Editorial - (2014) Volume 22, Issue 2
Community and Health Research Unit (CaHRU), University of Lincoln, UK
Pay-for-performance (P4P) is a popular means of incentivising behaviour change. As a result, P4P or contingent rewards are ubiquitously applied by people in a variety of settings, from parents rewarding their children for examination success, to employee remuneration packages in the workplace, to payments for healthcare or other organisations achieving quality targets. The issue of P4P in healthcare is hotly and widely debated.[1,2] Recent and increasing evidence from behavioural economics and from research on the effect of contingent rewards on healthcare organisations has led to a wider and deeper exploration of the ethical considerations of P4P.
The basic rationale for P4P from behavioural psychology is that rewarding performance with contingent incentives will heighten motivation and increase effort, leading to better performance. There has been a widespread assumption that the underlying reasoning is true. This leads us to ask a number of questions including the following:
. What sort of performance do/should we pay for?
. How should we define and measure performance?
Whom, how and how much should we pay for better performance?
Performance in health services is measured in a variety of ways. This includes the volume of care provided, including items-of-service such as influenza vaccinations administered by practices each winter, or socalled payment by results (PBR) applied to hospitals in England for the number of patients seen and the types of procedures provided. Capitation payments relate to numbers of patients cared for by general practices and to performance based on the idea that high-quality practices will attract greater numbers of patients.Most current P4P systems relate performance to quality of care, usually measured by one ormore quality indicators, for example, the Quality and Outcomes Framework (QOF) for UK general practice[3] or schemes such as Commissioning for Quality and Innovation (CQUIN), which provide payments to UK health providers for improving quality of care. Many policymakers agree that health services should be paid for by a combination of capitation, work volume and quality, although recently there has been an increasing emphasis on quality and a greater proportion of pay linked to this aspect of performance.
Fortunately, quality is better understood and most definitions have now converged to three main areas: effectiveness (and efficiency), safety and patient experience. [4] Patient experience incorporates notions of patient-centredness, timeliness, accessibility and equity.[5] What are less clear is how payments should be determined and what aspect of quality should be paid for.
The dimensions of quality may be translated into performance measures which may include structure, process or outcome measures.[6] Structural measures might focus on the quality of premises, equipment, staff (and training), protocols and disease registers. Outcome measures include survival (or mortality) rates, complication rates, adverse events and treatment success (or failure), as well as patient-reported outcomes or patient-related experience measures, including satisfaction. Process measures include indicators reflecting adherence to evidence, guidance or consensus on good practice. Payments in P4P schemes are often linked to achievements against targets, sometimes to performance compared with peers or similar organisations, and less commonly to improvements in quality.
Finally there is a question of who should be paid in P4P schemes. Should payments go directly to individuals such as the general practitioner partner, or should they go to the wider organisation, for example the practice? How should incentives be paid? Should this be as a direct financial reward or as Funding to increase the availability of staff or improve facilities? Finally, there is the thorny question of how much incentives should be worth and what proportion of pay should they constitute.
The research evidence from behavioural economics suggests that financial incentives almost always lead to worse performance. How does this paradox of greater motivation and effort leading to worse performance arise? It turns out, at least when it comes to individuals, that this is due to a number of mechanisms which are familiarly described as ‘choking under pressure’. This is reflected in the well-known Yerkes–Dodson law in which performance increases with anxiety until a threshold is reached above which performance deteriorates. By consciously thinking about a task automatic behaviours in experts can deteriorate when they come under conscious control or ‘overthink’ a problem.[7]
Incentives are also known to narrow an individuals’ focus of attention, which can be detrimental for activities that involve insight or creativity. Finally, monetary incentives can divert people to thinking about the reward they might gain and/or the fear of failing and losing the reward, distracting them from the task. Choking under pressure can be exacerbated by the effect of an audience, competition or personal traits such as competitiveness. A series of experiments conducted in the USA and India showed that very high incentives, in fact, lead to worse performance than low to moderate financial incentives.[7]
From an ethical perspective, P4P can be viewed in various ways including utilitarianism, deontology or principlism. Utilitarianism is a system of ethics that seeks to maximise happiness or utility for the greatest number. Deontology is based on the idea of our duty to others, the operation of reason and the idea that truths should be universal. Principlism argues that an ethical approach should try to maximise the four ethical principles of beneficence (doing good), nonmaleficence (not doing harm), autonomy (an individual’s right to self-determination) and distributive justice, that is, the equitable and fair allocation of resources.
