Desperately seeking cancer drugs: explaining the emergence and outcomes of accelerated pharmaceutical regulation
Address for correspondence: Courtney Davis or John Abraham, Department of Sociology, School of Law, Politics & Sociology, University of Sussex, Falmer, Brighton BN1 9QN.
e-mail: C.M.Davis@sussex.ac.uk or J.W.Abraham@sussex.ac.uk
Government regulators have increasingly accelerated new cancer drugs on to the market by granting them approval based on less clinical data supporting drug efficacy than permitted under standard regulations. With more lenient regulatory standards, pharmaceutical companies have keenly sought to develop cancer drugs. Focusing on the US, this article examines how the emergence and implementation of such accelerated approvals should be understood, particularly in relation to corporate bias and disease-politics theories. Drawing on longitudinal and case study data analysis, it is argued that the emergence of accelerated approval regulations for cancer drugs should be regarded primarily as part of a deregulatory regime driven by the interests of the pharmaceutical industry in partnership with all major aspects of the state, rather than as a response to patient activism in the aftermath of AIDS. Furthermore, even in cases when some patients successfully demand accelerated marketing approval of cancer drugs, such approval by regulators, while in manufacturers’ interests, may not be in the interests of patients’ health because the political culture of the regulatory agency is reluctant to uphold its own techno-regulatory standards of public-health protection when that would challenge the agenda-setting influence of manufacturers, including industry collaborations with patients and the medical profession.
Patients struggling desperately against state bureaucracies and regulations to gain rapid access to newly developed drugs to treat serious or life-threatening conditions have been a dominant narrative in many media depictions. Yet rarely does that narrative pay sufficient attention to crucial sociological factors, such as the role of the drugs’ manufacturers, the nature of the state’s mechanisms for accelerating drug approval, and the sort of evidence available that the drugs work. In this article, we aim to improve on this narrative by explaining the emergence and implementation of accelerated approval regulations in the US. We argue that the evolution of American government regulation of new drugs for serious or life-threatening conditions should not be understood as a state reluctant to deregulate market access in the face of patients’ demands, but rather as a consistent trend of deregulation to assist the pharmaceutical industry. We also present evidence explaining why regulatory outcomes of accelerated approvals are in the interests of drug firms, but may not be in the interests of patients, even when demands for such approvals by some patients are evident and influential.
The research problem and theoretical context
Internationally, the ‘standard’ process of drug regulation comprises laboratory and animal tests followed by three sequential phases of human trials before government agencies grant new drugs approval on to the market (known as marketing approval). After Phase I trials with a small number of healthy volunteers to ascertain drug safety/toxicity in people, Phase II involves medium-sized clinical trials, sometimes placebo-controlled, to collect preliminary evidence about drug safety and efficacy in patients intended for treatment. Phase III studies are larger, randomised controlled clinical trials (RCTs) whose purpose is to provide statistically significant evidence of a drug’s effects. Phase III studies are the most costly component of drug development for manufacturers (OTA 1993: 56). Clinical trials conducted after a drug is on the market (post-marketing), are known as Phase IV.
Normally the US drug regulatory agency, the Food and Drug Administration (FDA), requires at least two ‘pivotal’ Phase III RCTs to demonstrate drug efficacy before marketing approval. Typically this involves evidence of direct clinical efficacy (on symptoms or survival), but occasionally approval is based on a drug’s effects on an established surrogate marker/endpoint, that is, a laboratory/physical measure substituting for clinical efficacy. For example, a drug’s capacity to lower blood pressure is established (among medico-scientific/regulatory communities) as a valid predictor, and hence surrogate marker, of clinical efficacy to prevent strokes, albeit with a history of medicalisation (Kawachi and Conrad 1996).
In 1988 and 1992, the FDA introduced regulations to speed marketing of drugs intended to treat serious or life-threatening conditions and expected to address an ‘unmet need’. These (non-standard) regulations, known as ‘Subpart E’ and ‘Subpart H’ (Federal Register (FR) 1988, 1992), enable the FDA to approve such drugs with less data than normally required to support clinical efficacy. Specifically, approval may be granted without Phase III trials, with just one Phase II trial (Subpart E), and/or using non-established surrogate endpoints in trials ‘reasonably likely to predict clinical benefit’/endpoints, but not demonstrated to be valid substitutes for clinical endpoints (Subpart H). Drugs approved under Subpart H, based on non-established surrogates, are referred to by the FDA as ‘accelerated approvals’. Upon receiving accelerated approval, drug manufacturers must conduct post-marketing (Phase IV) confirmatory studies to demonstrate drug efficacy according to standard approval requirements. If such studies are not forthcoming from manufacturers with ‘due diligence’, the FDA is expected to remove the drug from the market.
These American accelerated approval regulations are the focus of this paper, with particular reference to cancer drugs. This is not irrelevant to the UK or Europe because, with the largest pharmaceutical market and regulatory agency, the US often influences developments elsewhere. Since 1993, and especially 2004, the EU introduced similar regulations allowing the European Medicines Evaluation Agency (EMEA) to approve marketing applications lacking comprehensive data on drug efficacy in ‘exceptional circumstances’ (EC 1993, OJEU 2004).
