In the 1930s, a movement had grown within the United Kingdom in favour of the establishment of a nationalized health service, with treatment free at the point of use.Footnote 1 A young medical student named Archie Cochrane heard about a local rally in support of this goal in a suburb of London. Most of his political activity to that date had been concerned with anti-fascism, including fighting in the Spanish Civil War as a member of the International Brigade.Footnote 2 But the idea of a national health service appealed to him as a trainee doctor, and as a socialist, and so he decided to go along with a home-made banner reading,
ALL EFFECTIVE TREATMENT MUST BE FREE
A statement about which he ‘had a deep inner feeling’ that it was ‘absolutely right’ because ‘the difference in the medical care of the rich and the poor was sufficient to touch the hardest-hearted student in the 1930s’.Footnote 3 The implications of this demand were not, however, uncomplicated. How would a nationalized health service determine whether a treatment was sufficiently effective to warrant taxpayer funding? Who would decide, and by what metrics? In short, the practical puzzle of how to measure effectiveness would become Cochrane’s central preoccupation for the rest of his career, with far-reaching effects on how the health service would come to be managed by the turn of the twenty-first century.
Cochrane served in the Second World War, including a formative stint practising as the chief medical officer in a prisoner of war camp in Salonica, where he found that the information distributed to doctors behind the front lines about treatment methods was lacking in justification and evidence for support.Footnote 4 On his return in 1945, Britain elected a Labour government, which eventually made good on its promise to establish the National Health Service (NHS), a process led by the fiery Welsh minister of health, Aneurin Bevan, in 1948 – a challenging ambition in an age of post-war austerity, but part of a post-war consensus, and a pillar of the welfare state.Footnote 5 Perhaps by no coincidence, 1948 was the same year when the first randomized control trial (RCT) was published, ‘Streptomycin treatment of pulmonary tuberculosis’, led by statistician Austin Bradford Hill (later to become a mentor to Archie Cochrane), and funded by the Medical Research Council (MRC) to determine the effectiveness of this novel drug.Footnote 6 The MRC had itself been founded in 1913 as a consequence of the National Insurance Act of 1911, and was charged with distributing the funds for research into which treatments should be made available through insurance, a decision made by the MRC as a government body, rather than by the insurance companies themselves.Footnote 7
Decades before the birth of the NHS, the question of validating which treatments should be eligible for funding was a prime concern of government, and this only became more urgent once the health service was nationalized, at a time when the country was still subject to food rationing. The discipline of clinical epidemiology, actively shaped by Bradford Hill and Cochrane, emerged in this context, with increasing opportunities for research funding in subsequent decades, the establishment of infrastructure in terms of MRC-led units, and university departments which trained new cohorts of researchers and embedded some of these principles in medical training and health economics. Later, in the 1960s (also with Cochrane’s involvement), health services research became a field of its own, incorporating epidemiology and other interdisciplinary approaches.Footnote 8
From the 1990s, evidence-based medicine (EBM) came to be a phrase – and indeed a philosophy and a practice – associated with both of these disciplines. Some would even start branding it a ‘movement’, one that had transatlantic appeal, especially in Canada, where the term was first coined by Gord Guyatt and David Sackitt at McMaster University.Footnote 9 The RCT, championed by the clinical epidemiologists of the early years of the NHS, came to be the gold standard in the British iteration and development of evidence-based medicine. Not only that, it also became enshrined in health care policy with the foundation of NICE, the National Institute for Clinical Excellence, in 1999.Footnote 10 This was led again by a Labour government under the auspices of the Secretary of State for Health, Frank Dobson, a politician whose obituary described him has having ‘devoted his lifetime … to giving practical effect to the socialist principles he had always held and from which he never deviated’.Footnote 11 Since that time, NICE guidelines have been established for many disorders: they draw from evidence from RCTs and cost–benefit analysis of treatments, and ultimately determine which treatments the NHS will pay for – and also those that it will not – from pharmaceuticals to psychotherapy. This is, of course, not without controversy, particularly among patient groups and pharmaceutical executives, although NICE has thus far been preserved by all subsequent UK governments, no matter their ideological orientation.
Britain is one of the few socialized and nationalized health systems in the developed world not to be modelled on an insurance system.Footnote 12 Although it is by no means the only nation with a medical community to embrace RCTs and evidence-based medicine, it is one of a minority for whom the results of RCTs have become concretized in government-led policy on a nationwide scale through guidelines, which in turn have gone beyond the role of ‘guidance’, and have actively restricted or enabled specific treatments through the decisions of health service procurement bodies.Footnote 13 Despite this, the British story has only had scant attention in the key sociological works on EBM (notably Daly, Berg and Timmermans, and Hanemaayer).Footnote 14 Thus this article draws on archival material, published oral-history witness seminars, and printed material (scientific articles, reports, policy literature and memoir) to draw together a long-term history of the relationship between the NHS and evidence-based medicine. Numerous actors were involved with these developments, but the article will have a particular focus on Archie Cochrane, his 1971 work Effectiveness and Efficiency, and its reception and impact. The principles laid out in this book, I argue, ultimately undergirded the establishment of the NICE guidelines (under the National Institute for Clinical Excellence) at the end of the 1990s.
