The problems with governance: budgets and blame avoidance

The term governance has risen to prominence over the past two decades, in essence to provide ‘an all-embracing concept capable of conveying diverse meanings not covered by the traditional term “government” (European Union, 2007). In practice, this means that the term governance is flexible enough to obscure questions of authority.

This focus on governance has been associated with the rise of the new public management. The new public management involves ‘decentralisation of management competencies’ (Vrangbaek, this issue) or a ‘commitment to letting managers manage and to entrepreneurship’ (Thompson, Reference Thompson2002: 362). In the words of the American National Performance Review of 1993, rather than being governed by rules and procedures, agencies should ‘delegate, decentralize, and empower employees to make decisions. This will let frontline staff and front-office workers use their creative judgment as they offer service to customers to solve problems’ (quoted in Considine and Lewis, Reference Considine and Lewis2003: 134).

Ironically, as noted by Tuohy in the quote above, the old management of health care, especially the Beveridge model in which the state sets the budget and leaves the physicians, in Burau and Wilsford's words, to ‘depoliticise the rationing of health care', looks remarkably like the ‘New Public Management’ in its emphasis on indirect methods. Certainly physicians were ‘empowered’, in the sense used by the American National Performance Review. If the general trend is away from hierarchy, why, then, would many health care systems appear, in this issue, to be moving (though quite haltingly) towards hierarchy, with increasing regulation of the ‘front-office workers’, otherwise known as physicians?

One answer involves a general trend in elite attitudes about social action: belief in what Vrangbaek in this issue calls the ‘audit society’ in which discretion must be justified by measurement of results. A second reason is that policy elites have come to believe that individual physicians and medical care organisations are not delivering quality care in a reliable manner. Elites are influenced by health services research that shows either a lack of evidence for treatments or instances of blatant mistreatment (the data about medical errors, particularly adverse events in hospitals). In some cases, this message was amplified by focusing events, such as the death of a reporter within a prestigious Boston hospital (Millenson Reference Millenson, Mechanic, Rogut, Colby and Knickman2005), or the Bristol scandal in the National Health Service (NHS).

Yet, a third reason should be emphasised. As Burau and Wilsford note, policy makers are interested in the ‘micro-level of governing health service delivery’ because ‘doctors control the delivery of services and are the pivot point around which the allocation of financial resources flows’. Talk may be of quality, and quality is a real concern, but quality remains subservient to cost control. In the US, managing treatments sounds attractive because it offers a blame-free form of cost control: patients should not complain because they are only losing unnecessary care, and providers should not complain if their incomes are reduced because they are not doing unnecessary care. In other contexts, where prices and capacity are regulated more successfully and costs are better controlled, a wish for further savings is supplemented by a hope that changes in health service delivery would help politicians avoid blame for service shortfalls, especially waiting lists.

In an era of fiscal stress, proponents of reinventing or re-engineering government argue that improved management can improve performance and so avoid the need to spend as much (Thompson, Reference Thompson2002: 362). Health care is just another example of the pattern. The original NHS internal market was an early attempt to respond to public complaints about what the public really cares about, service shortfalls as indicated by waiting lists, through management, variously rationalised as either hierarchy or market. In Denmark, Vrangbaek writes, ‘national level politicians were blaming the counties for emerging evidence of poor performance in regards to waiting times and more generally in regards to the relative decline in life expectancy’. The available evidence did not suggest that health care system performance was an important cause of the life expectancy results (Pedersen et al., Reference Pedersen, Christiansen and Bech2005: S53), while the central government had imposed on the counties a very tight fiscal policy that would be expected to squeeze capacity (Pedersen et al., Reference Pedersen, Christiansen and Bech2005: S45). However, one can hardly expect national policy makers to blame their own budgeting when they can blame county health administrators instead.

The legitimate case for concern about quality then provides the basis for elite rationalisations about policies whose real subject is cost. That pattern persists in odd ways. In both Canada and the UK, public objections to the results of cost constraint eventually convinced ruling parties to seek re-election by pledging to increase health care spending substantially. In each case, policy makers announced measures to, supposedly, make sure they would get value for the extra money, such as the National Institute for Clinical Excellence (NICE) and ‘pay for performance’ in the UK (Maynard, Reference Maynard2005), and fairly weak requirements that provinces use the money to hit policy targets in Canada (Wong, Reference Wong2005). The rhetoric of measurement and medical governance allowed politicians, central budgeters, and health policy elites to rationalise giving in to public demands for spending.