Our analysis is important because, between 1992 and 2008, the FDA approved 64 new drugs under accelerated approval regulations based on (non-established) surrogate endpoints – about 15 per cent of new molecular entities approved on to the US market (US GAO 2009: 15). Similarly, in the EU, between 1995 and 2000, 18 of 126 marketing approvals (14%) were accelerated under ‘exceptional circumstances’ (Garattini and Bertele 2001). Furthermore, a large and increasing proportion of accelerated approvals are cancer drugs. Forty-two per cent of accelerated approvals granted by the FDA based on surrogate endpoints from 1992 to 2008 were cancer drugs, rising to 59 per cent for 2001–2008 (US GAO 2009: 15–17). Meanwhile, in the EU, 48 per cent of cancer drugs were granted marketing approval based on non-established endpoints between 1995 and 2004 (Apolone et al. 2005).
Although pharmaceuticals is a growing field of study within sociology, the health significance of accelerated drug approvals has been neglected (Conrad 2007, Frasier et al. 2009, Williams et al. 2009, Abraham 2010). Scholars have examined drugs that rarely, if ever, receive accelerated approval, such as anti-inflammatories, anaesthetics, antidepressants, asthma medications, beta-blockers, psychotropics, and stimulants (Abraham 1994, Gabe and Bury 1988, 1996, Pearce 1996, Abraham and Sheppard 1999, Busfield 2004, Abraham and Davis 2005a, 2005b, 2006, 2007, 2009, Rafalovich 2005, Williams et al. 2008). Moreover, greater understanding of regulatory initiatives to expedite market access to drugs might imply that existing sociological theories of pharmaceutical regulation need modification.
The two main relevant theories of pharmaceutical regulation suggest that different actors and interests might drive regulatory developments and decisions pertaining to accelerated approval. According to Abraham, who has researched standard drug regulation in the UK, EU and US from the late 19th century, the pharmaceutical industry is the most important influence on regulatory change either by working in partnership with the regulatory state to advance joint government-industry interests from the outset, or by diverting regulatory agencies away from policies protecting public health that might conflict with industry’s commercial interests (Abraham 1995, Abraham and Lewis 1999, 2000, 2002). On this view, which we call ‘corporate bias theory’, the state typically prioritises the interests of the pharmaceutical industry over and above other interests both in the direction of regulatory change and the implementation of new regulations. How well that theory applies to non-standard regulation involving accelerated approval of drugs to treat life-threatening illnesses has not been investigated until now.
By contrast, in his study of 20th-century US drug regulation, Daemmrich argues that it is shifts in patients’ perspectives that drive regulatory change, including accelerated approvals:
During the 1980s and 1990s … [t]he American ‘patient’ evolved from needing state protection from industry and physicians to a free-market consumer who deserved access to still-experimental drugs. As a consequence, FDA placed fewer demands on manufacturers for lengthy testing and redesign of clinical trials than in the past (2004: 81).
For proponents of this view that patients and ‘disease-based’ patient groups drive regulatory change, the AIDS crisis of the 1980s is pivotal. For example, Edgar and Rothman (1990) assert that, before the mid-1980s, the FDA’s ‘heavy-handed paternalism’ was ‘heavily biased in favour of caution’ about approving new drugs and denied patients the right to their own risk-benefit calculus about pharmaceuticals. AIDS patient activism, they argue, reversed this. Similarly, Daemmrich claims that it was ‘largely the AIDS movement which led to procedural changes in US drug regulation’, including accelerated approvals (Daemmrich 2004: 30–2, 98, Daemmrich and Krucken 2000: 514, 523, 529). Reinforcing this perspective, and sharpening its distinction from ‘corporate bias theory’, Carpenter (2004: 52) declares: ‘Patients, more than pharmaceutical firms, shape the political costs to FDA of delaying drug approval’.
Furthermore, these theorists imply that post-AIDS regulatory changes toward accelerated approvals, are not only driven by patients’ demands, but they are also in patients’ interests, and that AIDS activists’ exposure of the ostensible adverse consequences of the FDA’s over-cautious approach before AIDS made government recognise the ‘legitimacy of demands for more rapid drug approvals’, vindicating deregulatory criticisms of the FDA made during the 1970s and 1980s (Daemmrich and Krucken 2000: 512–19).1 Thus, from this theoretical perspective, which we call ‘disease-politics theory’, the state’s introduction of accelerated approval regulations in the US was driven by demands from patients and it was a regulatory development that was in patients’ interests.
Corporate bias theory and disease-politics theory are not entirely incompatible. Nevertheless, the differences in emphases regarding the dominant actors and interests served are stark and unmistakeable. For example, both theories acknowledge that there has been a growth in alliances between pharmaceutical firms and patient groups in the last decade. However, disease-politics theorists assert that this is a strategic response by firms driven by the increasing independent power of patient groups to achieve regulatory decisions favourable to patients, whereas corporate bias theorists are sceptical that the apparent power of patient groups should be treated as a variable independent of industry interests (Carpenter 2004: 52–5, Abraham 2009: 962–5).
In exploring the plausibility of these theories with empirical evidence about US accelerated drug approval, we are specifically concerned with the following sociological questions. Is the emergence of regulations to accelerate approval of cancer drugs best explained by the influence of industry interests, or a new culture of patients’ demands, or both in equal measure? Irrespective of who drove the emergence of accelerated approval regulations, how are those regulations being implemented, with respect to the interests of industry and patients? Even when patients’ demands for accelerated approvals are part of the regulatory decision-making process, should it be assumed that those demands are in patients’ interests, rather than industry interests?
