Few areas of health policy debate stir such passionate debate as user charges for health services. To some, it seems obvious that user charges are a barrier to access to needed health services, and should be removed in their entirety, as a matter of humanitarian principle. To others, it seems equally obvious that there are limited funds with which to pay for those services, an absence of charges leads to excess utilization, and therefore that imposition of user charges is an inevitable consequence of seeking to match demand and supply. Historically, the debate was given added impetus by the vigorous promotion of user charges as a policy instrument by the World Bank, a stance it has since abandoned (World Bank, 1993).

To examine the role of user charges for health services, it is perhaps first worth recalling what most commentators now consider to be the fundamental goals of the health system. In addition to the improvement of health, these include protection of citizens from the financial consequences of ill-health. It has taken a surprisingly long time for the importance of this somewhat banal insight to be properly recognized. Yet it can be argued that financial protection – rather than health improvement – has been the driving force behind the creation of many contemporary systems of universal health insurance, one of the great triumphs of 20th century social policy in the developed world. By protecting people from many of the financial consequences of disability and illness, such insurance contributes to human welfare in two fundamental ways: it reduces the anxiety and uncertainty associated with medical expenditures, and it improves access to needed health care. The World Health Organization crystallized the importance of financial protection in its World Health Report 2010. It presented the concept of ‘universal health coverage’ (UHC) as a fundamental policy lever for improving health, defined as providing all people with access to needed health services of sufficient quality to be effective, without imposing financial hardship (World Health Organization, 2010).

Such has been the success of social health insuranceFootnote ¹ in most high income countries that much of the public discourse seems to assume that free access to health services is the default condition. Many debates are therefore framed in terms of threats to comprehensive coverage, such as the ‘exclusion’ of certain treatments from the insured health basket, or the ‘imposition’ of user charges on certain services. Yet it is equally valid to turn round the argument, to consider the default condition to be a situation without insurance, in which citizens would either have to pay market prices for needed services, or go without treatment. The social insurance debate might then be framed in terms of which treatments should be included in the public health basket, and which user charges should be subsidized from public funds. This formulation of the debate is more common in lower income settings. Yet, whichever perspective is adopted, the substance of the issues remains unchanged.

At root, policy concerning user charges and social health insurance is an issue of public finance. If participation in social health insurance is in any sense voluntary, then the insurance pool is likely to break down, as healthier individuals opt out (Rothschild and Stiglitz, Reference Rothschild and Stiglitz1976). And if contributions are in any way related to levels of sickness or previous utilization, they effectively become a type of user charge, and the financial protection sought by universal insurance is diluted. Financial contributions to social health insurance must therefore be mandatory, and unrelated to use of services – in the jargon of health finance, they should be ‘community rated’. Such financing arrangements should properly be considered as taxation, regardless of the precise institutional form that they take.

The debate on UHC and user charges should consequently be informed by the relevant theory of public finance. Among many other issues, this should enable us to determine:

• - What the size of the benefits package should be?

• - Which specific treatments should be included in the package?

• - Who should finance the package? That is, what are the implicit financial transfers between different social groups?

• - What subsidies to market prices should be introduced, for which social and clinical groups, for which services?

A full policy analysis would require a rigorous examination of the incidence of costs and benefits of the chosen benefits package on different social groups (Canadian Institute for Health Information, 2013). Its main focus would be on who pays for the package (the incidence of taxation) and who gains (in the form of reduced user charges, improved access to health services and improved health).

It can be argued that some of the insurance benefits of UHC could in principle be achieved by implementing systems of mandatory private health insurance, under which every person paid an actuarially fair premium, based on his or her expected utilization of insured health services. In its rawest form, this system of ‘risk-rated’ insurance would mean that a person's financial contribution to health services would depend on factors such as their age, genetic inheritance and medical history. Payments would be pooled and smoothed over time, but only within a given risk class. There would be none of the financial redistribution from healthy to sick people implied by the use of community-rated premiums. Of course, the level of private health insurance purchased would also be heavily dependent on an individual's ability to pay, and – other things equal – the range and quality of services insured by a poor person would be inferior to those insured by an otherwise identical richer person. In short, private insurance would secure some elements of financial protection, but without the financial redistribution between social groups implicit in social health insurance.