The article also explores the philosophical underpinnings of the validation practices of clinical epidemiology, making a broader contribution to this issue on regulation and validation in the sciences of health. Crucially, it places these developments in their context, arguing that the history of EBM needs to be understood not only as a search for a more ‘scientific’ means of validating treatments in medicine, but also as one which has been inextricably bound up with questions of health economics and funding. I acknowledge the contribution of Jeanne Daly to the history of these debates, drawn from oral histories with key figures, comparing the histories of the EBM movement across the USA, Canada and the UK.Footnote 15 I go beyond the chronology of their work to evaluate how NICE also became a realization of Cochrane’s earlier EBM principles in the late 1990s and early 2000s. Unlike the substantial philosophical interventions on EBM, I am concerned primarily with the social, political and economic context of its emergence, and with the implications of its implementation in these same terms.Footnote 16 While no author has yet sought to chart the long-range history of the rise of EBM in Britain, a number of scholars have addressed key moments or periods in this story. As such, I will engage with their work in each section as appropriate.
I argue that the social-democratic politics of a number of key figures in the history of British EBM (epidemiologists Archie Cochrane, Richard Doll and Iain Chalmers, and the later Labour health minister Frank Dobson, who supported the establishment of NICE) needs to be recognized, as EBM is often cast as a product of neoliberal forces. The motivating factors for a number of the key historical actors in the history of British EBM were a commitment to equality of access to health care, a presumption that a taxpayer-funded health service free at the point of access to all citizens was a necessity, and, among policymakers, a need to distribute resources on the basis of evidence and cost–benefit analysis in a way that limited the disproportionate profit-driven influence of the pharmaceutical industry. As such, regulation and validation fit closely with social-democratic concerns, and these motivations shaped both the construction of their mechanisms and their implementation. The contemporary politics of health care provision is beyond the scope of this article, but I suggest that current debates about EBM from a left-wing perspective would benefit from a better understanding of some of the historical left-wing logics that played a key role in their construction.
This creates a tension between the British case and arguments made by scholars such as sociologist Ariane Hanemaayer about the ‘neoliberal’ quality of evidence-based medicine in the Canadian context.Footnote 17 EBM, as it has played out in the British NHS, has actively limited both the doctor’s and the patient’s ability to engage in consumer ‘choice’ within the health care system – and deliberately so. I argue that the mode of governance that ensued is better thought of as a mode of technocracy. This resonates with a late twentieth-century shift noted by sociologists and STS scholars such as Frank Fischer, Jürgen Habermas and Sheila Jasanoff in terms of the rise of ‘expert’ knowledge in policy decision making in Western democracies, with an alleged detrimental effect of democratic exclusion.Footnote 18 However, this situation is a more nuanced technocratic formation: one that paradoxically diminishes the power of certain types of expertise. Instead, the combined technology of the RCT and the cost–benefit analysis calculation were intended to strip experts (be they individual physicians or pharmaceutical company executives) of their subjective authority, generating a form of evidence that was presented as transcendent in its claim to objectivity. The social realities of this assemblage of validation, regulation, policy and practice have, of course, been complicated, with unintended effects, as social scientists have documented.Footnote 19 More ethnographic and sociological work would be valuable to understand how guidelines have been developed, and how subsequent policies play out. This paper will engage with this literature as far as it exists, but at its centre it argues for the importance of understanding the intentions and motivations for EBM and its application within the health service.
The emergence of randomized control trials
Historians of science are trained to eschew hagiography, but sometimes the stories of the ‘saints’ of a particular intellectual tradition do provide us with coordinates for understanding how scientists themselves have created mythologies that have meaning for professionalization and identity, shaping communities of consensus around specific ideas and practices. The EBM mythology is one of the decades-long frustration and hard toil of a few individuals: of long-delayed enlightenment and eventual mass conversion to a cause, which took almost half a century to achieve in the face of conservative (with a small ‘c’) clinical resistance.Footnote 20 While Cochrane is often cited as the ‘father’ or ‘pioneer’, Cochrane himself commemorates 1948 as the key origin point: the year in which statistician Austin Bradford Hill, Geoffrey Marshall and D’Arcy Hart published the outcome of the first randomized control trial to examine the effectiveness of streptomycin in the treatment of tuberculosis.Footnote 21 Epidemiologist Richard Doll, a key later collaborator of Bradford Hill and disseminator of evidence-based values in clinical epidemiology, described 1948 as ‘the watershed moment’ as a result of this publication. He also pointed out in retrospect that although the TB trial was the first to be published, it was actually preceded by the MRC’s field trial to compare the effectiveness of four different whooping cough vaccines, which began in 1946 and was completed in 1948, although the full publication did not come to light until 1951.Footnote 22 For our purposes, in terms of understanding the impact of the trials on epidemiology and medicine as practices, the publication date is the more significant factor.