Adopting a realist meta-theoretical stance, we distinguish between people’s desires/demands and health interests. Such realism entails that clinical/scientific evidence need not be merely social constructions by scientists, but can partly reflect objective reality about people’s bodies, such as tumour growth or death (Abraham 2008). That reality about the scientific efficacy of pharmaceutical products can have a bearing on the health interests of patients independently of the discourse of some patients’ demands. Consequently, it must form part of our sociological analysis of the dynamics of accelerated drug regulation.
First we investigate the origins of expedited development and regulatory review initiatives for cancer drugs. That investigation is mainly concerned with whether such regulatory changes occurred primarily in response to demands from patients or industry/industry-supporters. Empirical analysis of this employs longitudinal documentary data analysis so that the role and influence of various actors can be adequately characterised over time. Then we explore the implementation of accelerated approval regulation primarily to examine whether, and how, it operates in the interests of patients or industry in the presence of patients’ demands. A case study analysis is necessary to do that in order to capture the nature, depth, and scope of evidence and actors involved in the regulatory decision-making process, and relate them to decisions taken.
Case study analysis has limitations regarding generalisation, but these can be mitigated by choosing a case strategically with respect to the theoretically-driven research questions. Accordingly, chose the AstraZeneca drug, gefitinib (brand-name Iressa), as a case study for several reasons. First, it is approved as third-line treatment for patients with non-small cell lung cancer who have exhausted all approved pharmaceutical treatment options. Effective third-line treatment is their last hope from medico-pharmaceutical intervention against a highly lethal condition. If accelerated approval regulations for cancer drugs operate in the interests of patients, then one would expect to find that manifested most in a context where patient need for access to treatment could scarcely be more urgent. Second, patients’ demands were influential in accelerating Iressa on to the market, rather than absent from the regulatory process as might be the situation in other cases. And third, the manufacturer did conduct confirmatory studies of Iressa in a timely manner. At the time of our fieldwork, that was true for only 25 per cent of cancer drugs granted accelerated approval by the FDA (Roberts and Chabner 2004, Fleming 2005). One would expect accelerated approval regulations to serve patients better in those cases (like Iressa) where attempted confirmation of drug efficacy is forthcoming in a timely manner, rather than in the 75 per cent of cases where it has been neglected. For these reasons, the Iressa case presents sociological conditions under which one would expect disease-politics theory to shine more than with most other accelerated cancer drug approvals, if the theory is sound.
We report part of a five-year research project spanning 2003–2008, including several years of US fieldwork. All relevant government documents, including those from the FDA, Congress and the Executive, from 1980–2009, were scrutinised and coded, as were publicly available materials from the pharmaceutical industry, specific manufacturing firms, public health advocacy groups, patient organisations, litigation, and media coverage. This was complemented with 45 tape-recorded and coded semi-structured interviews (cited anonymously) with ‘informants’ from these organisations. The interview data were read comprehensively, not merely electronically searched, by both authors. On reviewing the interview data a second time, they were analysed along five main dimensions: ‘drug’ (e.g. Iressa); ‘theme’ (e.g. ‘accelerated approval’); ‘factual contribution’ (e.g. ‘timing of a clinical trial’); ‘policy/value judgement/implication’ (e.g. ‘approval should have been delayed’); and ‘affiliation of informant’ (e.g. ‘the FDA’). Data analysis was checked by both authors independently. The research met all ethical requirements of the UK Economic and Social Research Council (ESRC), who funded the research.
The origins and emergence of expedited drug development and review
Proponents of disease-politics theory contend that AIDS activists ‘succeeded in doing what earlier critics of the FDA were unable to do’ by expediting patients’ access to needed new drugs (Edgar and Rothman 1990: 137). For example, Daemmrich and Krucken (2000: 515–16) acknowledge that in the mid-1980s, President Reagan’s pro-business Task Force on Regulatory Relief pressed for expedited FDA drug approval, but imply that only after patient activists joined forces with the Task Force did FDA procedures change. However, an analysis of regulatory reform initiatives throughout the 1980s to early 1990s provides evidence that FDA management was responsive to industry and government pressure to reduce data requirements and shorten regulatory review times for companies’ new drug applications, and that shifts were occurring before demands by AIDS activists.
In 1981 Reagan was elected President of the United States and the Republicans gained control of the Senate against the background of a large federal budget deficit together with reported deterioration in the international competitive position of US industries (NAS 1983: 1, Hilts 2003: 210). Believing that America’s economic and industrial decline resulted from excessive government interference in the private sector, the New Right within the Reagan Administration and Congress declared itself committed to radical deregulatory agendas (Hilts 2003: 210–11). In this context, Reagan’s health policy adviser, Joseph Stetler – former president of the Pharmaceutical Manufacturer’s Association – invited the pharmaceutical industry to make proposals for FDA reform (Anon. 1981a). That same year Congress convened a Commission on the Federal Drug Approval Process to address ‘regulatory overkill at the FDA’ (Anon. 1981b). In 1982, the Commission recommended reforms designed to promote more rapid approval of new drugs, which were subsequently endorsed by the National Academy of Sciences on the grounds that such reforms would boost market innovation (NAS 1983: 86).
FDA management responded between 1982 and 1987 with proposals to streamline the new drug application process, incorporating many of the Commission’s recommendations. These included narrowing the scope of the FDA’s regulation of Phase I studies; replacing individual case report forms with company summaries of data; accepting foreign data collected in countries with lower regulatory standards as a basis for approval; and allowing closer and more frequent contact between industry and FDA officials (Anon. 1982a, FR 1985, 1987). Closer communication between the FDA and drug sponsors was justified on the basis that it would increase the likelihood of trial designs acceptable to the agency, thereby facilitating rapid new drug review (Anon. 1985).
