What market-based patient choice can’t do for the NHS

Competition and markets

The theory and evidence of how choice works in health care


1. Patient and user choice is at the forefront of the debate about the future direction of the provision of health and other public services in many industrialised countries. Specifically, in publicly funded and provided health care systems, where choice has been, or is perceived to have been historically lacking, increasing it has become a key policy objective. Introducing choice is also often seen as a response to a demand for individualised services on the part of the middle class. In other systems, however, the rising costs of health care have led policy makers to curtail rather than extend patient
choice. This was the rationale for the ‘managed care’ reforms introduced in the late 1980s in the market-based system in the USA, where insurers decide what services will be provided to patients, and for linking choice to higher co-payments in European social insurance-based systems.

2. In the English NHS, legislative changes have introduced an increased amount of choice in all aspects of patient care over the last decade. The first attempts to inject elements of choice date back to experiments with quasi-market reforms throughout the UK in the early 1990s. In 2003, under New Labour, patients in England (Scotland, Wales and Northern Ireland have followed different non-market approaches after devolution), were offered a choice of five providers for elective treatments – such as hip operations or cataracts; this was expanded to about 150 approved providers from public, private, or not-for-profit sector within an Extended Choice Network (ECN) in 2008. The Health and Social Care Act 2012 broadened the range of treatments
which patients could choose beyond elective services, mandated the inclusion in this of Any Qualified Provider (AQP), and assigned the exercise of choice to patients and their General Practitioners (GPs).

3. Two rationales typically underlie measures for widening consumer choice in publicly provided health systems: first, as a means to an end, stimulating providers to improve the quality of services offered; and second, as a good in its own right that is valued and desired by patients. The idea of patients as choosers shaping service provision has been transplanted into NHS policy from mainstream economic thought, but it also responds to long-standing demands by patient and user groups for autonomy and for greater control over the health care resources available to them.  Choice is also important whenever it enables patients to take better care of their lives, for example in dealing with long term conditions.

4. The aim of this report is to critically review the theory and the evidence of how choice works in health care. After presenting a brief outline of the NHS choice policies implemented in the last three decades, mainly in England, it discusses the theoretical premises underlying such policies and scrutinises the limited evidence of the impact of choice on efficiency, improved care outcomes and patient empowerment. It also explains why market based choices cannot work in health and puts forward proposals for alternative models of choice that are closer to the reality of patient care. The report concludes by recognising the value and importance of the variety of aspects involved in patient choice in health care and proposes a more balanced framework for the exercise of choice, taking account of users’ diverse needs and the resources they can actually draw on when making their health care-related decisions.

Choice and competition in the NHS in England

5. Choice was not on the NHS policy agenda until the introduction of market-oriented reforms in the 1990s. Until then patients could be referred wherever they or their General Practitioners (GPs) chose, although this option was little used. The introduction of market competition into the public sector in the UK was primarily designed to keep providers ‘on their toes’ and not at increasing patient choice. In the NHS specifically, the aim was to make services more responsive to users’ needs by giving Health Authorities a budget to contract services from hospitals which had to compete for contracts. At the same time, GP practices were encouraged to take up a portion of the budget to purchase some services for patients on their own lists, again requiring providers to compete. This was known as GP Fundholding. In practice this kind of choice was not vigorously pursued, resulting in lukewarm and isolated responses rather than a choice revolution.15 If anything choice of provider is likely to have diminished, because the internal market of the 1990s set up contracts with
specific hospitals, so that GPs and patients could only choose from among these.

6. Choice became a standard health policy objective under New Labour after 2002 following the introduction of a handful of pilot project for elective conditions with long waiting times such as coronary heart disease, followed by cataracts and orthopaedic surgery in the London Choice pilots. Patients were given an opportunity to make individual decisions where and when to have treatment. According to advocates of this policy giving the NHS patients free choice over these aspects of care would lead to more responsive care and promote their autonomy. Although there was no strong groundswell of opinion asking for choice of hospital, there was considerable public concern about waiting times and this choice has benefitted patients where the existing
services were very poor and had long waiting times.