In short, the emphasis on medical governance, whether spending is going up or down, appears to have more to do with budget politics and blame avoidance than with serious concern with quality. One reason is that purchasers simply care more about costs. In the US, productivity incentives are much more common and significant (from the doctors' perspective) aspects of contracts with physicians than are quality incentives, for many reasons including that it is a lot easier to measure productivity or costs than quality (Reschovsky and Hadley, Reference Reschovsky and Hadley2007). Employers are not so interested in ensuring quality through the market if high-performance networks do not save money (Draper et al., Reference Draper, Liebhaber and Ginsburg2007).

Similarly, in the countries studied in this volume, policy makers care somewhat about quality, but much more about other things. In both Italy and Denmark, health policy concerns are partially subservient to the ideological politics about devolution or centralisation, respectively. Where measures are adopted and even enforced, they tend to emphasise costs; thus, in Italy, the focus is on standards for spending on prescriptions. When conflicts occur about details, pursuit of clinical governance measures tend to be traded off or subsumed in deals about other values. In the UK, the advocates of stricter measures backed off in the context of negotiations over a new contract for consultants. In Germany, support for measures was in the context of the broader difficulties of establishing some sort of competition among sickness funds (as it also was related to choice among hospitals in Denmark). Public attitudes appear to have been relatively insignificant unless some other elite actor came along and whipped them up, as in the Fifth Report of the Shipman Inquiry. Fenton and Salter, in a footnote, report that this upset the applecart in the UK, putting the settlement described in the main body of their text in flux as of mid-2007.

What the non-American cases reveal is that approaches to governance may appear to be principles, but instead are often tactics to win short-term advantage. The discourse of the time – ideational regimes – may encourage certain kinds of rhetoric, or even approaches. However, when initiatives are also rationalisations and excuses, seeking to characterise governance by reviewing a small set of initiatives for their ‘fit’ with models is a risky business.

Policy capacities

Successful policy requires three kinds of capacities. The first is technical capacity: knowledge of how to do something. Much of the clinical governance agenda founders on the fact that it is very hard to define conditions and treatments in ways that reduce uncertainty enough to expect adherence to guidelines. The second is institutional capacity, or power. Even if guidelines can be defined, there may be no institution with the power to oversee their application. This is especially the case if physicians are practicing in widely-dispersed offices with sketchy records. The third is political capacity, or will. Even if guidelines can be defined and administered, this form of governance will not occur if the objects of governance resist and the advocates of governance cannot force acceptance.

Governance of medical treatments reaches policy agendas because medical providers have trouble rationalising opposition to some of its ideas. Policy makers who have been battling providers over other cost-control measures, or contemplating the political pain of benefit cuts, hope that making care more appropriate will be less difficult. Yet, the difficulties become more obvious if we compare medical governance to the two mainline cost-control methods: regulation of prices and limiting physical capacity.

Price setting is as technically hard as the price setter wants to make it. If one pretends to need to have a right price based on perfect evaluation of inputs and value, it is very hard; but that approach is rare. Prices are normally set as increments or decrements from previous prices, or by analogy to similar services, by bargaining to set relative values and by mandating conversion factors (Glaser, Reference Glaser1994). Once a fee schedule is established, enforcement is relatively easy. The primary payer can tell what price it pays, and at least some patients will complain if providers try to extra-bill. Providers may try to game the system by mis-categorising services, as with the ‘bracket creep’ in the American DRG (diagnosis related group) system. However, this can be handled, although imperfectly, by various methods of volume-adjusting fees.

Now contrast the technical and institutional challenges of price setting to the challenges of managing care. It is much harder to create rules about treatments than to declare a price. Moreover, price setting does not require that the diagnosis be correct, or that technical judgements about the specific needs of patients be reviewed. Enforcing guidelines is hard because it requires a much more detailed oversight of what physicians do. It is even difficult to come up with a plausible definition of some of the desired methods. As the German Advisory Council for the Concerted Action in Health Care reported, ‘a clear and unequivocal definition of disease management is to be found neither in the law nor in the international literature’ (2003: 79). Similarly, reform of the care processes within institutions requires far more power – the power to change behaviour and the power to observe compliance or deviation – than is needed to set a budget for the same clinical service.

