug against existing treatment alternatives to demonstrate relative therapeutic advantage (Smith and Hillner, Reference Smith and Hillner2010).

While these are not challenges unique to cancer drugs per se, for a variety of reasons the issues are exacerbated in the area of oncology. In cancer care, for example, clinical uncertainty in measuring comparative effectiveness leads to economic uncertainty, as for many cancers, there is no dominant treatment alternative and thus each attempt to treat incurs a separate cost, especially towards the end-of-life when additional, more aggressive lines of therapy are employed (Virgo et al., Reference Virgo, Vernava, Longo, Kissling, Lin and Johnson1995). It is also difficult to accurately capture patient preferences and assessments of impact on quality of life when, in advanced cancer, there are notable uncertainties about survival, side effects of drugs and personalised responses to care that frequently change along the course of treatment. Along these lines, existing preference-weighted quality of life measures (often used in QALY determinations) may not be sensitive enough to pick up changes in the health status of cancer patients or reflect preferences driven by specific characteristics or severity of the disease (Garau et al., Reference Garau, Shah, Mason, Wang, Towse and Drummond2010). In addition, the ubiquitous use of oncology drugs off-label makes it difficult to assess their value, as there is often insufficient or poor quality evidence from clinical trials to determine benefit, as previously discussed (Mullins et al., Reference Mullins, Montgomery and Tunis2010).

There are also political challenges to measuring value and, in particular, to applying resulting evidence in decision making on the availability and use of cancer drugs. The emotional debates preceding passage of the health reform legislation illustrated strong disagreements about what should constitute ‘value’ [i.e. whether both clinical effectiveness and costs (or other factors) should be considered], who should evaluate value and what role for CER in decision making (Garber and Sox, Reference Garber and Sox2010). Some legislators opposed using CER evidence for federal decisions about health insurance coverage, a stance supported by intense lobbying from various interest groups and dramatised with references to ‘death panels’, the dreaded R-word (rationing), and the heartless British NHS who leaves grandma to die to save money. These debates highlighted the political risk of giving PCORI even the resemblance of authority over health care decisions and, consequently, the PPACA prohibited the Institute from assuming the responsibility or authority to measure value and limited the ways in which CER can be used by Medicare (Garber and Sox, Reference Garber and Sox2010). Moreover, it barred Medicare from using cost-effectiveness as a factor in making coverage and reimbursement decisions.Footnote ¹⁰ Such provisions were largely supported by the general public (Gerber et al., Reference Gerber, Patashnik, Doherty and Cowling2010).

Research specific to advanced cancer drugs suggests oncologists concur; very few desire explicit resource allocation with government assuming a role in determining what care constitutes good value for money (Neumann et al., Reference Neumann, Palmer, Nadler, Fang and Ubel2010). While current provisions rendered CER more politically palatable and fostered stakeholder acceptance of PCORI, there is a risk that without some policy ‘teeth’, such research will have limited impact on clinical practice and costs.

Given the existing scenario where CER is restricted to more decentralised and implicit use by physicians, patients and other actors in the health system, there has been a surprising lack of discussion of how CER evidence on the value of treatments will get from ‘bench to bedside’. Simply making such evidence available will not likely be sufficient to ensure its consideration in patient care, including end-of-life treatment – that is, evidence is only as good as what you do with it (Freemantle and Drummond, Reference Freemantle and Drummond2010). First, in Neumann et al.'s study (Reference Neumann, Palmer, Nadler, Fang and Ubel2010) of oncologists, physicians reported insufficient time to consider available evidence, which may be heightened in end-of-life care when time to treat may be limited (Sullivan et al., Reference Sullivan, Peppercorn, Sikora, Zalcberg, Meropol, Amir, Khayat, Boyle, Autier, Tannock, Fojo, Sideroy, Williams, Camporesi, McVie, Purushotham, Naredi, Eggermont, Brennan, Steinberg, Ridder, McCloskey, Verellan, Roberts, Storme, Hicks, Ell, Hirsch, Carbone, Schulman, Catchpole, Taylor, Geissler, Brinker, Meltzer, Kerr and Aapro2011). Moreover, only 42% of oncologists said they felt well prepared to interpret and use such information in their treatment decisions. Second, even if one assumes that physicians and patients will have access to CER information on cancer drugs and use it in their decisions, there is evidence suggesting that the cost issue will not be adequately addressed. The most obvious reason is that as currently fashioned, CER evidence may not include the costs of therapy. Furthermore, patients are often reluctant to raise concerns regarding costs because they feel awkward addressing the topic and worry about how such discussions might affect quality of care (Alexander et al., Reference Alexander, Casalino and Meltzer2003; Kim, Reference Kim2007). Oncologists feel similarly uneasy, which stymies open discussion of costs with patients (Schrag and Hanger, Reference Schrag and Hanger2007; Neumann et al., Reference Neumann, Palmer, Nadler, Fang and Ubel2010). This discomfort may be more pronounced in the context of end-of-life treatment decisions, when it could feel that a life is being priced (Schickedanz, Reference Schickedanz2010), especially if costs have not entered into patient–physician discussions previously.

In addition, the existence of a “cancer taboo” may pose additional barriers to value-based decision making on cancer drugs. Cancer holds a privileged position – the “emperor of all maladies”Footnote ¹¹ – in US health care, as evidenced by significant public and private research investment in understanding and treating the disease, a powerful industry, the multitude of influential interest groups and national associations with considerable political and financial capital, and, of course, the human element that almost every citizen knows of someone who has suffered from or died from cancer. One only needs to envision the bevy of organised sporting events in support of various types of cancer and the ubiquitous coloured rubber bracelets that have now become a global phenomenon to understand the extent of public support for ‘fighting the war’. In the context where available CER evidence demonstrates a lack of benefit against the costs, such influence, coupled with national sensitivities towards rationing, will render saying ‘no’ extremely difficult. Inevitably, payers will fear negative publicity and potential lawsuits, policymakers will want to avoid ‘disease politics’, and the entire CER enterprise could be placed under threat.

While not specific to advanced cancer drugs, tensions between evidence, access to treatment and political and social preferences played out in the recent breast cancer screening guidelines. The guidelines, if followed, would reduce the use of mammography among women in the age group of 40–49 years; substantial opposition ensued, which lead to a provision in the health reform legislation requiring the government to ignore the recommendations (Gusmano and Gray, Reference Gusmano and Gray2010). The experience of the UK, as discussed by Chalkidou (Reference Chalkidou2012), further underscores the difficulty in balancing such tensions. Even in a long-standing and centralised system of CER via National Institute of Health and Clinical Excellence, where the use of evidence to support access decisions is well known and formally integrated into policy, the Institute has faced significant negative media attention and public backlash to ‘no’ decisions for advanced cancer drugs, which arguably contributed to its change in policy for end-of-life treatments, the promulgation of a dedicated cancer fund and discussion of potential removal of the Institute's decision-making authority among government officials.