7. In addition to improving quality and efficiency, the policy of offering ‘choice to all’ was intended to extend the opportunity to choose different health service providers beyond the articulate and those who could afford to access private health care. This second attempt at creating a market within a single payer public health system was justified on the basis of having to keep up with the alleged demands of patients who were increasingly thought of as consumers and who were expected to ‘reveal their preferences’ through choice. It was based on little evidence of whether market-based choice could work in health care services.

8. The Coalition government took this expanding commitment to market choice yet further, by requiring the newly created commissioners of health care services (CCGs) to contract for health care with a range of providers on a competitive basis in almost all circumstances. One of the aims of the Health and Social Care Act 2012 is to allow patients to choose the best services for their needs for any kind of service – whether from an NHS, third sector or independent sector provider. New providers entering the market will operate under the system of Any Qualified Provider and will be subject to the licensing requirements imposed by Monitor, the Care Quality Commission (CQC) and local commissioners (CCGs).

How far has choice improved service delivery and outcomes of care?

9. The evidence on the effects of choice in the 1990s Conservative internal market, in Labour’s subsequent choice reforms and in similar reforms in other comparable health systems introduced from the 1990s onwards, suggests at best a very limited impact in terms of efficiency and quality. There are also some indications that the measures intended to promote choice have had a negative impact on equity. Each of these issues is detailed next.

Choice as a way of driving up efficiency

10. In current policy debates in England competition between health care providers is seen as central to improving the efficiency of the NHS. Patient choice is a means to making competition work. That the users and commissioners of care will choose services from a competing provider, it is argued, will motivate providers to improve their service in order to attract patients. However, the analogy from the commercial sector does not readily apply in health care, where the introduction of economic incentives such as competition tends to have perverse effects. This is partly because the health sector is an imperfect market; many assumptions of economic supply and demand therefore do not hold. For instance, health care is rarely consumed for the sake of it, and users often base their choices on insufficient information or may be induced to make choices that suit providers (medical care of marginal if any benefit may be and often is provided, especially when there is a financial incentive to do so).

11. In marketised health systems administrative expenses are higher while additional losses of efficiency occur due to gaming of the system. For example providers may classify treatments as being more risky and expensive than they actually are in order to generate additional revenues.26 Some of these phenomena were observed under the internal market in the NHS in the 1990s. Thus a study of the effects of NHS reforms on hospital efficiency in England by a team from Bristol University, found some relatively small productivity gains during the period 1991–1994 for trusts that competed with each other when compared with those that did not.28 However, the same study showed that some hospitals became intentionally less productive shortly before
obtaining trust status, so as to be able to look more efficient under the new arrangements.

12. A review of the evidence on the impact of patient choice commissioned by Civitas concluded that increases in efficiency that occurred after 2002, (e.g. the increase in number of elective surgery patients treated as day cases, the decrease in the length of inpatient stays, and reductions in avoidable admissions), could not be attributed to patient choice as there were also other policies and trends which could have also encouraged such results. Recent research by the Centre for Health Economics at York University reveals that the NHS productivity took a clear step forward in the two financial years to March 2012, contradicting the views of inefficient public providers.

13. The evidence from other public health systems committed to universal health service provision (e.g. the Netherlands and Sweden) suggests that implementing regulated competition and choice is associated with an increase in costs. An increased supply of services followed the introduction of public competition in Sweden which, combined with freedom of choice for patients, made strategic priority setting and resource allocation by County Councils very difficult market in health care creates new threats to efficiency.

14. In the Netherlands, patients were also encouraged to choose among insurers in the context of regulated competition, but the expected benefits of competition such as cost containment and quality improvement were not achieved since the insurance companies focussed on expanding their market share rather on reducing costs and improving services. There is also evidence from many market-based systems (with the USA as a chief example), of providers tending to compete on quality by introducing expensive technology (particularly when they do not face hard budget constraints),
that leads to higher costs and squeezes out cost-effective care. Thus the evidence that choice leads to greater efficiency is not persuasive. It is also
difficult to single out a specific policy initiative as a ‘cause’ of a specific ‘effect’. Further, introducing competition between providers to improve efficiency relies on an implicit belief that existing NHS providers are intrinsically inefficient which has little basis in evidence.