The standard treatment choice at the time was bed rest. All of the patients who participated were given the standard treatment – but they were randomly assigned to either the S group (to receive streptomycin in addition to bed rest) or the C group (control, meaning that they received no streptomycin, only bed rest). The assignation was done by means of envelopes handed to patients’ consultants from the central trial organizers (inside it would be indicated S or C, with the order of the envelopes generated by a random number sequence). The trial was blinded: the patients were not aware of which group they were in, and the doctors assessing them by chest X-ray or microbiological sampling also did not know. Ultimately the trial showed that streptomycin had some effectiveness: only four out of fifty-five patients in the S group died, by comparison with fifteen out of fifty-two in the control group. Unfortunately, it became clear that bacterial drug resistance was also a factor, and patients deteriorated after the bacilli became resistant to streptomycin after only a few months.Footnote 23
The idea of a trial to evaluate the effectiveness of a treatment was not new, and the Medical Research Council had been in existence since 1919, having evolved from the Royal Commission on Tuberculosis under David Lloyd George. Martin Edwards, in his thorough research on the MRC’s work prior to 1948, notes that the term ‘controlled’ had a variety of different meanings in the trials conducted prior to the 1948 TB and whooping cough trials: it was the systematization of these two designs that created a model for the ‘gold standard’.Footnote 24 That tuberculosis became a serious object of study for one of the first RCTs is apt, given the long-standing profile that the disease held for British public health. Streptomycin was first isolated in 1943 by Alfred Schatz at a laboratory at Rutgers. It showed efficacy in the trial as an antibiotic, but was far from an ideal treatment, with significant toxicity and acquired antibiotic resistance. George Orwell, who procured a course of treatment through his American contacts in the same year, described his experience in vivid detail: ulcerated throat, skin flaking away, blistered lips, loss of hair and nails. But it was the first drug to show significant promise in the treatment of TB.Footnote 25
A number of historians have drawn a causal link between Bradford Hill’s training as an economist and statistician, including his authorship of a key textbook on medical statistics, and the decision to randomize patients. They argue that the decision to randomize patients was an application of the work of R.A. Fisher’s statistical principles to trial methodology, following from Fisher’s 1930s work showing that randomization was crucial for testing statistical significance.Footnote 26 Certainly, Bradford Hill was well versed in Fisher’s work, which was incorporated into his own textbook, Principles of Medical Statistics, based on a series of articles published in The Lancet – indeed, he corresponded with Fisher directly in 1937 in relation to his preparation of the book.Footnote 27 Epidemiologist Iain Chalmers (who later worked to further Cochrane’s agenda), challenges this line of causality, pointing instead to Bradford Hill’s concerns about selection bias in previous MRC trials.Footnote 28 Randomization through the use of random number schedules was chosen specifically as a means to conceal the identity of the control groups from the researchers conducting the trial in order to minimize bias as far as possible. The choice was pragmatic, rather than an application of statistical theory, although the statistical principle was later meshed with the practice, increasing the justification for the RCT as a gold standard validatory method. Alan Yoshioka posited that the choice also made sense from a more immediately political perspective: patients had a strong sense of wanting to be treated fairly, and the fact that they were randomized without the research team knowing who was on which list meant that the method could be presented as one with fairness at its core.Footnote 29 If some individuals were to be denied streptomycin for the greater good of validating a treatment, this deprivation should at least follow an equitable ethical code. The imperatives for randomization were multiple.
Historian Theodore Porter notes that both the use of statistics for standardization and cost–benefit analysis had much longer histories in Anglo-American society prior to the 1940s, emerging in the eighteenth and nineteenth centuries to manage distrust in individual judgement.Footnote 30 Furthermore, they had a pragmatic value, enabling the functioning of disciplines and professions through communities who could use agreed, quantified data to cooperate in the administration of public life – something that was particularly relevant in medicine where physicians traditionally had operated as private individuals.Footnote 31 Nevertheless, it took until the late 1940s for the RCT as a technology to be taken up within medicine, and, as Porter asserts, while doctors and statisticians led the streptomycin trial, the initiative and funding for its inception came from the state.Footnote 32 We can note that the subsequent appropriation of the Bradford Hill model in Britain also happened in parallel with the first decades of the establishment of a health service, which required the mass cooperation not only of doctors, but also of other health care professionals – managers, procurement departments and policymakers – with communication across different levels of primary, secondary and tertiary service providers. The value of standardized, validated data was not only a question of clinical and economic regulation. It also assisted with the shift from services provided largely by individual clinicians and hospitals to a large-scale and highly complex, collectivized health care system faced with novel challenges from post-war industrialization, consumer culture and novel public-health concerns.
Later, with Richard Doll, Bradford Hill applied the principles of the RCT to another post-war concern, lung cancer, demonstrating a clear link with smoking, which had higher statistical significance than car exhaust fumes, road tarmacking or the by-products of coal fires.Footnote 33 Doll was active in the Socialist Medical Association (later the Socialist Health Association, a Labour Party-affiliated group that campaigned for the NHS) and would establish (with Richard Peto) the first MRC Clinical Trials Unit in 1975 at Oxford, with a focus on smoking-related diseases.Footnote 34 The delay here is notable. Despite the success and the retrospective importance accorded to the 1948 MRC trials as a model trial methodology, the movement to make trials a core part of health service research did not emerge fully until the 1970s. Referring back to Theodore Porter, this time lapse might be explained to some extent by the fact that the 1960s and 1970s saw the zenith of the ‘overwhelming success of quantification in the social, behavioral, and medical sciences during the postwar period’, both within the academy and in public life.Footnote 35 The reinforced value of the RCT by the 1970s was part of this overall trend, and the MRC Clinical Trials Unit’s foundation marked the institutionalization of RCTs in both the academic research community and the health service.