Proponents of P4P argue that payment will inevitably lead to better performance. However, the evidence for this is mixed. A recent systematic review of the effect of P4P on quality found that most (13 of 17) examined process measures, often for preventive care, saw partial or positive effects; however, four studies also found unintended effects of incentives.[8] Another review of theUKQOFfor general practice found slight improvements in process measures for diabetes and asthma above secular trends, but no definite improvement in care for coronary heart disease (CHD). There has been no real effect on outcomes apart from epilepsy where there was a significant increase in the recording of annual reviews and this was associated with a reduction in hospital admission.[9]
Increased quality might at first sight be seen to be a utilitarian approach until we consider different aspects of quality and the unintended consequences of some P4P initiatives. Consequentialism requires the analysis of intended and unintended consequences and their effects on utility or happiness. Possible harmful consequences of P4P include medicalisation and a biomedical rather than holistic approach, leading to greater use of excessive or unnecessary treatments.
For example, a focus on individual indicators related to the control of hypertension or hyperlipidaemia can lead to excessive drug treatment and greater potential for adverse effects, particularly in the frail, elderly or in those with comorbidities. Sometimes increasing drug treatment to achieve biochemical targets can lead to worse outcomes such as the increased mortality associated with low glucose levels in diabetes. The focus on a narrow range of indicators may lead to worsening in non-incentivised areas, the ‘lamp post’ effect, rather than ‘spillover’ effects leading to an improvement in other areas, For example, overall, there has been better recording for QOF domains, but not for untargeted areas.[9]
P4P with its basis in reason might fit with a deontology approach, but contingent incentives seem to fundamentally conflict with the idea of doing something because of duty to others. It is also at odds with the ‘categorical imperative’, which states that principles should be able to be applied in all circumstances, also termed universalism: P4P undermines quality as an end in itself when the universal law that derives from P4P is that everything should be paid for. Respect for the person rather than the end outcome can become lost in P4P systems.
When we look to the framework of principlism, the evidence for the benefits of P4P on quality and efficiency is there, albeit limited.[8,9] Doctors and nurses believe that the person-centeredness of consultations and continuity were negatively affected by the QOF and indeed patients’ satisfaction with continuity has declined, with little change in other domains of patient experience.[9] Autonomy can be adversely affected if patients are unaware that their doctor is incentivised to treat certain conditions, encourage specific interventions or undertake particular tasks, and this can undermine trust.[10]
From an equity perspective, early evidence of the QOF suggests that it has helped to narrow the gap in quality indicators in areas of socio-economic deprivation compared with non-deprived areas. There have also been reductions in disparities for older people with cardiovascular disease or diabetes, but some differences remain, for example, worse care for women with CHD. [9]
Despite evidence that P4P may have had benefits, the ethical issues raised by contingent rewards in healthcare mean that we need to build in safeguards to reduce potential adverse consequences, involve stakeholders in developing valid measures of quality, reduce the proportion of pay allocated to contingent rewards, and develop patient-centred measures which are cautiously introduced and carefully evaluated.[11,12]
This editorial is based on a workshop entitled ‘Payments by performance: the good the bad and the ugly’ given at the 8th Annual Institute of Medical Ethics (IME) Conference in London on 7 February 2014. I am grateful to the IME chair and organisers and the workshop attendees.
Commissioned; not externally Peer Reviewed.
None declared.