Choice as a means of improving quality

15. Quality is an intrinsically difficult concept to define, having to do with process (e.g. waiting times) or experience as well as the outcome of care (e.g. improved clinical indicators). The economic assumptions driving the current NHS policy in England, is that where prices are fixed, providers will strive to attract patients by improving quality if the market contains a sufficient number of competitors: hospitals in these instances will compete in terms of quality and not price. However, studies measuring the relationship between competition and quality of care suggest that there are positive as well as negative consequences, while there is an agreement that the estimated impact of competition is small.

16. There are well-publicised cases under the recent reforms where competition between health care providers seeking patients who now have the freedom to choose is said to have improved quality. One such example is the ‘LSE research’ referred to by the Prime Minister in support of the market-based choice and competition enshrined in the Health and Social Care Act. This research suggests that the post-2006 choice reforms led to an increase in hospital quality, measured as slightly lower patient death rates from the acute myocardial infarction (a clinical condition commonly referred to as a heart attack) after they received treatment in geographical areas with greater potential competition between hospitals. These competitive pressures are attributed to the effects of patient choice initiatives, although patients exercise it mainly in relation to elective care, which was not the subject of the evaluation in this specific study. However, another piece of research assessing the impact of the internal market in England in the 1990s by researchers at Bristol
University found an association between the introduction of the market and increased, rather than reduced, mortality for patients admitted with the same condition.

17. There are ways in which quality can be driven up in hospitals other than through competitive mechanisms. For example, a study evaluating the introduction of pay for performance in all NHS hospitals in the North West region of England – whereby the top performing hospitals were given ‘bonus payments’ for outcomes in certain treatments, which was associated with a clinically significant reduction in mortality. Many other factors besides competition influence the quality of hospitals’ services including price structure, payment methods, internal organisation and pre-existing culture in addition to quality regulation systems and protocols. In reality, any impact on quality will depend on the precise institutional setting and on the regime of regulation. Overall, evidence of the impact of market choice and competition in terms of improved clinical outcomes is inconclusive since, alongside methodological weaknesses, the reported improvements are small or derive from a very narrowly defined set of indicators, and are often conducted under specific conditions that may not be universally replicated.

The adverse impact of choice on equity

18. Individual patient choice, introduced in England in 2003, had, as one of its aims, improving equity by removing barriers to access. It sought to do this by offering a choice of five providers directly to patients and not their agents (i.e. not to GPs, as was the case in the quasi-market reforms in the 1990s), and the choice was subsequently included for all providers on the Extended Choice Network (ECN). The idea was that giving all patients the option to choose a hospital or Independent Sector Treatment Centres (ISTC) run by private companies to have their treatment would offer them access to services they needed without incurring long waiting times. The New Labour policy makers claimed that their choice reforms enhanced equity of access by permitting those unable to afford private health care, a choice of provider already enjoyed by those who could afford to pay for it.

19. However, evaluations of the pilots introducing choice in various regions in England in 2002-2003 found that age, class, income and family obligations affected patients’ ability to travel to a non-local provider, and therefore their choices. Other studies reported no evidence of inequalities of access for patients participating in the same projects but these studies did not consider patients who were not offered choice. In many cases, choice was only offered to a minority of patients in the pilots (e.g. excluding older and sicker patients).23 For instance, the number of excluded patients who could potentially benefit from choice at Strategic Health Authority level ranged from under 1% to 70%, in December 2004. Another study found that many patients
did not experience the degree of choice that Choose and Book was designed to deliver as only a minority (approximately one third) of patients were given choice of referral at their first outpatient appointment.

20. Recent evaluations of the market reforms of the 1990s suggest that socioeconomic differences that lead to variations in health care utilisation are deeply ingrained, and that in the context of universal and comprehensive health systems small doses of ‘quasi market’ competition have little or no effect on socio-economic inequality in health care. Nevertheless, the risk for creating new inequalities over and above those that already exist is real. This can happen either because some patients receive preferential access and treatment under certain schemes (as was the case under the internal market in the UK with the patients of GP fund-holders obtaining a preferential access to hospitals with shorter waiting times); or because physicians are likely to modify their behaviour in order to fit the market, which could benefit some patients more than others.