Archie Cochrane, evidence and the health service
Three years prior to the foundation of the MRC unit, Archie Cochrane’s Effectiveness and Efficiency: Random Reflections on the Health Service was published. In large part due to the reception of this publication, Cochrane’s name has become synonymous with EBM. Effectiveness and Efficiency was based on his Rock Carling Fellowship lecture at the then Nuffield Provincial Hospitals Trust (known since 1998 as the Nuffield Trust, an independent charitable policy think tank specializing in health services research).Footnote 36 The trust was one of a number of independently governed charitable organizations linked with health and education established with philanthropic funding from Lord Nuffield of the Morris motor car manufacturers. It held considerable influence with NHS leaders and researchers.Footnote 37 In the words of the trust’s then secretary, Gordon McLachlan, the spectre of NHS reforms was looming by 1970, confirmed by the incoming Conservative government. Greater efficiencies would be required in the NHS, with a reasonable government research and development budget being made available.Footnote 38 Britain was significantly impacted by the economic downturn of the OPEC crisis and, as Eleanor MacKillup and Sally Sheard point out, health care costs were rising due to technological innovations in areas like organ transplantation, renal dialysis and the macroeconomic effects of an ageing population.Footnote 39 The Nuffield Trust’s members knew, however, that politicians often paid little heed to understanding how ideas of effectiveness and efficiency were designed or measured.Footnote 40 It was in this atmosphere that the trust decided to award Cochrane the fellowship.
Cochrane had been committed to trials from an early point – as early as 1941 as a prisoner–doctor in the POW camp in Second World War Salonica in a situation of dire hunger and rampant typhoid, he attempted an early trial to understand whether nutritional yeast could help alleviate oedema (using vitamin C tablets as the control). While it wasn’t adequately randomized, this was the germ of an idea for future research.Footnote 41 His experience in the camp also prompted reflection on the nature of evidence in treatment. Propaganda pamphlets circulated, and one, according to Cochrane, was aimed specifically at POW medical officers and discussed the concept of ‘clinical freedom and democracy’:
I found it impossible to understand. I had considerable freedom and choice of therapy: my trouble was that I did not know which to use and when. I would gladly have sacrificed my freedom for a little knowledge … I knew that there was no real evidence that anything we had to offer had any effect on tuberculosis, and I was afraid that I shortened the lives of some of my friends by unnecessary intervention.Footnote 42
The fetishizaton of ‘clinical freedom’ had facilitated what he described as the doctor’s ‘God complex’, a sense of entitled authority based on a vague sense of ‘clinical experience’ accrued over the course of years, which was not, however, in fact based on clear data – something he would particularly allege about cardiologists who believed in the face of evidence that recovery in hospital after myocardial infarction was preferable to home.Footnote 43 His wartime experience narrowed his political activist ambitions from a broad commitment to socialism, to a sense that his aim to improve the lives of the population was more likely to be achieved within medicine, but through a radical version of it, and one committed to social justice: ‘I have given up on any attempt to change the world as I once wanted to do … I feel that I should just concentrate on changing a small bit of it.’Footnote 44 Cochrane’s peacetime work was predominantly conducted in Wales: following completion of a diploma in public health at the London School of Hygiene and Tropical Medicine (LSHTM), he took up a position running trials at the new Pneumoconiosis Unit at Cardiff, working on the celebrated Rhondda Fach studies, surveying coal-mining communities about their lung health. He said that working with communities gave him ‘a worm’s-eye view of the NHS’.Footnote 45 This perspective, shaped by working in one of the country’s poorer areas, with a significant health burden, committed him to a public-health perspective on the health service: ‘In particular I believe that the cure is rare while the need for care is widespread, and that the pursuit of cure at all costs may restrict the supply of care.’Footnote 46 In the 1960s, he found clear resonances between his own ideas and the work of philosopher Karl Popper, to the extent that he felt ‘taken over by his thought’: RCTs conformed, Cochrane believed, to an experiment that accorded broadly with Popper’s model of falsification.Footnote 47 Like Popper (but unlike Richard Doll), he became deeply sceptical of Soviet communism, primarily through his own experience of observing the Soviet forces in the Spanish Civil War, and later wrote a complex critique of Soviet medicine and the fact that communist countries had tended to eschew the evidence that could be derived from randomized control trials as a result of what he saw as both political and clinical authoritarianism.Footnote 48 In spite of his opposition to communism, his commitment to the ideal of socialized, universally accessible, state-funded health care as a base necessity for democratic society was a constant, and his enduring ideological position should be interpreted as a social-democratic stance. This centre left position is reinforced by reported feedback from Cochrane’s American colleague, Roger J. Bulger. When he visited the US in 1973, Cochrane reportedly quipped that it was ‘not fair to label him a socialist just because he advocated that all effective health care should be free’, and that a socialized health care system need not be an expensive endeavour: ‘Because so little of health care had been proved to be effective, there was plenty of room for capitalism after one distributed all the effective health care for free.’Footnote 49 As much as Cochrane avoided grand politics in later life and sought to use a rational experimental approach to minimize the effect of ideology and ‘bias’ in clinical epidemiology, his overall position on the normative nature of the health service still belied deep political commitments.