In addition to formal initiatives to reduce regulatory requirements, FDA management pursued informal measures to shorten agency review times. Faced with successive budget cuts throughout the early to mid-1980s, FDA Commissioner Young responded by transferring staff and resources from other FDA departments into the new drug review divisions (Anon. 1986a). Evidence from the period suggests agency reviewers were under intense pressure from managers to decrease review times, with one FDA scientist from the oncological drugs division expressing concern before a Senate Committee that this high-level pressure to expedite approvals might ‘compromise the scientific integrity of [new drug] reviews’ (Anon. 1981c, 1982b). Meanwhile, the Secretary of State for Health praised the ‘new spirit’ of FDA-industry co-operation (Anon. 1986b).
Evidently early FDA reforms to accelerate, and increase the likelihood of, new drug approvals were initiated without reference to the AIDS crisis and several years before AIDS activists targeted the FDA. Although sometimes presented by government and industry as in patients’ interests, these changes did not result from patients’ demands or evolution of ‘the American patient’ as Daemmrich (2004) asserts. Rather, a partnership between the regulatory state (including the FDA, the Executive, and the Legislature) and the pharmaceutical industry was established by the mid-1980s dedicated to accelerating drug approvals.
Early in the AIDS epidemic no approved treatments were available for patients. However, this was hardly due to ‘regulatory overkill’. Rather, what is notable about the FDA’s response to AIDS is the speed and flexibility with which the agency acted once the first promising investigational drug (AZT) appeared (Hilts 2003). In March 1987, the new drug application for AZT was approved in just three-and-a-half months based on just one phase II double-blind, RCT (Anon. 1988a, 1989a). However, AZT’s accelerated development/review was achieved through enormous investment of agency resources, prompting pharmaceutical firms to complain that the ‘super-priorities’ accorded to AIDS drugs would divert over-stretched agency resources away from reviews of other new drugs (Anon. 1987).
While FDA actions to facilitate approval of the first AIDS drugs set the template for later reforms, the initial push to extend this to other drugs, such as cancer drugs, came from the Task Force for Regulatory Reform and industry, rather than patient advocates (Anon. 1988b).2 After consulting with the Pharmaceutical Manufacturers’ Association and patient/consumer groups, the FDA implemented the ‘Subpart E’ regulations, which extended the template for AZT’s approval even beyond drugs for life-threatening conditions to include products to treat ‘severely debilitating diseases’ (Anon. 1988c). Subpart E included key industry recommendations, but did not include proposals to which the industry had objected, such as conditional approval status for expedited drugs and mandatory post-marketing (Phase IV) commitments (Anon. 1988d). Thus, instead of a response to patient activism, this was more plausibly a continuation of the previous corporate bias in the 1980s, shaping the FDA as a collaborator with industry to give companies more confidence that drug development investments would pay off with marketing approval.
AIDS activists, industry and some scientists wanted the FDA to approve new AIDS drugs based on surrogate measures that could be assessed earlier in the drug development process (Epstein 1997). Cancer patient activism about such matters was, by contrast, absent, but efforts to reform the FDA by the Republican Administration and industry continued. In 1989, President Bush established the Committee to Review Approval of New Drugs for Cancer and AIDS with Dr Lasagna (a long-standing deregulatory critic of the FDA) as Committee Chair. Lasagna immediately announced that the Committee would examine the role of surrogate endpoints in expediting drug approval (Anon. 1989b).
Clinical trials using surrogate endpoints are in manufacturers’ interests because they require fewer patients, have shorter duration, and are cheaper, than trials tracking direct clinical effects. However, in the late 1980s, no surrogate markers were proven to predict increased survival in AIDS patients, while in 1985 the FDA’s Oncology Drugs Advisory Committee concluded that tumour response (shrinkage) should not be the basis for approval of cancer drugs because it did not correlate with increased survival or clinical benefit in numerous cancers (Johnson and Temple 1985, Johnson et al. 2003: 1404). Thus, in 1989, the FDA advised the Lasagna Committee against approving cancer drugs based solely on the surrogate endpoint, namely, tumour shrinkage; but the National Cancer Institute (NCI) argued in favour, while the National Institute of Allergy and Infectious Diseases proposed CD4(T)-cell counts as surrogate endpoints for approval of AIDS drugs (Anon. 1989a, 1989c). In September 1990, the Lasagna Committee recommended that the FDA should approve AIDS and cancer drugs based on surrogate endpoints of CD4(T)-cell counts and tumour shrinkage, respectively (Anon. 1990a).
AIDS activists lobbied the FDA to accept the Lasagna Committee’s recommendations so that HIV/AIDS patients could access drugs, still undergoing Phase II trials, that showed dramatically promising effects on CD4(T)-cell counts in Phase I (Anon. 1990b). In March 1991, Vice-president Quayle’s Council on Competitiveness, which was committed to removal of ‘regulatory burdens’ on industry, endorsed the Lasagna Committee’s recommendations and subsequently the FDA began investigating how ‘conditional’ new drug approvals based on surrogate endpoints could be granted (Anon. 1991a).