21. The latter trend has been already noted in other health systems funded on the basis of universalism. Following the introduction of competition and choice, both primary care and hospital doctors in Sweden felt these enhanced their autonomy, income, and employment prospects but could also weaken their commitment to the normative foundation of the system namely equal access according to clinical need. A study from the Netherlands found that surgeons felt they had to sell themselves by advertising or marketing their performance after patients could choose between them. As this could be easier to demonstrate for relatively minor routine afflictions such as varicose veins and hernia and since they represented significant source of income for
hospitals, surgeons began to pay more attention to patients with such conditions following their preferences rather than medical need. The overall conclusion is that patient choice may exacerbate inequity of access due to the pre-existing inequalities of income, class and individual circumstances while the risk of individual choice leading to new inequalities also cannot be ruled out. Although users may be generally attracted to the idea of having a choice, research shows that not all groups of patients are able to exercise it in an equal measure.

Do patients want choice and need it to feel empowered?

22. In addition to choice being used as an instrument for achieving the policy goals of efficiency, better quality or even equity, choice is also seen by policy makers as promoting user empowerment and autonomy.51 The development of the active, critical consumer is considered as an important end in itself, even if they cannot always act as a perfectly informed agent.

23. There is some evidence supporting this claim. A 2012 survey of patients’ involvement in health across the EU found that a desire for a more balanced
relationship with their doctors allowing patients to participate more actively in their care, was particularly strong for younger people, well-educated patients and people with chronic conditions, and more so in Western than in Eastern or Southern Europe. However, patients’ need for more information and a more equal relation with health care professionals does not necessarily imply a demand for more choice. Though ‘having choice’ might be desired by working-class respondents in England at least as much as by middle class respondents, since this in theory might give them access to high quality of care they may otherwise be lacking, patients from disadvantaged backgrounds are less able and less likely to benefit from it.

24. In reality, patients’ individual characteristics and their circumstances are likely to not only influence their choices but also decide whether they will exercise them at all. Thus an evaluation of the London Choice project found that old age, low education level, family commitments or low income all had impact on patients’ choice of a nonlocal hospital – meaning that they are less likely to travel to a non-local hospital if they are offered the choice. Distance remains an issue for many – patients want services locally for themselves and for family reasons while lack of public transport can make choice difficult for people who are unable to afford a car. One exception to this are older patients who already use health services more frequently, and people living outside urban centres discovered in another study investigating patients willingness to choose primary care physicians. Both groups were more likely to opt to travel if offered a choice of hospital, probably because they have already had a history of travelling and/or have faced longer commutes to health facilities.

25. Furthermore, research by the King’s Fund shows that differential access to information by less educated and those for whom English is not their first language, could lead to variations in uptake of choice according to social class, education level and ethnic group. Another recent empirical study by the King’s Fund concluded that patients in England who are not highly numerate and health-literate are less able to use the available information to make complex decisions regarding hospital choice without some expert support. Thus comprehending the options and making tradeoffs between quality, safety, patient experience and location posed difficulties but the way information was presented also made a difference to how patients used it.

26. The type and degree of choice patients want and value is not self-evident either. Research by Which? in 2005 found that choice is a relatively low priority for many patients compared with other aspects of health delivery in the NHS. Rather than choice of provider the majority of patients in England are more concerned to have safe, good quality services provided locally. Contrary to policy makers’ beliefs, patients tend to favour a provider they know and trust and opt for choice only when no such provider is available. Put differently, policy-makers conceive of choice as an ability to select different providers, whereas for many patients it is about choosing the same local provider over and again. Often patients appear to be more interested in choosing treatments. These attitudes are however often influenced by the severity of the medical condition and the complexity of the procedure involved: The more life threatening the disease and technologically advanced the treatment, the less is the patient’s desire for choice.

27. Retaining the public and universal aspects of the health system is another concern overriding any desire for choice for patients across the UK. When ranked on a scale of one to five in a recent MORI survey, for the majority (63%) of the British population fairness in public services comes first, whilst choice and the personalisation of services is last. Although users of health services may see policy initiatives aiming to introduce choice as offering scope for improving services or their access to them, changes that seem to undermine the founding principles of the NHS are unlikely to find favour. Overall, in MORI polls, public satisfaction with the NHS at a national level, and patient satisfaction, have remained relatively stable since 2000, though patient ratings of their treatment were far higher than ratings for the NHS as a whole. Both were high and on an upward trajectory when choice was rolled out in the NHS in 2008.