These, then, were the values that shaped Effectiveness and Efficiency, along with the hardened belief that the RCT was the best means of evaluating treatment. As the title implied, his objective was to go beyond the question of how to measure treatment efficacy in a trial context, to understanding how such evidence could contribute to better organization of the health service in terms of treatment choice and delivery. Cochrane targeted particular specialities, such as psychiatry and obstetrics, as problematic fields in terms of the evidence base for their treatments (not out of keeping with other British critiques of psychiatry in the same decades).Footnote 50 But the overall argument was not about the failures of specific examples. Its primary aim was to convey three important priorities for the NHS. First, treatments must have a secure evidence base, determined through RCT where possible, to be available within the service. This was as much about ethics as about any other criterion, to prevent harm and ensure that only effective treatments were available. Second, treatments must be cost-effective in terms of health economics: restrictions might be necessary in terms of rationing some approaches. Ultimately the service was taxpayer-funded and attention to efficiency in delivery of treatment and service was vital for the NHS to survive. Finally, the book was consistent with the original Bevanite vision of the NHS in that it advocated health equity, regardless of region or background. This argument also had implications for procurement: if citizens were entitled to equal health care in all parts of the nation, it followed that there had to be standardization of treatment options:
I have been convinced for some time about the final form in which any analysis of the over-all result of the various activities of the NHS should be expressed. If we are ever going to get ‘optimum’ results from our national expenditure on the NHS we must finally be able to express the results in the form of the benefit and the cost to the population of a particular type of activity, and the increased benefit that could be obtained if more money were made available.Footnote 51
Effectiveness and Efficiency was also timely: the 1970s saw the first major reorganization of the health service since its birth, and an increasing role for health economics in the Department of Health.Footnote 52 The argument for cost effectiveness and spending efficiencies was already being promoted from other quarters, not least from government. Cochrane’s book linked this up with a philosophy of evidence and treatment trials, providing a blueprint for how R & D budgets might be well spent. Nevertheless, it did not have a widespread positive reception at the time of publication. When it was reviewed, the focus was on the sense of exposé of inefficiencies in the health service, over and above the proposed solution.Footnote 53 It took two decades for the key message regarding evidence and the important of RCTs to become institutionalized in practice in a meaningful way.
Reception of Effectiveness and Efficiency and the Cochrane Collaboration
A major, lasting impact of Effectiveness and Efficiency was that it spurred the establishment of what came to be known as the Cochrane Collaboration. Obstetrician and epidemiologist Iain Chalmers encountered the book shortly after its publication, having recently returned from clinical service for the United Nations in a Palestinian refugee camp in Gaza. During his time there he reflected that ‘some of the things that I had been taught in medical school were probably lethally wrong. This came as a very sobering realisation: how could it be that health professionals acting with the best of intentions could do more harm than good …?’Footnote 54 Specifically, he was referring to the practice of not prescribing prophylactic antibiotics to infants with measles, which could be life-saving (as evidenced by later RCTs).Footnote 55 Upon return to hospital training at a Cardiff hospital in 1970, he was disconcerted by the fact that his consultants had differing opinions on the most appropriate treatment interventions in pregnancy. Discovering Effectiveness and Efficiency in 1972 confirmed his scepticism about the vagaries of clinical ‘authority’ unless it was backed up by trial data. He sought out contact with Cochrane, as well as training in the evaluation of the efficacy of treatments and services through the MSc in social medicine at LSHTM.Footnote 56 Like Cochrane and Doll, Chalmers became active in the Labour-affiliated Socialist Health Association (previously the Socialist Medical Association).Footnote 57
Chalmers worked to systematically gather and review trial data in perinatal medicine from the 1970s onwards, beginning in Cardiff. By 1978 he had secured funding from the World Health Organization and the Department of Health to establish a National Perinatal Epidemiology Unit at the University of Oxford, with the goal of assembling a register of controlled trials and later the circulation of up-to-date reviews of data via floppy disk. By 1992 this groundwork had led to support from Michael Peckham, then head of R & D for the Department of Health to establish the first Cochrane Centre at Oxford, expanding the systematic reviews to other aspects of health care.Footnote 58 The publicity and promotion of the initiative in the 1990s was very clearly linked to an implementation of Cochrane’s vision.Footnote 59
The Cochrane Collaboration was formally launched in 1993, a network that led to the establishment of an online Cochrane Library; support for the establishment of Cochrane centres in Germany, Spain, China and elsewhere; and a series of conferences involving international participants.Footnote 60 It also involved substantial community building, including the regular publication of a photocopied black-and-white newsletter with updates on work in each country, relevant meetings and invitations to participate in questionnaires about the development of the field – all illustrated with amusing line-drawn cartoons depicting people climbing the Mountain of Evidence, or researchers inputting search requests into 1990s desktop PCs.Footnote 61
A fundamental part of the collaboration’s intervention was the application of information technology and library science within the health service. As the number of trials and research articles proliferated, it was no longer possible to keep up with developments in a medical speciality purely by reading a small number of print copies of core journals. Scientific publishing was expanding internationally, and a significant challenge that needed to be overcome was access: according to an article in the British Medical Journal (BMJ) by Fiona Godlee in 1994 it was estimated that MEDLINE, the primary computerized database for medical research literature, was only successful at finding around half of the 40,000 RCTs that had been published in the previous nine years. Hand searches with the assistance of librarians continued to be required while digitization was still not the default for the publishing industry. It is worth noting that this article by Godlee, who was then a Harkness fellow at Harvard University, argued strongly for governments to support the aims of the Cochrane Collaboration. Godlee would go on to become the BMJ’s first female editor, for sixteen years, from 2005 to 2021, a position that gave her a considerable platform to further the agenda for EBM in Britain and beyond.Footnote 62 During the 1990s the scale of the Cochrane centres’ and the collaboration’s dissemination of systematic reviews of trial data went from circulated floppy disks to CD-ROMs to an online Internet platform. Its achievements encouraged Michael Peckham as the head of R & D for the Department of Health to believe that the establishment of a National Institute for Clinical Excellence (NICE), a body that could use large-scale data to evaluate the efficacy and cost-effectiveness of treatments, was a feasible project.Footnote 63