The final 1992 Accelerated Approval (Subpart H) regulations were negotiated between the FDA, the Secretary of State for Health, and Quayle’s Council, but it was Quayle’s industry-responsive Council whose proposals dominated. The FDA wanted marketing approval based on surrogate endpoints after completion of Phase II trials to be restricted to drugs for life-threatening diseases where no alternative therapy existed, and on condition that such drugs would have restricted marketing until Phase IV studies confirmed a positive benefit-risk ratio for direct clinical endpoints (Anon. 1991a). However, following industry intervention, the restricted marketing proposal was uncoupled from the provision to allow approval based on surrogates and accelerated approval was extended to drugs to treat not only life-threatening or severely debilitating, but also ‘serious’, illnesses – the formal definition of which includes reversible morbidity (Anon. 1991b, FR 1992, FDA 1998a: 4, Willman 2000).
Initially, the FDA’s oncology division was reluctant to approve cancer drugs under subpart H due to disagreement among oncologists about which surrogate endpoints were appropriate (Clinton and Gore 1996: 3). Once again, there is little evidence that pressure for the FDA to accept tumour shrinkage as a surrogate endpoint in cancer trials came from patient advocates (Anglin 1997: 1407–9). Rather, the Clinton Administration’s ‘Reinventing Government’ initiative was decisive, within which President Clinton and Vice-president Gore directed the FDA to accept tumour shrinkage as a valid surrogate marker for accelerated approval (Clinton and Gore 1996). The principal explicit goal of the Reinventing Government initiative was to cut regulatory ‘red-tape’ for the benefit of business.
AIDS activism significantly helped to establish CD4(T)-cell counts as an acceptable surrogate endpoint, as disease-politics theory suggests, though some of that impetus came from the industry-friendly Lasagna Committee (Epstein 1996: 265–308, FDA 2005). However, there is little evidence that patient activism drove the extension of those regulations to surrogate endpoints beyond AIDS. Extension was achieved through joint efforts of industry and the state, particularly the Executive branch. Thus, corporate bias theory seems to account well for the development of these accelerated approval regulations – certainly beyond AIDS drugs, and even to some extent including AIDS drugs.
In 1994 the Republicans gained control of both House and Senate in Congress and told industry it had a unique opportunity to lobby Congress for changes at the FDA (Anon. 1994). By mid-1995, the pharmaceutical industry drafted legislative proposals including a new approval standard based on data from just one clinical trial, especially for ostensibly ‘breakthrough drugs’ and a greater emphasis on use of surrogate endpoints (Anon. 1995a, 1995b). When FDA reform bills sponsored by members of Congress, which underpinned the 1997 FDA Modernization Act, came before the House and Senate, it was reported that sections repeated verbatim industry’s proposals (Anon. 1995c). Indeed, an informant closely involved with this legislative process claimed that a draft of the Act contained blank sections with the words: ‘Industry inserts language here’ (Public Citizen representative 2003).
By contrast, many patient/consumer groups, including AIDS and cancer activists, were opposed to any further lowering of the regulatory standard – particularly approval based on one clinical trial unless in very exceptional circumstances (Patients’ Coalition 1996). Consequently, they formed the Patients’ Coalition, which by 1997 comprised over 100 organisations representing patients with serious, rare or life-threatening illnesses (Patients’ Coalition 1997). According to Coalition representatives, no ‘genuinely independent’ (non-industry-funded) patient advocates supported the Act.3 Additionally, the Patients’ Coalition lobbied Congress for the Act to empower the FDA to impose fines on pharmaceutical firms that failed to conduct confirmatory studies. Significantly, the final legislation rejected that proposal, which was opposed by industry, and accepted many of industry’s demands, including those opposed by patient activists, such as the single-trial rule (Anon. 1997).
Such lobbying of Congress about the Modernization Act shows the evolution of organised patient activism noticed by disease-politics theorists. However, this was not an evolution of the American patient away from state protection to free-market consumerism with fewer regulatory demands on manufacturers, as disease-politics theorists imply. Rather, this patient activism opposed such deregulation, suggesting that disease-politics theory may have exaggerated the homogeneity and ‘consumerism’ with which patient advocacy insists on accelerated access to new drugs. Furthermore, the fact that such opposition was unsuccessful in reshaping accelerated approval regulations when contrary to industry interests indicates, once again, that the emergence of accelerated approval regulations cannot be adequately explained as a response to the demands of patient activism per se. Instead, the experience of the Modernization Act implies that the demands of industry were paramount.
Implementing accelerated approval: a case study of Iressa
Lung cancer is the most common cause of cancer death in the US. Two-thirds of patients who present with non-small-cell lung cancer are beyond curative surgery so effective drug therapy is vital. Companies seeking standard approval of drugs to treat this cancer must demonstrate that a drug provides a meaningful therapeutic benefit for patients such as increased survival or symptom improvement. However, the FDA also stipulates that tumour response can qualify as a surrogate measure for accelerated approval under subpart H ‘in refractory cancer settings, where therapies with meaningful benefit are unavailable’, and that evidence to confirm true clinical benefit can be obtained through post-approval Phase IV studies (FDA 1998b: 3).
On this latter stipulation, in August 2002, AstraZeneca sought accelerated approval for Iressa based on one small uncontrolled, un-blinded, Phase II trial in 139 patients for third-line treatment of non-small-cell lung cancer – that is, in patients refractory (unresponsive or intolerant) to both platinum-based (first-line) and docetaxel (second-line) therapy (Cohen 2002: 10). Of the two primary measures of efficacy in this trial, the FDA found that one (symptom improvement) could not be accurately evaluated and the other (tumour response/shrinkage) was reached in only 10 per cent of patients (FDA 2002: 117). This fell below the Lasagna Committee’s standard that accelerated approval of cancer drugs should be based on tumour regression in at least 20–30 per cent of patients (Anon. 1989c, 1990a).