28. The evidence across Europe suggests that in addition to satisfaction with the health system, perceptions of choice are influenced by personal health situation, age and gender. For instance, older patients in Sweden were both interested in choice of primary care doctor and happy about the amount of choice offered, while highly educated young people, and women in particular, were found to both exercise and favour choice more when compared to other population groups. These age and gender factors are also confirmed for England, but patients from less advantaged backgrounds express very high support for choice. In sum, choice is desirable but not essential for patients to feel empowered, and their need for high quality accessible services provided locally supersedes it. Patients’ desire for more involvement in their health care does not automatically translate into having more options to choose from but rather for more support to make informed decisions.

The limitations of the market choice: How patient choice works in reality?

29. The market choice introduced in England and in comparable health systems with commitment to universality and equality of access was shown to produce few, if any benefits, while creating potentially undesirable effects. The theory of market imperfections in health care and accounts of how choices are made in reality, demonstrate the problems of replicating simplistic economic choice models in health care. Choice policies, as empirical results demonstrate, rarely lead to more efficiency, are likely to have negative consequences on equity and fail to meet patients’ most important interests for a local provision of services and choosing treatments.

30. The necessary pre-conditions for a competitive market rarely apply in health care since health is not a commodity that can be easily sold and  exchanged, health care markets are rarely competitive, and patients often lack information needed to make choices. This may apply more in respect of some types of care than others. For example, patients with long term conditions may be more able to make informed choices. However, the narrative of knowledgeable users of public services exercising their preferences via acts of consumption overlooks something that is actually central to health care choice in real life: the patient’s need for trust-based relationships with care providers. Precisely because patients lack the information needed to make
informed choices about their care, they need medical professionals they can trust; this overrides their desire to ‘shop around’.

31. More to the point, even in material markets people are seldom rational choosers and least of all in relation to health services. Individuals do not always choose what is in their best interest even if they are able to identify it – allowing them to make decisions which are acceptable to them but which may not be entirely rational – a reality that economists have now come to acknowledge. For patients, the severity of their medical condition amplifies the bias in processing information that the human mind is prone to even further.

32. Choice means different things to different or the same people at various points in time because users of services share multiple identities as citizens, family and community members, members of religions, and much more. Patients’ ability, and even their willingness to make choices, is influenced by their beliefs, cultural values and expectations as well as their life circumstances, personal characteristics and their experiences of health care services. Put differently, the individual choices we make are socially constructed.

33. Although it is possible to treat people who seek professional help as customers this is incompatible with ways of thinking and acting that are crucial to health care. Good care grows out of collaborative and continuing attempts to attune professional knowledge and technologies to diseased bodies and complex lives. When making complex health decisions, patients rely on their intuition and emotions involving the avoidance of regret as well as trusted networks, rather than objective, impersonal data. To sum up, framing the issue of choice in the context of market competition roots it in old-school neo-classical economics and involves a significant narrowing of the concept of choice, and of the users of health services as rational ‘choosers’ exercising their preferences. Choice and independence are indeed powerful concepts, but interdependency is an essential part of social life and never more so than in the relationship between patient and clinician.

What type of choice should be available to patients in the future NHS?

34. Despite its commitment to introducing the market fully in health care the current government is unlikely to be able to offer patients costly choices. The decision to hold down NHS spending over the coming years makes this an inescapable reality. The ability of patients to choose from a range of alternatives for any treatment and at each stage of their treatment is therefore likely to be constrained. In the competitive health care market that the Health and Social Care Act has expanded some patient choice will exist (under the Any Qualified Provider process) but it will still be subordinated to commissioners’ priorities, as has often been the case in the past (notably in experiments with the quasi-markets in health systems across Europe).

35. Yet the desire of service users for more autonomy and greater control over the health care they receive should not be discarded together with the consumerist market model but rather addressed on its own terms. Patients are often obliged, and in many cases are increasingly willing, to make health-related decisions as coproducers of their health together with health care professionals, and also as citizens and community members co-designing health services; but these choices are governed by social values and the need for co-operation and recognition, not by mere self interest. As individuals they are involved in co-producing their care; as taxpayers they are concerned about the allocation of resources, and as community members they rely on family and friends for support regarding their health-related decisions. However, patients’ involvement is most effective when used as part of a broader ethos of care.