Robust neoclassical economists might then argue that – if the concern is with redistribution from the rich, healthy and young to the poor, sick and old – then taxation policy alone could secure that redistribution, by transferring income between social groups in line with such equity concerns. Disadvantaged citizens would then have the means to purchase health insurance if they so wished, or spend the transfer on other preferred goods and services. Why, they might argue, should mandatory health insurance be used to secure such redistributive goals?

However, the manifest market failures associated with private health insurance – such as information weaknesses – have made this practically infeasible, and so almost all developed countries have favoured systems of social health insurance, under which mandatory financial contributions are made independently of expected utilization, and usually according to ability to pay. Such social health coverage has three distinct roles (Smith, Reference Smith2013):

• - An insurance function, reducing households’ financial risk associated with sickness.

• - A health improvement function, through enhanced access to needed health services (by reducing financial barriers).

• - A redistributive role, by contributing to societal equity objectives (through the implicit financial transfers from rich to poor and healthy to sick).

These benefits are likely to be immensely attractive to large sections of most electorates, and it is therefore hardly surprising to find that progress towards UHC enjoys very high levels of public support almost everywhere it has been tried. Once it is in place, politicians find it hard to challenge the principles underlying UHC, or to be seen to be diluting the benefits that it yields.

However, inexorable health care expenditure growth is a feature of countries at all levels of development. This is not necessarily a cause for concern if it accurately reflects popular preferences, and if potential symptoms of market failure, such as supplier-induced demand, can be held in check. However, expenditure growth can place extreme pressure on publicly funded health services, especially in the financially straitened circumstances in which many countries now find themselves. In most countries the growth in taxation (or equivalent revenue sources) needed to fund health care expenditure increases has therefore met resistance, leading some commentators to question the ‘sustainability’ of current social health insurance arrangements. Loosely interpreted, this argument suggests there is a limit to the extent to which society is prepared to countenance the financial redistribution towards sick or poor people implicit in any further growth of such insurance. It implies a judgement that it may become infeasible to sustain current pledges of universal and comprehensive coverage into the future.

So how should health systems respond to the challenges posed by these trends? Of course, seeking out efficiency improvements is an obvious way of mitigating the impact of the growth in demand, and there is consequently widespread interest in how to squeeze more value from existing resources (OECD, 2009). However, efficiency improvements on their own are unlikely to stem the expenditure rises. Attention is therefore also focusing on the potential for limiting the range of health services covered by the social insurance. If some treatments, or indications for treatment, can be excluded from the public package, then the level of coverage offered could be tailored to the resources available. Indeed many systems have accepted that some criterion such as cost-effectiveness should in principle be applied to decide which treatments and beneficiaries should be covered. However, it is proving immensely difficult in practice to apply the cost-effectiveness principle to the choice of the contents of the entire ‘health basket’ (Rumbold et al., Reference Rumbold, Alakeson and Smith2012).

With a few exceptions, such as Chile's Plan AUGE, systematic codification of the health basket has therefore been avoided. Instead, most health systems have used cost-effectiveness analysis to assess a relatively small number of treatments at the margin (especially new drugs), and have found it necessary to accept that the great majority of health services cannot be subject to rigorous assessment. Systematic efforts to exclude treatments have often been characterized as ‘rationing’, and have attracted widespread hostility. Policy makers have therefore preferred to pretend that coverage is comprehensive, in the sense that it secures access to all ‘needed’ treatments, where the definition of ‘needed’ remains vague (Schreyögg et al., Reference Schreyögg, Stargardt, Velasco-Garrido and Busse2005). However, when resources are limited, and no explicit limitation to the benefits package is in place, some sort of covert rationing must take place. This might take the form of diminution of the quality of the social coverage, non-price rationing (for example, in the form of waiting time), or implicit denial of treatment (for example, failure to make services available).