The National Institute for Clinical Excellence
NICE was established in 1998, supported by the Secretary of State for Health, Frank Dobson.Footnote 64 It brought together evaluations on the basis of systematic reviews, safety risk assessments and both published and unpublished trial data from medical-technology companies and pharmaceutical companies for the approval of new interventions (as well as sometimes evaluations of existing treatments on offer in the NHS, such as electroconvulsive therapy in psychiatric care). Panels with representation from all major stakeholders, including patient groups, would evaluate the evidence, along with analyses from health economists based on cost–benefit analysis (using quality-adjusted life years as the means of measurement).Footnote 65 A set of principles for the calculations were agreed from the outset, including guidance for decision making and definitions of how to estimate ‘total cost to the NHS’ and ‘maximum likely benefit’.Footnote 66 A key aim was also to limit the effects of the so-called ‘postcode lottery’, by which patients in different regions were not given access to the same treatments. Standardization across the nation was a requirement for equity.Footnote 67
Timmins, Rawlins and Appleby point out that there was not a legal obligation for enforcement of the guidelines – nevertheless they had a very significant impact on health service delivery:
NICE had been set up to be an advisory body to the NHS. There was nothing in legislation that required the service to implement its recommendations. However, in a cash-limited service and with individual doctors and health authorities, in the main, poorly placed to assess the value-for-money of new treatments, the service and its clinicians were only too happy – certainly in this instance – to be advised.Footnote 68
A number of historical witnesses involved in NICE’s establishment also remind us that the idea of setting up committees to consider the cost-effectiveness of treatments in the NHS dated back to the earlier 1990s Conservative government.Footnote 69 The decision to establish NICE went beyond the Department of Health, however, and had significant, visible support from the higher levels of the Cabinet, and was deliberately presented to the public as a valuable innovation in NHS management. This resonates very clearly with the wider policy context of the early New Labour years, where the agenda was oriented to what they actively described as ‘What Works’. This technocratic shift to evidence-based policy in the round was part of the party’s self-conscious reinvention, away from the more ideologically driven rhetoric of its socialist history, and instead towards a philosophy that was still based broadly on social-democratic principles of equity, but where the methods for achieving that had to be based on clear, validated data (a philosophical trajectory perhaps not unlike Cochrane’s own personal journey a few decades earlier). As such, the roles of experts in the departments, and scientific advisers, became far more important under Tony Blair than had been the case under John Major, and there was a greater platform for policy initiatives based on the implementation of evidence-based decision making, with a higher likelihood that they would be taken up by ministers. There was, nevertheless, bipartisan agreement across both political parties for the establishment of NICE.Footnote 70
Looking back over a longer period in the history of the NHS, Harrison and Ahmad argue that NICE was one of a number of NHS innovations in relation to management systems within the health service that eroded clinical autonomy, predominantly starting in the mid-1970s but growing in pace by the 1990s.Footnote 71 With regard to NICE in particular, they note the lack of opposition from clinicians to its introduction, which was perhaps mediated by media attention in relation to ‘the discovery in 1998 that paediatric cardiac surgeons in a Bristol hospital had a poor survival record for particular surgical procedures and that attempts by colleagues to draw attention to this had been obstructed’.Footnote 72 The linking of failures in practice to poor evidence for specific procedures was a compelling argument in favour of higher levels of regulation and validation, one which chimed with public opinion.
While clinical and public opinion may have been sympathetic at the start of NICE’s implementation, one stakeholder for whom the body’s decisions were extremely contentious was with the pharmaceutical industry. An oft-told anecdote circulates in the policy literature about a visit to Number 10 from Richard Sykes, chief executive of the pharmaceutical conglomerate then known as Glaxo Wellcome in October 1999, following NICE’s decision not to approve the inhaler-delivered anti-flu drug Relenza on the basis that the evidence for its claims to efficacy was too weak (having not been sufficiently tested with older patients or those who were more vulnerable, such as asthma patients). Given that an annual flu vaccine was available, this was not considered cost-effective given the slim evidence base. Prime Minister Tony Blair stayed the course and backed up the NICE decision in the face of Sykes’s threat to move the company out of the UK, saving the NHS £100 million in the process.Footnote 73
Certainly, this incident has been celebrated as a triumph of regulation over free-market interests. However, sociologists Abraham and Balendran have queried whether pharmaceutical stakeholders’ influence may be weighted more heavily than others’ in the panels, and Moncrieff and Tamimi, in relation to the guidelines on ADHD, have alleged that the panels are limited in their ability to challenge the clinical status quo.Footnote 74 John Boswell’s study on the guidelines for bariatric surgery suggests that NICE panels and the stakeholder process that leads to the guidelines serve as a ‘useful mythology’, which continues to elicit ‘buy-in’ from the respective stakeholders, enabling the process to function, albeit imperfectly and within limits.Footnote 75
EBM, NICE and the NHS: concluding arguments
The story of EBM in the NHS is one of ascendance by the turn of the twenty-first century, to a current situation where the RCT has become a norm for good practice in clinical and health services research, and cost–benefit analysis is ingrained into health care management through a guideline system which, in the British case, has more regulatory effect than guidelines used in other contexts, actively restricting choice and availability where it is deemed ‘fair’ to do so.