FDA scientists reviewing the Iressa data questioned whether a 10 per cent response rate was ‘likely to predict clinical benefit’ (FDA 2002: 110), and even if it did, they were concerned it overestimated the actual number of patients who might respond to the drug once marketed because trial patients were atypical – 70 per cent had ‘less aggressive, slow-growing tumours’ than the US population with this cancer (Cohen 2002: 13–15). Nonetheless, confident of accelerated approval, AstraZeneca began confirmatory trials, known as INTACT, to demonstrate the efficacy of Iressa on survival, even before the FDA had granted marketing approval. Consequently, the INTACT trials, initially intended as Phase IV confirmatory studies, became pre-market Phase III studies by default. In fact, further doubt, rather than confirmation, was cast over Iressa’s efficacy when the INTACT results became available to the FDA in early September 2002.
INTACT comprised two well-conducted, double-blind, placebo-controlled, randomised trials evaluating the effect of standard chemotherapy combined with Iressa on overall survival in more than 2000 first-line patients with non-small-cell lung cancer (FDA 2002: 129).4 If Iressa had significantly improved survival in these trials, then its application status could have been upgraded from accelerated approval to regular approval because survival is a key clinical endpoint (Williams 2002: 4). However, INTACT failed to demonstrate that Iressa prolonged survival (FDA 2002: 132). Indeed, patients survived a month longer on placebo (Williams 2002: 4). Moreover, there was ‘no difference between Iressa and placebo with respect to tumour shrinkage’, even though this ought to have been easier to demonstrate in first-line patients because their cancer was less unresponsive to therapy than third-line patients (FDA 2002: 132). One FDA scientist characterised the scientific evidence from these Phase II and III trials as: ‘It’s 2000 patients saying Iressa doesn’t work versus 139 saying it works marginally’ (FDA Medical Officer 2003).
Meanwhile, AstraZeneca keenly ran an ‘expanded access programme’ involving over 12,000 patients with such cancer. Such programmes are un-blinded, non-controlled, and not designed to evaluate efficacy (FDA 2002: 19, Williams 2002: 5). The FDA permits patients, who have exhausted all treatment options and are ineligible for clinical trials because of poor fitness/illnesses, to receive new drugs that show therapeutic promise in Phase II or III trials (Stahel et al. 2003). While such programmes can have financial costs for companies, industry executives acknowledge that firms ‘benefit from a lot of goodwill from patients’ by offering them (Baldwin 2002: 1669).
The FDA deferred making a decision on Iressa and sought the advice of its expert Oncology Drugs Advisory Committee. On 24 September 2002, in addition to analyses of the clinical trials by AstraZeneca and the FDA, the advisory committee heard testimonies from advocacy groups and individual patients on the access programme. Some patients were funded to attend the meeting with AstraZeneca money distributed through a patient group (Goldberg 2005: 15). The National Organisation of Rare Diseases, which ran the access programme with AstraZeneca, spoke supportively about Iressa, as did the Cancer Research Foundation of America and The Wellness Community, both of which received ‘unrestricted educational grants’ from the company (FDA 2002: 22–52). Six patients recounted, anecdotally, how Iressa‘began to eliminate cancer symptoms in seven days’, made tumours ‘90 per cent gone in three months’, was ‘a wonderful drug’, ‘light-years better than previous treatment’, and ‘will save lives’ (FDA 2002: 22–52). Despite the negative INTACT trial results, the advisory committee voted 11–3 in favour of accelerated approval for Iressa.
Our interviews with patient groups and FDA scientists attending the meeting indicate that some advisory committee members were swayed by the testimonies of patients on the access programme. One FDA scientist said: ‘I think they [patient testimonials] definitely have an influence over advisory committees. That’s what Iressa proves’ (FDA 2003). Another informant recounted how two committee members acknowledged their votes in favour of accelerated approval were influenced by patient testimonies despite recognising that such testimonies do not constitute sound, scientific evidence of drug efficacy (Director, Center for Medical Consumers 2005). This is significant because FDA scientists believed that the advisory committee’s recommendation was a key factor in the agency’s decision to grant accelerated approval to Iressa.
Hence, there is certainly evidence of patients influencing FDA decision-making to accelerate approval of the cancer drug, Iressa, but whether this illustrates a new disease-politics in which the regulatory agency is primarily responsive to patients’ interests, with the interests of pharmaceutical firms merely a secondary consideration, is highly questionable. For instance, AstraZeneca largely determined which patient voices the committee heard. Thus, ‘pressure’ felt by the committee from patients to approve Iressa should not be regarded as entirely separate from ‘pressure’ exercised by the manufacturer. Industry shaping of patient testimonials extends to other cancer drugs as well as Iressa (Goldberg 2005: 12). Moreover, the manufacturer also influenced oncologists directly, such as those on the committee. Reminiscent of Brown and Webster (2004) on sociology of technological expectations, the same FDA scientist commented:
I’m not saying it was a lot of pressure that caused Iressa to get approved. But there was a lot of pressure. AstraZeneca did a fantastic job of marketing Iressa to the oncology world, and everybody was thinking it was a wonder drug! When it clearly isn’t (FDA 2003, informant’s emphasis).