Comparing management of individual treatments to managing the capital stock of a system tells the same story. Much of the research that can be used for analysis is about averages. Technically, it is more appropriate to make the argument that there is not enough need, in total, to justify a new MRI machine, than to establish whether scans are appropriate for an individual patient. It is much easier for outsiders to analyse the relative values of investing in hip replacements or cardiac surgery capacity in the aggregate, than for them to set rules about how physicians should choose among services for individual patients. In short, the knowledge base for investment decisions is stronger than the knowledge base for clinical governance. Moreover, there are many fewer investment decisions and, in most systems, clinicians have to go to some sort of managers or policy makers to get the funds and approval for investments. In contrast, there are millions of clinical decisions made by physicians every day. Thus, managers and policy makers have much more technical and institutional capacity to influence capacity choices than to shape clinical choices.

So, does the common rhetoric about the needs for evidence and quality suggest that the political advantages of that clinical governance outweigh the technical and institutional disadvantages? Apparently not, since price setting and capacity limits still appear to be implemented far more fully and effectively (though by no means in an ideal manner!). One reason is that even the will to implement medical governance may be weak. That brings us to the subject of authority.

Management authority

The problem is how to influence the behaviour of care givers in a desired direction. Economists might say the problem is how to create incentives for care givers to behave in the desired way. I think it is more fruitful to see this as a question of authority.

Following Herbert Simon (1997), we may say that one person exercises authority over another when the first sets premises that the second uses in making decisions. These decision premises may involve questions of fact or questions of value. For example, both views of clinical evidence described in Burau's chapter on Germany – that evidence should be based on clinical studies, and that evidence can be based on the experience of senior medical practitioners – involve situations where clinicians will find fact premises about treatment effects. They are different versions of expert authority. A simple version of authority over value premises occurs when system managers impose a fee schedule, forcing physicians to accept somebody else's judgement of what their services are worth.

People in any system or organisation accept some organisational restrictions in order to receive rewards (or avoid sanctions). Authority, then, is provisional because the object of authority can always opt out, sacrificing the reward or enduring the sanction. That is why, as Simon argues, authority is defined by the ‘zone of acceptance’ of its object; authority is defined from below (what its object will accept) rather than above (because the object of authority has the final choice). Physicians, for instance, can quit the profession.

As my example of fee schedules suggests, we should not imagine that authority is accepted gleefully. In all systems, people are more likely to accept things that they don't think they can change. When conditions endure long enough, people may adjust their expectations of what is right to fit what is possible; thus, in a highly capital-constrained NHS, it may have been easier for doctors to adopt conservative views about treatment (Payer, Reference Payer1996). We also should not imagine that zones of acceptance are stable. In the US, in the mid-1990s, physicians and hospitals grudgingly accepted terms that, when perceptions of the relative powers of the contracting parties changed, they then stopped accepting (White, Reference White2007). In negotiations between the state and the medical profession, physicians may recognise that the government of the day is unfriendly yet secure, so they accept policies they would not accept under other circumstances. Moreover, authority over one domain can sometimes be translated into power, if not yet accepted authority, over another domain, through trade.

At the same time, the ability to supply fact premises has natural limits. Engineers may accept from a corporation's accounting staff premises about how they should account for spending on a project, but not calculations of the stresses on a bridge. Anyone who seeks to influence physicians by setting fact premises about treatments must have some plausible claim to expertise. Authority to be effective must also be exercised with confidence, which means those who set premises must be sure of their own standing to do so.

What happens then, when policy makers of some sort try to alter the existing division of decision power in health care by exerting some new authority over medical practice? The first thing we should expect is that, because it is a new effort, it is not within the pre-existing zone of acceptance, so it will be contested. That applies to all the initiatives in this volume.

Second, supply of fact premises means, at best, that policy makers need to get some doctors to supply premises for other doctors. In spite of the rise of health services research, this means that ultimately clinical guidelines end up being validated in some way by panels that include clinicians. Similarly, any validation process gives a major role to physicians. The dispute about revalidation in the UK centred on which peers would review performance and the extent to which the system could be overseen in any way by outsiders (thus, the consultants did not want explicit standards). But it had to be operated by clinicians in either plan. In Italy, hospital doctors are evaluated every three years by panels of physicians. That, in turn, brings the internal governance of the physician community to the forefront.

Thus, third, divisions within the clinician community may be exploited so that policy makers can use one group to help develop standards for another group, which appears to have been at least an implicit threat in Germany. Yet, even that is difficult, and in the German case, acceptance did not extend as far as any system that would involve meaningful sanctions.

