Barriers to performance management

The technical barriers are daunting. Most approaches to performance management require clear definitions of quality and reliable methods of measurement, but many aspects of medicine defy easy quantification. Although this argument is sometimes overplayed by the professions (such as when appealing to the power of clinical intuition, despite its demonstrable fallibilities; Dawes et al., Reference Dawes, Faust and Meehl1989) it is in many respects justifiable. For example, the ability to raise expectations and to marshal patients’ psychological and physiological defences through effective patient–practitioner interactions is an essential clinical skill (Benedetti, Reference Benedetti2009). Capturing the quality of these interactions across an entire health care system is not currently feasible, and the effectiveness of this and many other aspects of care can only be measured indirectly through the assessment of patient outcomes. Outcomes, however, are influenced by a host of factors beyond the control of providers, making judgements of their abilities and fair comparisons between them problematic. The potential solutions – risk adjustment, larger patient panels and clinical exclusion criteria – introduce new problems of their own, such as overly complex measurement frameworks, loss of specificity and opportunities for providers to game.

The ideological and philosophical barriers can also be formidable. In particular, definitions of high quality may differ radically between providers and purchasers. To date, performance management has concentrated on incentivising safety, effectiveness and timeliness, as these dimensions of care are generally easier to measure than patient centredness and equitability. More recently, efficiency and ‘value’ have become the focus of quality improvement initiatives, particularly in the United States, where health care costs have spiralled out of control. Clinicians, however, may be committed to aspects of care that are of lower priority to purchasers, or that are more difficult to measure. This creates a conflict for clinicians, caught between doing what they believe is in the best interests of patients and doing what is required in order to maximise income or avoid censure. As Oliver points out, involving clinician representatives in constructing incentive frameworks is therefore essential, but limits to the extent of their involvement are required in order to avoid provider capture. Although it is important to have clinical involvement in identifying quality indicators, it is equally important not to have the subjects of incentives setting their own performance targets. Failure to recognise the latter point hobbled the UK’s Quality and Outcomes Framework, as general practitioners’ representatives were able to block proposed increases in achievement targets for several years, during which time quality improvement stalled (Doran et al., Reference Doran, Kontopantelis, Reeves, Sutton and Ryan2014).

Using incentives in health care

The incentives under pay-for-performance programmes are never purely financial, as they are invariably linked to some form of public reporting or ranking. Separating financial from reputational and professional incentives can be difficult, but it is conceivable that for a generally well-remunerated profession marginal financial gains are less of an incentive than maintaining professional standing. Under the UK’s Quality and Outcomes Framework, most practices performed above the level required to secure maximum remuneration, suggesting motivations beyond avarice were at work, although a financial component may have been required to draw practices into the scheme and make them aware of their relative status (Oliver and Brown, Reference Oliver and Brown2011).

Whatever the underlying mechanism, financial incentives are generally effective in improving effort for routine, repetitive activities. However, when greater cognitive ability or cooperation is required, the results of financial incentives can be disastrous: they subdue creativity; nurture narrow, short-term thinking; and inadvertently reward selfishness and all manner of devious behaviour. As health care is ideally a creative, cooperative and continuing endeavour, built on strong ethical foundations, many within the caring professions find financial incentives repugnant. These objectors have new allies in the church of identity economics, whose high priest George Akerlof has criticised pay-for-performance schemes for signalling to providers that they cannot be trusted (Akerlof and Kranton, Reference Akerlof and Kranton2010). The close scrutiny required by incentive schemes risks alienating providers, converting them from committed ‘insiders’ to demotivated ‘outsiders’ who might then seek to game and defraud their way to apparent high performance, as occurred in several offices of the Veterans Health Administration (Chokshi, Reference Chokshi2014). Even honest triers may change the way they record their activities under incentive schemes, with serious consequences for the long-term monitoring of quality of care (Bevan and Hood, Reference Bevan and Hood2006; Farmer et al., Reference Farmer, Black and Bonow2013).

The ongoing confusion about the merits of incentive schemes is partly related to diversity in both design and impact. The same set of incentives can have diametrically opposite effects on the motivations of different providers: some will thrive, others will adopt outsider status, particularly if they perceive the system of rewards and punishments to be irrelevant or unfair. The danger then is that the resulting variations in care will be concealed by gaming and misreporting. Programme design can increase or reduce this risk: for example, competitive frameworks can stimulate providers located towards the top of the rankings (particularly those who just missed out on bonuses in previous rounds; Ryan et al., Reference Ryan, Sutton and Doran2014) whereas fostering disillusionment and discontent in those at the bottom, who may decide that they do not want to play – or at least not play fair – in a game they cannot win. In contrast, criterion-referenced systems, which set targets for attainment or improvement in advance, give all providers an opportunity (albeit not necessarily an equal opportunity) to obtain financial rewards, and remove some of the uncertainty and resentment associated with competitive schemes.

That monster, custom

According to identity economics, high levels of performance can be secured over the long term by harnessing the altruistic and professional motivations of providers and aligning them with the aims of purchasers. Insiders increase their identity utility through exerting effort, and are therefore self-motivating and self-monitoring with no requirement for additional financial inducements. Following this logic, rather than funding financial incentive schemes, purchasers would be better advised to invest in recruiting or creating likeminded insiders. Bruce Vladeck made similar points a decade ago, arguing that values of professional responsibility need to be nurtured in a health service where providers identify with an organisation that has the provision of high-quality care as its primary aim (‘Unless we can continue to assume that most providers and administrators want to do the right thing for most patients most of the time, we are all sunk, and no amount of economic incentives can salvage the situation’) (Vladeck, Reference Vladeck2004).

This nurturing process needs to begin with medical schools selecting entrants on characteristics such as altruism, commitment and compassion, and not just on academic ability. In the United Kingdom, the traditional selection procedure of summing grades and vetting personal statements – which may have been ghosted – admitted many bright but unsuitable candidates to the profession, and deprived the NHS of potentially exceptional doctors. Selection procedures have increased in sophistication, but it remains to be seen whether, for example, the ‘situational judgement’ component of the UK’s Clinical Aptitude Test (UK Clinical Aptitude Test, 2014) can – in under half an hour – successfully identify the adolescents most likely to develop into caring and competent clinicians. Once selected, candidates need to be provided with educational and work environments that do not then obliterate all traces of benevolence, commitment and responsibility (Report of the Mid Staffordshire NHS Foundation Trust Public Inquiry, Reference Francis2013).

Despite the best efforts and most supportive environments, however, the health care system is too large to populate exclusively with saints. There will always be room for rogues and maniacs (Hamblin, Reference Hamblin2007), and there will be many saints who are roguish in their approach to some of their duties. Effective performance management must take account of individuals at all points of the spectrum, whilst guiding the majority towards sainthood. Proponents of incentive schemes argue that they do just that by providing clear guidelines for high-quality care, rewarding compliance and punishing failure. Pay-for-performance schemes can also counter perverse incentives to provide sub-standard care associated with other forms of remuneration, thereby forming a key component of balanced payment models (Roland, Reference Roland2012). Opponents, unimpressed by the irony of stimulating better care for patients by appealing to the more selfish motivations of their physicians, worry that incentive schemes replace beneficence with fear and greed, ultimately degrading the relationship between doctor and patient. This is an enduring dispute, with no immediate end in sight.