One striking aspect of this narrative, given the contemporary dominance of the RCT, is how many decades it took for it to become the consensus model of evidence – which is now the case not only within the UK but globally. The government policy landscape is obviously a clear explanatory factor for the incentivization of the use of RCTs since 1998. Prior to that, the self-styled ‘EBM movement’ had become prominent within the UK, the US and Canada, three of the main anglophone scientific nations which also hosted, collectively, the largest share of journal editorships, and had considerable international influence in terms of philanthropic and governmental donor funding for health research. The norms in these nations became cemented and extended through membership on funding panels, patterns of peer review and editorial practices that favoured the anglophone global North, and postcolonial and Commonwealth networks in global health training, cooperation and funding (such as through the international reach of programmes at, for example, LSHTM, Johns Hopkins and McMaster).
Jeanne Daly mobilizes an argument about ‘charisma’ based on Max Weber’s concept, accounting for the rise of the self-styled EBM ‘movement’ as linked to specific personalities. The memoir and oral-history sources, as well as the celebratory editorials and articles in mainstream journals which self-consciously commemorate Bradford Hill, Cochrane, Doll and latterly Chalmers as fathers or saints in the hagiography, certainly provide some confirmation of this hypothesis. However, the half-century gap between Bradford Hill’s streptomycin trial and the establishment of NICE queries this argument for charisma-as-cause. I posit that the crucial turning point was in significant part due to the content, format, style and context of Effectiveness and Efficiency as a document with appeal to a particular audience, its far-reaching reception, and the fact that it emerged from the support of a donor-backed, third-sector think tank whose board and membership also had considerable influence within academia and the health service. As a piece of scientific campaign literature, which is what it should be seen as, it was highly successful.Footnote 76 The existence of a consensus within a significant proportion of the medical and epidemiological community is, of course, significant, but the establishment of a national initiative on the scale of NICE also required a policy environment that was favourable to the arguments: Labour’s ‘What Works’ agenda could not have been more amenable.
Iain Chalmers’s own memoir note in the appendices to Cochrane’s One Man’s Medicine should also give us pause for thought. He reminisces about the process – and the amount of funding required – for conducting meta-analyses in the early stages of working towards the Cochrane Collaboration. Prior to the digitization of journals and trial data, funding was needed for a research assistant to facilitate the mechanics of collating the data from catalogues and printed journals before the analytic work could proceed – this labour-, cost- and time-intensive activity was in itself a barrier to realizing the goal of EBM, and one that the rapid implementation of computer-based technology in health care research was soon to remove. Even with the best of intentions, implementing hierarchies of evidence was something of a pipe dream until the IT resources were available to allow the work to be done. Neither the policy imperatives of NICE, nor the broader clinical emphasis on hierarchies of evidence, would have been possible to impose from the late 1990s without the contemporaneous availability of technological infrastructure.
While evidence-based medicine in terms of clinical effectiveness has become a ubiquitous global priority, the power invested in NICE to regulate approval on the basis of cost-effectiveness and health equity has something of a different quality to other nations in terms of the implications for governance of the health service. How should we characterize this mode of governance? It does not conform to technocracy in the normal sense of the word, as straight governance-by-experts. Experts drawn from clinical practice, epidemiology, health economics and patient groups as experts-by-experience, of course, populate the panels that deliver the guidelines, and there is an underlying technological system of RCTs, meta-analyses and cost–benefit analysis calculations that take the role of a non-human actor in the Latourian sense, with its own agency in shaping knowledge and practice.Footnote 77 This technological system exists, to use Cochrane’s words, to mitigate the ‘consultant God-complex’, to go beyond the problems created by simple governance-by-experts by providing as-near-objective-as-possible data that attempt to marginalize the subjective opinions of human actors, whose ‘expertise’ might be drawn from untested clinical experience and a sense of authority affirmed through social capital (in terms of holding senior positions in the hospital or university hierarchy, fellowship of professional bodies, and so on). In principle, EBM in the Cochrane mould actively problematizes the idea of traditional clinical expertise, undermining its authority and essentially denying its capacity for self-validation. In reality, however, as anthropological studies have shown, there are question marks over how far clinical orthodoxy is actually challenged, reminding us of the value for ethnographic work of querying how systems function in practice.Footnote 78
The way EBM is inscribed into NICE also, in theory, minimizes the persuasion and advertising tactics of pharmaceutical companies, evaluating value for money in a way that cuts against the principals of a free market. Yet the guidelines generated, and their consequences, are not simply governance-by-algorithm. The evidence generated by the non-human actor in EBM is still mediated by a panel of stakeholders. It aspires to be simultaneously democratic and authoritarian, involving patient and clinical representatives in the deliberation process, while its effects restrict their choice as consumers. It is designed to be in keeping with the principles of socialized health care with a view to delivering health equity and regulated markets, and limiting profit. At the same time, it respects the value of private enterprise in generating new health technologies, and provides a pathway for their conditional approval.