More generally, fourth, the ability to regulate doctors depends on the structure of authority within the medical profession as a whole, or in specific medical delivery institutions. The concept of professional self-governance includes two very different situations. One is essentially professional non-governance; professional status insulates some aspects of each clinician's activity from any outside authority, and the profession itself does not pick up the slack. The second involves serious efforts by the profession, essentially as a guild, to exercise authority over the performance of its members. Long before the clinical guideline movement, citizens and policy makers expected the medical profession as a guild to ensure the quality of its members, through regulation of entry (e.g., medical education and licensure) and some monitoring of performance (e.g., hospital medical boards). However, nations vary in the extent to which they have developed authority structures within the profession that can meet those expectations.

An American reader may look at the descriptions of Denmark and Germany in this volume and develop a slight case of authority envy. There is no American equivalent of the Danish National Board of Health, with a corps of health officers and an administrative system of malpractice regulation. This reader would like to know more about how the system performs, though it is clear that extending this regulatory structure to creation and enforcement of clinical guidelines is difficult. Yet, there is at least a structure with which to start. Similarly, the degree of governing responsibilities exercised by the German medical profession far exceeds any policy maker's dream in the US. In principle, ‘the service guarantee for the medical outpatient treatment lies within the Federal Association of SHI Physicians and the regional physicians’ associations’ (Worz and Busse, Reference Worz and Busse2005: S134). As Burau writes, in this corporatist health care state, physicians are directly involved in the implementation of the insurance system. While physicians will resent some policy initiatives, at least organisational leaders have come to expect to participate, and thus to value participation.

Leaders of either interest groups or corporate bodies must always balance the need to be a player with the need to maintain support from their members. Where participation is more of a norm (so accepted by the members), leaders have more room to cooperate. Thus, the Federal Chamber of Doctors, which is not involved in the joint self-administration, was less willing to participate than was the Federal Association of Sickness Fund Doctors. Yet, in spite of the value they set on participation, the Federal Association's leaders also reached a point where they could not concede further. Self-governance does not mean the profession abdicates to outsiders. It can mean that there is a greater prospect for developing and implementing policies regarding practice than exists if the self-governance structures are weaker.

Finally, authority is most easily extended if it can be exercised in tandem with a pre-existing structure of authority over other matters. The sickness fund physicians in Germany have other stakes in negotiations with the insurers and (at one step removed) the federal government, and this makes them more willing to engage in negotiations over disease management. This occurs in a context where the state has long engaged in what Burau calls ‘steering at a distance’ on other matters. Just as the sickness fund and insurer parties to the negotiations about disease management worry that the state could step in and impose its own terms, negotiations about cost control have long been within a context of guidance about what contribution rates can be allowed. Although combining issues in a pre-existing arena may extend authority, it does not always do so; for example, in the UK, negotiations over the consultants’ contract provided a context for weakening revalidation initiatives.

To some extent, therefore, authority patterns for medical governance replicate patterns for other matters. In Italy, as Tousijn and Giorgino describe, everything becomes a matter for negotiation and perhaps party politics. In a context where, ‘personnel was not managed, or it was managed using criteria based on political and trade union affiliation’, it should not be surprising that ‘intense negotiations even occur regarding the evaluation of all medical managers and eventual sanctions for non-attainment of targets’. Yet, the difference in the ability to set fact premises between governance of treatments and governance of prices and capacity remains fundamental. The former is simply not possible without involving some segment of the medical profession in making all the technical judgements, and then relying on another portion of the profession for implementation. Political leaders can enunciate basic principles and directions, but implementation is delayed for years, and may never occur, as the details are worked out (or not). We see that pattern in all four cases.

Similarities and differences in the US

The technical and sociological aspects of medicine are much the same around the world. As a result, much of the policy conversation and some of the policy problems are the same (White, Reference White, Marmor and De Jong1998). The technical capacity aspects of clinical governance are the same everywhere, and that explains much of the common weak implementation. Institutional and political capacity, however, does vary across systems, and that causes variation on other policy dimensions. The most glaring cross-national difference is the absence of any national health insurance guarantee in the US, as a result of a difference in political will (as operationalised by the system's decision rules). But there are also differences in clinical governance, and they may be useful for thinking about the topics of this issue.

From a comparative health policy perspective, the US is unusual not just in its overall results, but also in the variety of its institutions. The US system is a little bit of the UK in its Veterans Health Administration system, a little bit of Canada in its Medicare system for the elderly, and a series of other options in the small portion of integrated group practices such as Kaiser–Permanente and the conglomeration of other insurance arrangements. Because the US is so large, proportionally small aspects of its system, such as the Kaiser–Permanente system or Veterans Health Administration (VHA), may be larger than other nations' health care systems; for instance, each is significantly larger than the Danish system.