Indeed, FDA scientists emphasised manufacturers’ determination of knowledge about, and availability of, cancer drugs – something entirely neglected in disease-politics theory and dominant media narratives (FDA Drug Evaluation Director 2003, FDA 2003). The FDA’s Oncology Division asks, but cannot force, firms to conduct trials with survival, rather than surrogate, endpoints. FDA scientists pressed AstraZeneca to conduct trials comparing Iressa with second-line therapy, but the company refused for two years due to fear of negative results (FDA 2003). This suggests that the corporate bias of industry priorities built into accelerated approval regulations, frames the FDA’s regulatory options significantly, long before patient testimonials at advisory committees.
The FDA had a clear option to reject its advisory committee’s recommendations of September 2002 to approve Iressa on to the market. In October 2002, the agency became aware of reports of fatal acute-interstitial pneumonia associated with Iressa during the drug’s marketing in Japan (Anon. 2002). About two per cent of Iressa patients were affected, a third of whom died within one or two weeks (FDA 2003: 107–120). Insofar as a small proportion of third-line patients might benefit from Iressa, no such patients could be identified prospectively either on the product label or by prescribing physicians. Consequently, a majority of users would be exposed to potentially fatal reactions without deriving any therapeutic benefit from the drug. Nevertheless, on 5 May 2003, the FDA approved Iressa as third-line treatment for non-small-cell lung cancer.
As a condition of that accelerated approval, AstraZeneca agreed to conduct a post-marketing confirmatory RCT (known as ISEL) in 1700 third-line patients with survival as the clinical endpoint. In December 2004, the company reported that ISEL showed no survival benefit with Iressa compared to placebo (FDA 2005: 11–12). Some regulatory experts believed that patients’ interests would have been better served by Phase IV trials identifying/studying the small group of third-line patients with Iressa-responsive genetic profiles, but noted that ‘that would have been an economic disaster because it would reduce [drug] utilisation to very few patients’ (EU regulator 2005).4 Such frustration exploded publicly at an FDA advisory committee meeting in March 2005, discussing the ISEL results, when one expert declared:
Development of this drug has been mishandled by AstraZeneca and this committee. I take some blame for that because I voted for approval two years ago … this drug has been available for seven years, and we still haven’t figured out how it should be used in lung cancer. Perhaps if we had held off in getting it available to people two/three years ago, those studies would have been done (FDA 2005: 124).
Despite the failure of ISEL to confirm efficacy, the FDA did not withdraw Iressa from the market as would be expected under accelerated approval regulations, even though by early 2005 an alternative lung cancer drug, Tarceva, became available on the US market after demonstrating survival benefit in a RCT in third-line patients (Anon. 2009). While, AstraZeneca withdrew its EU marketing application for Iressa in January 2005 after the EMEA indicated it would not be approved in Europe, the FDA’s initial response to ISEL was to allow continued marketing of Iressa in the US provided AstraZeneca sent doctors letters advising against starting new patients on the drug (FDA 2005: 12–13). However, according to the health advocacy group, Public Citizen, who petitioned the FDA not to approve and then to remove the drug from the market, there were 331 new prescriptions for Iressa in one week of February 2005 alone (FDA 2005: 94). In June 2005, the FDA altered Iressa’s labelling to limit its distribution to patients in clinical trials or those who were benefiting/had benefited, from the drug.
Maintaining Iressa on the market after failing to confirm clinical efficacy when an alternative drug with proven efficacy to treat the same condition was available was in AstraZeneca’s interests, but inconsistent with patients’ health interests and the FDA’s own techno-scientific regulatory standards. Indeed, even marketing approval of Iressa in the first place was inconsistent with those standards and barely in the health interests of patients given that tumour response was unconvincing, trial patients survived longer on placebo, and there was a two per cent chance of being killed by the drug almost instantly.
This shows that the dichotomy within disease-politics theory of an over-cautious FDA acting against patients’ interests by denying them timely market access to drugs before AIDS, on the one hand, and an FDA responsive to patients’ interests by accelerating access to drugs after AIDS, on the other, is inadequate because Iressa highlights FDA’s accelerated approval against the health interests of patients. The question then raised is why the FDA did not reject the testimonials from patients on company-funded access programmes, uphold its own techno-regulatory standards, and refuse approval. Disease-politics theorists, such as Carpenter (2004), might argue that this occurred because the FDA, motivated by reputational considerations, did not wish to be accused of ignoring the demands of some patients and the advice of its expert committee. While possible, this explanation is tenuous because a compassionate-use programme could have made the drug available to patients already benefiting from it without permitting the drug on to the market. Furthermore, that argument does not explain why the FDA exhibited no such reputational concerns about transgressing its own standards in the face of public calls by health advocacy groups not to approve or maintain Iressa on the market.
Our interviews with FDA scientists point to a different explanation for this divergence from health interests, namely, a political culture – especially among FDA management – which prioritises industry interests. For example, that institutional context has involved internal organisational pressures from FDA management who had assimilated the ideology that pharmaceutical companies were FDA’s clients, with whom agency reviewing divisions should work to accommodate new drugs on to the market:
This philosophy has come down [from FDA management] that we work with industry to get drugs approved. Within the agency, there are people who view drug companies as their clients … it’s ‘how do we give them [drug companies] what they want?’… the balance has tipped so that the benefits of the doubt will be given to the company and there’s greater reliance that we can take care of things after drug approval (FDA Drug Safety Scientist 2003).