36. All these factors need to be taken into account. On the individual level, there is considerable scope for developing existing shared decision-making methodologies and supportive devices that can enable patients to reach optimal clinical decisions to meet their needs within available resources.75 Extending choice could be achieved by, for instance, capitalising on patients’ pre-existing experience with chronic disease management programmes and their desire for more involvement in their own health. The London Choice Projects used two strategies to improve decisions with some success: patient care advisers, and decision aids.

37. There are also various practical ways of proceeding with the implementation of non-market choice by strengthening elements of ‘voice’ in the system. This way of giving voice and allowing patients to become co-producers of their health is different from the personalization agenda invoked by politicians. Collective mechanisms involving users’ in the governance of choice-based health care arrangements could be implemented through user organisations or by setting up bodies with an ombudsman function as another means whereby ‘voice’ can be strengthened.

38. The example of co-production of public services could also provide further guidance. In such arrangements service users are active asset-holders of resources and experiential knowledge rather than passive consumers; collaborative rather than paternalistic relationships between staff and service users is promoted and the focus is on the delivery of outcomes rather than the services. In the UK’s health and social care services co-production has gone beyond user consultation towards developing ways of service delivery intended to impact on service users and on wider social systems.

39. The expansion of choice can empower patients, if it is properly linked to their direct participation in the decision-making processes, for instance by involving them (individually or collectively) in deciding about the allocation of available health care resources within a community. This would also apply in the case of patients with long term conditions for which co-management by patients and professionals works best. In all cases patients and users of services should be clear what is involved in their choices, and what the potential consequences are – not just for their immediate care, but for
the future provision of care for them and their families and community. This is a different way of appealing to users’ ‘rationality’ – as co-producers of their health, citizens-taxpayers and community members.


40. Promoting market-based individual patient choice, first introduced in the 1990s, has now become a standard health policy objective in the NHS in England. The passing of the Health and Social Care Act 2012 means that this trend is set to continue. The idea of patient choice in health services is founded on two general assumptions: one is that it will aid competitive markets in their tasks to improve the efficiency of providers as well as improve quality; the other is that the exercise of choice is an important good in itself for patients. But the assumptions on which the policy rests have been found
wanting. Their applicability is either severely limited or invalid when applied to health care, for both theoretical and empirical reasons.

41. Market choice is therefore not likely to provide more efficient services and any improvements in quality will come at a cost and will depend on the precise institutional setting where it is implemented. It also carries a risk of reproducing old inequalities while introducing new ones having to do with health literacy and access to information. However, financial constraints make it unlikely that the support mechanisms needed for choice to be exercised equitably will be funded. Overall, and with the exception of those whose access to services is not constrained by their educational status and ability to pay, choice is unlikely to benefit patients with low health literacy in the absence of appropriate institutional arrangements.

42. The type of patient choice that contemporary health policy draws on almost exclusively is based on a flawed account of what choices actually mean for patients. Such conception of choice rests on the simplistic and erroneous assumption that appealing to patients’ self-interest will make them behave as consumers in a market place. This model has now been widely discredited, even among economists. Policies based on this assumption ignore the reality of patients having various needs, more so in times of dislocation, vulnerability and stress. Lastly, instrumental approaches to choice tend to neglect the intrinsic value trust has for care relations and its fragility that is easily undermined by egoistic action.

43. For choice to work, policy design needs to consider users’ prior experiences, including their experience of health services, and other social and psychological factors affecting their health-related decisions, in addition to their relational and social bonds as family and community members. Policies must also foster public trust in the health system and health organisations, and also protect trust between patient and doctor or nurse, without which care is impossible. In order to achieve this, policy makers should draw on interdisciplinary frameworks and alternatives to market mechanisms in health
care offering a more balanced view of how choice works in reality, and what choices matter to patients. Only then can choice be employed in the provision of responsive and effective health care.

This article was first published by the Centre for Health and the Public Interest and is reproduced with their kind permission. The version on their site is fuller and referenced.