A number of social scientists, usually focusing on the Canadian case, have asked what kind of political economy EBM generates, citing managerialization as an outcome of the diminution of clinical autonomy.Footnote 79 The fact that health services research and the rise of health services management as a profession were developments contemporaneous in Britain with much of Cochrane’s work does not conflict with this assessment. In the most in-depth book-length analysis to date of the McMaster EBM research, sociologist Ariane Hanemaayer views Ontarian EBM as ultimately remaining within a mode of liberal governance, given that clinical guidelines are ultimately guidelines, and clinical autonomy remains protected, with practitioners themselves a self-regulated profession.Footnote 80 I would argue that the British case – despite self-regulation also being a fact of the medical profession – has aspects that are decidedly illiberal and centralized. NICE has the power not to approve the use of specific drugs or technologies within the health care system (thus prohibiting certain practices unless in private practice). For those that are approved, a managerial procurement level above individual clinicians (whether it be integrated care boards or previously clinical commissioning groups, or hospital formularies) can restrict the use of specific drugs or technologies on the basis of guidelines, cost-effectiveness measures and their own procurement practices. Clinical autonomy is eroded, and patient choice is restricted – but in the name of being able to deliver equitable health care free at the point of access.
For some more contemporary authors, the acceptance of scarcity as an unchallenged fact itself places an argument within the realms of neoliberalism. Technocratic modes of governance that limit treatment options on the basis of measurable efficacy, in a one-size-fits-all model, according to this view, inherently support neoliberal, small-state governance where responsibility is shifted onto the individual in the face of lack of state resourcing for health care.Footnote 81 The regulatory aspect of NICE as a technology approval body, which was linked to limiting the power of pharmaceutical companies in the health services, has been somewhat forgotten in this type of argument. The original intention was in many ways intended to defend socialized health care in opposition to neoliberal economics of choice and deregulated markets, and a number of key proponents of EBM in Britain, as we have seen, came from a social-democratic or even overtly socialist background. But policies often have unintended consequences, and a separate empirical article would be required to trace the sociological consequences of NICE in the health care service: research that would be significantly worthwhile.
The left-wing orientation of many British EBM proponents, and the active promotional role of the Labour-affiliated Socialist Medical Association (later the Socialist Health Association), of which many were members, does not diminish the fact that the NHS, despite being a socialized service, does not exist within a socialist political economy.Footnote 82 I argue that the most accurate characterization of EBM’s role within the NHS conforms with what sociologists Stephen Harrison and Waqar Ahmad have characterized as part of a ‘post-Fordist’ set of reforms which happened in the UK welfare state from the 1970s, drawing from the use of the term by Robin Murray in 1988. This accurately describes NICE as a body established by government with the managerialist purpose of negotiating between the corporate world and the health service, and accounts for how the impetus towards the establishment of a regulatory, validating body for medical technologies arose prior to the New Labour government, under John Major’s Conservative government, whilst at the same time conforming very clearly to New Labour’s What Works agenda.Footnote 83 NICE can simultaneously be read as a body limiting the free market, but also as one that facilitates the role of private pharmaceutical and medical technology companies – within reason and (in theory) only when in the public interest. This is the essence of post-Fordist regulation. Indeed, one of the critiques levelled at NICE more recently from Huseyin et al., published in The Lancet, rests on the argument that the money invested in these medical technologies in fact reduces public-health benefit, as the estimated quality-adjusted life years gained by investing the funds instead in primary care would far exceed those gained from the new technologies approved by NICE.Footnote 84 This argument overlooks the costs saved by the introduction of cost–benefit analysis in regulating what drugs and technologies can be used and which cannot, instead of leaving the decision in the hands of individual clinicians, hospitals or GP practices.
Debate remains about the implementation of the guidelines in terms of fairness, the quality of evidence, and how different stakeholders’ interests are weighted on the panels. Naci et al.’s critique opens up alternative ways of investing funding that might even be deemed more coherent with a social-democratic vision of health care, through emphasis on primary care and prevention, reminding us that choices about interpretation of evidence and allocation of resources are always intimately bound up with political and social values. To quote evidence-based general-practice advocate and Socialist Health Association member Julian Tudor Hart, ‘there is no evidence without ideology … Scientific evidence is derived from hypotheses conceived within an ideology – that is, a set of prior assumptions about the real world, established by previous evidence, by faith, or by both’.Footnote 85 Regulating cost-effectiveness, validating evidence of efficacy and ensuring health equity (at least at a regional level) were the driving principles of NICE when it was established in 1998. These ideological principles cohere very closely with the tripartite agenda of Cochrane’s 1971 Effectiveness and Efficiency, to the extent that they can be said to be the realization of this manifesto in policy terms: one which was guided by a faith in twentieth-century social democracy.
Acknowledgements
I am grateful to the special issue’s editors Lara Keuck and Angela Creager for involving me in the network and workshop at MPIWG that led to this special issue, and to Trish Hatton and Jahnavi Phalkey for their support, as well as the two anonymous reviewers for their feedback. Thanks go to Eoin Fullam for support with editing and formatting. The article would not have come about were it not for more than two decades of heated debates with Michael Marks (London School of Hygiene and Tropical Medicine), and for my stint as archivist at the Nuffield Trust for Health Services Research in 2010. The research was supported by a UKRI Future Leaders Fellowship (MR/X011143/1/10), the Wellcome Collaborative Award ‘Connecting Three Worlds: Socialism, Medicine and Global Health after WWII’ (221321/Z/20/Z), and an Honorary Research Fellowship at UCL Science and Technology Studies.
Competing interests
The author declares none.
Use of artificial intelligence (AI) tools
AI has not been used to generate content in this article. Claude.ai was used to gather secondary literature on the history and theory of technocracy.
Open access