Which brings us back to user charges. A user charge is the most obvious rationing device to put in place in the face of budgetary constraints. If a treatment is excluded from the public benefits package, perhaps because of poor cost-effectiveness, it means that a patient must pay 100% of the price to secure access to treatment, in contrast to the zero charge attached to treatments in the public package. User charge policy can be refined to include intermediate copayments for treatments that are insufficiently cost-effective to be fully subsidized, but which justify some subsidy from public funds. These principles inform recent trends towards ‘value-based health insurance’ (Chernew et al., Reference Chernew, Rosen and Fendrick2007; Thomson et al., Reference Thomson, Schang and Chernew2013), and can perhaps be characterized as ‘value based user charges’. In theory, they are readily formalized mathematically to maximize the value derived from a fixed pool of insurance funds (Smith, Reference Smith2005).

Models such as this demonstrate the very close link between treatment charging policy and health technology assessment. In effect, treatments accepted for inclusion in the public benefits package by health technology assessment (HTA) agencies will be subsidized at 100% for users, while use of those rejected can be secured only by payment of the full market price. The only refinement required to align HTA models with public finance models of optimal user charges is to allow the option of intermediate (partial) payment by the user. Adoption of this variant would raise interesting issues for both HTA agencies and the suppliers of health technologies. HTA agencies would have to determine the price at which a technology would be fully subsidized, given the assessed level of benefits it confers. Any price in excess of that threshold would have to be borne by the user, along the lines of ‘top-up’ payments considered a policy option for certain treatments in England (Richards, Reference Richards2008). Also, suppliers of health treatments would have to determine an optimal price from their perspective, given their estimates of the price elasticity of demand among patients.

From a societal perspective, people deterred from treatment by such user charges represent a loss of potential health gain. And people who pay the user charge to secure treatment suffer an income-related loss of utility. However, the user fees secure income for the public benefits package, and the social insurance fund also experiences reduced expenditure because of those patients deterred from treatment. These expenditure benefits offer the opportunity to spend public insurance on patients seeking other treatments that offer higher levels of health gain. The (fiendishly complex) policy problem is to secure an optimal balance of charges and health gain across all potential treatments (Smith, Reference Smith2005).

Of course there are many other policy issues associated with user charges for health care. For example, they are an instrument for mitigating problems of moral hazard, supplier induced demand, and other inappropriate utilization of the public package of care. Or they could be levied on those seeking higher quality of care, for example in the form of reduced waiting or improved hospital facilities. Charges used for such purposes may lead to unintended consequences, and would require careful appraisal before being considered for policy.

Furthermore, even if a policymaker wishes to pursue a policy of ‘value based’ user charges, there are immense practical difficulties associated with implementing any scheme under which different treatments, and even possibly different patients, attract different levels of charge. Quite apart from the administrative challenge of such schemes, it would be necessary to engage providers in informing patients of the financial as well as clinical consequences of their choices. And perhaps most crucially the information requirements of developing a set of treatment-specific user charges are enormous, requiring information on cost-effectiveness that is often currently not available. It is for reasons such as this that most user charge schemes have hitherto been very crude, such as levying a fixed rate for a clinic visit, or for each day of a hospital stay, possibly with exemptions for low income or otherwise disadvantaged patients. It is highly unlikely that such crude systems of user charge will secure an optimal use of limited public insurance funds.

I nevertheless believe strongly that, as the call for UHC gathers strength, in a simultaneous context of growing demand and fiscal austerity, there is a need for a much more nuanced debate about user charges. This should recognize that public subsidy of health care should be used to maximum effect, and that allowing indiscriminate access to publicly funded health care may undermine that objective, for example by allowing privileged access to those who shout loudest, or those who have best knowledge of how to navigate the system, or even those who are able to offer ‘informal’ payment to providers. Instead, if limitations to access are deemed necessary, charges should be put in place to encourage optimal use of the available public funds.

Critics of value-based charging will argue that they would inhibit access to treatments by poor people. However, the real enemy of such access is the constrained size of the public package, which is a separate issue. My personal view, passionately held, is to advocate a generous budget that will allow the important triple objectives of health improvement, financial protection and financial redistribution to be pursued, with enormous gains for social welfare. However, that is my own value judgement. What is nevertheless clear is that progress towards UHC will be undermined if it is seen to be leading to dysfunctional outcomes, such as diminution of quality in public health care, or resistance to taxation among the young, the healthy or the wealthy. Therefore, if policymakers wish to make progress towards UHC it is imperative that they implement user charges appropriately, and in particular take careful note of the associated opportunities and pitfalls.