There were cases where reviewers would have reservations about a drug that were ignored. If the reviewers tried to do something about it then someone from the higher management would get involved … So a lot of people took the path, ‘it’s not worth the effort’ (former FDA Biostatistician 2003).
Discussion and conclusion
Our longitudinal analysis of the emergence of accelerated approval regulations demonstrates that, contrary to disease-politics theory and consistent with corporate bias theory, regulatory reform following the high-point of AIDS treatment activism continued, rather than departed markedly from, earlier attempts by the FDA to reduce regulatory requirements mainly in response to, and driven by, industry concerns. The emergence of accelerated approval regulations were part of a broader deregulatory regime in the US, pre-dating AIDS (Francis 1998). Nonetheless, we found a growth in American patient activism beyond AIDS from the mid-1990s hitherto neglected by corporate bias theory. While disease-politics theory has identified this growth, it has mischaracterised it solely in terms of ‘consumerism’. In fact, one of the largest mobilisations of American patients’ groups during the 1990s was in opposition to deregulatory measures in the Modernization Act designed to enable pharmaceutical firms to develop and market their products faster and cheaper. Significantly, however, we found that such patient activism lacked influence when opposing pro-industry reform proposals and/or proposing anti-industry reforms. This is further evidence that, with the exception of some aspects of AIDS drugs regulation, industry interests, rather than the independent power of patient activism, are the primary sociological explanatory factor for the emergence of accelerated drug regulation.
The theory of corporate bias is largely applicable to the emergence of accelerated approval regulations because industry and the state worked in partnership to reshape the FDA’s policies to more fully reflect industry’s commercial priorities spurred on by the wider ‘neo-liberal’ political environment (Navarro 1994). However, it is notable how the pharmaceutical industry stretched its influence across all major aspects of the state, including the ‘bureaucracy’ (FDA), the ‘Executive’ (Reagan, the two Bush, and Clinton Administrations), and the ‘Legislature’ (Congress) – reminiscent of ‘horizontal policy networks’ (Salter and Jones 2006). Rather than a vertically centralised form of corporate bias conventionally associated with European pharmaceutical regulation, what seems to have emerged in the US is ‘tentacled corporate bias’ in which the industry seeks a range of strategic partnerships with various elements of the state and, as our case study shows, with some patients’ groups and medical professionals.
We do not claim that our case study of Iressa entirely confirms or falsifies existing theories, but it points to their limitations and potential improvement. It demonstrates that accelerated approval of a cancer drug is not necessarily synonymous with patients’ interests, and that a more cautious FDA would not necessarily be contrary to patients’ interests, even when some patients support such approval. It also explains that such a scenario occurred primarily because FDA decision-makers were inclined to consistently give the company the benefit of the doubt, albeit an inclination strengthened by support for the drug from some patients and expert advisers.
While disease-politics theory, which characterises the FDA as primarily responsive to patients, cannot account for this scenario, our case study also suggests that corporate bias theory must take account of a more complex and indeterminate representation of industry interests. This includes a ‘culture-ideology of consumerism’ in the form of indirect influences on the regulatory state via patient-group and/or professional alliances with pharmaceutical firms, as well as direct government-industry relations (Sklair 2001). For example, a modified theory of ‘tentacled corporate bias’, in which the regulatory state tends to embrace demands for accelerated drug approvals from patients and the medical profession because those demands help to cement a smooth partnership with industry could account for the Iressa case, especially in an institutional context where a political culture timid about antagonising manufacturers has percolated down to regulatory scientists.
Acknowledging that one should not generalise from a single case study, it is important to note that the Iressa case does not stand alone. It shares many regulatory commonalities with other cancer drugs granted accelerated approval. In interviews, some FDA scientists expressed scepticism that FDA management would ever remove a drug approved under accelerated approval regulations (FDA 2003, FDA Oncology Leader 2005). As of September 2009, the FDA had never withdrawn from the market a drug granted accelerated approval despite failure to confirm clinical benefit because Phase IV trials were either negative or never conducted (US GAO 2009: 32–3). Such policy has implications not only for current cancer patients, but also wider public health because of the longer-term effects of lowered efficacy standards on incentives of firms to develop future cancer drugs that might truly provide therapeutic benefit.
In Carpenter’s case, the argument is that accelerated approvals are in the interests of the FDA’s reputation to protect/promote public health.
Daemmrich and Krucken (2000: 506) claim cancer patient activists ‘pressured the FDA to approve Taxol for advanced breast cancer’, but acknowledge that this was not until the early 1990s, years after subpart E’s extension to cancer drugs. In Daemmrich’s (2004: 86–99) account of cancer drug (Proleukin), from 1988–1992, no patient activism ‘pressured’ the FDA for it, and in 1993 the FDA established Cancer Liaison Programmes to increase communication with cancer patients.
Often companies switch from third-line to first-line trials when designing confirmatory studies because, if successful, approval grants access to the larger market of first-line (plus third-line) patients.
In 2009, the EMEA approved Iressa for a small number of non-small-cell lung cancer patients with Iressa-sensitive genetic mutations, who showed survival benefit in RCTs in 2008 (Anon. 2009).
We are grateful to the ESRC (grant no. L218252001) for funding the research upon which this article is based, and to two anonymous referees and the editors for comments on a previous version of this article.