In the terms of this special issue, the US has far more market-based governance of medical care than in the other systems discussed here. Hospitals are managed, largely, by entrepreneurs who compete with other hospitals for patients and philanthropic revenue, and must rely on financial markets and retained earnings for much of their capital investment. This means financial market analysts set some premises about how hospitals should be managed (White, Reference White2007).

An even more significant difference involves the role of selective contracting. Although government fee setting shapes much of clinicians' income, a large part is determined through contract with individual insurance companies. In the European systems described here, virtually any provider in a geographically defined area has the right to serve all patients in that area (though there may be an intermediate step of rostering with general practitioners (GPs), and then gate keeping by GPs that shapes access to specialists). The source of a patient's insurance does not matter. In the US, however, the terms by which patients can see specific care givers, and vice versa, are determined by the contracting between insurers and both the insured and the providers. There are thousands of different contracts, each with different terms as to which providers the patients can see, or which they can see with which levels of cost sharing, and which rules about referrals.

Within these contracts, insurers might try to manage treatments in one of three basic ways: third-party utilisation review, risk-bearing gate keeping, or integrated group practice (White, Reference White1999). Third-party utilisation review includes measures in which a bureaucracy controlled by the insurer must pre-approve procedures, such as hospitalisations or the number of days in the hospital. In this situation, the providers will normally refuse to provide the service if the insurer says it will not pay the bill. Risk-bearing gate keeping occurs when a provider, normally a primary care physician, can see its income reduced if it lets too many expensive procedures through the referral gate. For example, a capitated fee could be reduced if the average prescription costs per patient exceeded some budget; or bonuses would be awarded if hospitalization costs for the rostered population were kept below some limit. Integrated group practice is essentially the Kaiser–Permanente model, in which physicians operate both within a system of administrative oversight and a practice culture that encourages less intensive treatment styles. In such a group, for example, paediatricians who share responsibility for call duty in the hospital, will notice if one of their colleagues is unusually likely to hospitalise patients, and will exert social pressure to bring that colleague into conformity with the norms. In the US, selective contracting – contracting only with some providers – is central to implementation of any of these controls because the threat of being excluded from the contract is used to convince providers to accept the restrictions of the bureaucratic review, threats to their income from the risk-bearing, or overall structure of the integrated group practice.

For a short period in the mid-1990s, this system led to highly publicised implementation of managed care within American insurance contracts. At the same time, cost increases slowed dramatically, which led to the common belief that managing care was saving money. This experience, in turn, may have encouraged the belief in Europe that clinical governance was a promising way to control costs, without as much pain as in the usual methods of cost control.

In practice, this story of success was misleading. Implementation of management methods was less thorough than the spread of managed care organisations would suggest. In fact, most of these contracts took the form of preferred provider organisations (PPOs), in which there was very little management of treatments. The number of enrolees in the most managed form of insurance, the traditional group/staff health maintenance organization (HMO), declined significantly after 1993. Savings were obtained mainly by negotiating lower prices during a period when providers for organisational and psychological reasons were weaker, and costs burgeoned again when providers got the upper hand in the contracting. Such measures of treatment management as were implemented, mainly methods of third-party review and putting risk on providers, were rolled back, in part due to provider resistance, in part because of bad publicity, and in part because insurance company managers decided the costs of administering the measures often exceeded the savings from reduced services (see White, Reference White2007, for an extensive discussion).

Nevertheless, the American experience showed that selective contracting has one huge advantage for implementation of management methods. Insurers need not convince all clinicians to accept some set of rules; they only need to recruit enough providers to fill out a network and so satisfy their enrolees. When both patients and providers have the option of ‘exit’, it is easier to resist political pressures that oppose a set of rules.

Failures to create a stable regime of treatment management also reflected some disadvantages of the selective contracting approach. The biggest problem was that the vast majority of providers operated under many different contracts, which made it even harder than usual to accept fact premises within the contracts. When care givers face different guidelines from different insurers, then physicians who accept any given guideline ‘facts’ must reject all the others – and will have good reason to suspect they are all false. Moreover, utilisation review was enforced largely by nurses and clerks rather than by other doctors. Profiling data was suspect because it was not based on a provider's entire practice, thus creating a ‘small number’ problem, with data not only being unrepresentative but, for many conditions, involving too few cases for plausible assessment (White, Reference White1999; Reschovsky and Hadley, Reference Reschovsky and Hadley2007). In short, all the problems of authority associated with clinical governance in the cases in this issue were magnified in the context of selective contracting, as long as physicians and hospitals were contracting with a wide variety of insurers. Not only were physicians unwilling to accept that authority, but they had exceedingly good reasons to resist.

Moreover, the disadvantages of having a system of selective contracting are also severe. It creates expensive administrative overhead from the contracting: the well-known phenomenon of much larger administrative expenses on both the insurer and provider sides of the American system. It disperses payer power over prices, so weakens that form of cost control, as is shown by the much higher prices per service in the US than in other fee-for-service systems, and even the higher prices paid by US private insurers compared to America's quite large governmental payers. In addition, it provides no platform for capacity regulation, since no single payer or consortium of payers can prevent investment by refusing in advance to pay for it. Therefore, any advantages for the micro regulation of medical practice are overwhelmed by the disadvantages of such market-based governance for broader regulation of the health care system.

Two exceptions to the normal US pattern, however, are instructive: the remaining integrated group/staff HMOs and the VHA system. In the ideal HMO, physicians only serve that HMO's patients, who only (except in special cases) go to those physicians. Thus, the HMO can profile accurately. Its guidelines can be developed through its own processes (especially when a system has millions of enrolees, as Kaiser does), and so gain legitimacy from the fact that colleagues of the physicians on whom guidelines are enforced, developed the guidelines. More importantly, attitudes towards standards of care can be developed and enforced by members of the practice observing each other in the practice setting, and the management has far more opportunities to monitor and sanction the physician staff. Most importantly, the physicians who go to work for an organisation like Kaiser are self-selected; those who are not willing to accept its restrictions choose another job.

For all these reasons, management of treatments is much more likely to be implemented within a group/staff HMO. Unfortunately for the vast corps of health policy analysts who wish the US health care system would turn into a system of competing Kaisers, however, these institutions are extremely difficult to build, have major marketing disadvantages, and require a great deal of capital, so are inferior investments. Market-based governance allowed a niche for Kaiser, but apparently no more (again, see White, Reference White2007, for a much more extensive discussion).

The second example, the VHA, is not market-based at all, but it has two key similarities to Kaiser. First, physicians work within a highly organised and even bureaucratised system, often in an employment relationship, with an extensive management apparatus. Second, it is very large, which gives the VHA the resources to perform virtually any imaginable management or analytic task. As a result, it shifted from being widely perceived as a low-quality system to being a leader in adopting improvements in process quality, including extensive measurement of the organisation's activity and dissemination of information to its staff. As Adam Oliver summarises, ‘the improvement in VHA performance is universally attributed to reforms introduced in 1995 that could perhaps more readily be implemented in an integrated, centrally-controlled system than in the ordinary fee-for-service US health sector’ (2007: 6).

In short, experience within the more market-oriented US system suggests that large hierarchies, such as Kaiser and the VHA, are the preferred platform for medical governance. They simply provide a better basis for exercising authority over clinicians. Yet, it remains the case that these systems are still a small part of the much larger US system, and selection of both providers and patients into these systems may mean that the political capacity to operate them could not be extended much further. Moreover, the fact that they can be platforms for management does not mean that in any particular case they will be used either effectively, efficiently, or equitably for that purpose – as critics of the traditional NHS will surely argue.

Conclusion

What, then, is the future of clinical governance measures? We can safely predict that the efforts to improve clinical governance will continue across the world. The criticisms of clinicians are now institutionalised in the profession of health services research. Efforts to measure performance are certain to continue and, given the complexity of medicine, failures are sure to be common. Medicine is also likely to become more and more organised as the expansion of knowledge makes solo practice even less viable. Increased organisation will provide further opportunities to measure performance and create authority.

Over many years, changes in medical education may also make physicians more likely to accept forms of management such as computer-assisted guidelines. In some cases, they will be viewed as helpful (e.g., as ways to avoid prescription errors). Some measures that were not within the zone of acceptance for traditional solo practitioners will be taken for granted by younger physicians who practice in more organised settings.

Hence, systems of authority are likely to evolve in ways that allow somewhat more medical governance. The prospects for any particular measure, however, will depend on its specific characteristics, because the capacity and authority challenges vary. The challenges will always be severe, as the difficulty of creating structures for disease management shows. Thus, it is highly unlikely that governing health care through governing medical practice will ever be less difficult than measures such as governing prices and capacity.