Discussion Paper - (2010) Volume 18, Issue 2
2Research Fellow in Health Policy
Health Services Research Unit, London School of Hygiene and Tropical Medicine, London, UKReceived date: 14 January 2010; Accepted date: 16 February 2010
Background Pay for performance (P4P) schemes have become increasingly popular innovations in primary care and have generated questions about their effect on improving quality of care. Aims To provide a brief outline of the international evidence on the relationship between P4P schemes and quality improvement. MethodWe conducted a literature search using relevant databases and reference lists of retrieved articles which discussed P4P schemes, quality in primary care and the Quality and Outcomes Framework (QOF). These included two recent systematic reviews of P4P schemes. Results Evidence on the effect of P4P on quality is limited. What we can say is that P4P schemes can have an effect on the behaviour of physicians and can lead to better clinical management of disease,but that there is cause for concern about the impact on the quality of care. ConclusionP4P schemes need to take more account of broader definitions of quality, as whilst they can have a positive impact on incentivized clinical processes, it is not clear that this translates into improving the experience and outcome of care.
pay for performance, primary care, quality improvement
There is increasing international interest in pay for performance (P4P) in primary care. The introduction of such performance systems reflects concerns about three interlinked issues – the variation in performance and quality, the emphasis on driving improvements in performance and ensuring high quality primary care.[1] This paper briefly examines the current evidence on P4P schemes and the extent to which such schemes can contribute to quality in primary care services. We start by identifying key aspects of quality relevant to primary care and then examine the relationship between P4P schemes and quality criteria
Primary care by its very nature is likely to encompass substantial variation in practice due to the nature of the delivery and organisation of care (for example, different staff mixes and training levels) and also to external contextual factors (socio-demographic factors, geography etc). In the UK, variation in practice standards has been observed over many years.[2] In 2004 the Government renegotiated the general practitioner (GP) contract incorporating a P4P element – the Quality and Outcomes Framework (QOF) – which could account for around 20–30%of practice income. The QOF financially rewards GPs for the quality of care they deliver to patients across four domains: clinical, organisational, patient experience and extra services (such as maternity and child health). It is a practice-based, practitioner scheme with practices awarded points for performance against set indicators within each domain. Other P4P systems for primary care have been developed in a number of countries although, as a recent survey by Schoen and colleagues demonstrates, the degree to which financial incentives are employed to improve the quality of primary care varies, with primary care physicians in theUSA (33%), Sweden (10%) and Norway (35%) less likely to be receiving financial incentives while primary care physicians in countries such as the UK (89%), the Netherlands (81%), New Zealand (80%), Italy (70%) andAustralia (65%) aremore likely to receive financial incentives.[3] The design of P4P schemes varies but essentially they tend to be based on clinical activity targets.
However, performance as a measure of quality depends upon what the performance standards are and how they are measured. This point is particularly important in relation to P4P schemes. Hogg et al (2007) argue that the performance domain in such schemes is divided into two main components: health care delivery and the technical quality of clinical care.[4] This is an important distinction, as most P4P systems focus on aspects of clinical care rather than including delivery systems. Giuffrida et al (1999) also caution that it is important not to confuse performance indicators with health outcomes.[5] Perhaps a key point is to examine the match between performance targets and those criteria generally seen as central to the provision of high-quality primary care.
There have been a number of attempts to define quality in primary care. Wilson et al (2006) suggested that there are four broad areas upon which the performance of primary care, and general practice specifically, should be measured.[6] These are equity, quality of clinical care, responsiveness to patients and efficiency. Their review of practice suggested that UK practices score highly in all four domains although there have been recent concerns about a lack of support for selfcare [7,8] and poor support for people with long-term conditions.[9] The fact that inequalities in health at a primary care level persist6 also raises questions about whether general practice can retain this strong position. However, clinical aspects of the management of chronic conditions do feature in many P4P schemes. For example, the UK QOF includes a number of clinical criteria related to cardiovascular disease, diabetes and chronic obstructive pulmonary disease.
While a clear and uncontested definition of quality primary care is unrealistic there are specific components that are widely agreed as central to the idea of quality. For example, Starfield identified four unique features of a primary care service: first contact access, personfocused care over time, comprehensiveness and coordination. [10] Hogg et al suggest that other important aspects of primary care include patient–provider relationships as defined by communication, holistic care and an awareness of the patient’s family and culture.4 They also argue that primary care performance needs to be set within a broader structural environment that recognises the wider healthcare system, the practice context and the organisation of the practice. This reflects an increasing acceptance of the role of the healthcare delivery system, including issues of governance and accountability, resources and interrelationships between primary care and other health and social care services and person-centred care.[11] Furthermore, in a review of outcome indicators for primary care, Sans-Corrales et al identified key attributes linked to patient satisfaction, health outcomes and cost of services provided.[12] They found that improved satisfaction and health outcomes were associated with continuity of care, patientcentred care, longer appointments and a good patient– doctor personal relationship. These factors were also associated with lower overall health costs. Continuity of care is consistently reported as a key attribute and quality indicator of good primary medical care (general practice/family medicine).[13] These aspects of quality are less likely to be included in P4P schemes, given their clinical focus, despite the evidence highlighting them as key components of primary care performance and quality. Indeed, of further interest is the way P4P schemes are starting to redefine how quality is conceptualised in practice. A number of recent articles reporting on P4P, especially in the UK and the QOF, appear to equate quality with the P4P criteria.[14–17]
Despite some scepticism about the evidence base of the effectiveness of P4P schemes in improving quality,[18] recent systematic reviews[19,20] have concluded that P4P contracts do affect physician behaviour and increase the number of primary care services provided – although often in complex and limited ways.[21] The actual effect depends on factors such as the age and sex of physicians, previous experience of financial incentives, the uptake of continuing professional education, the type of payment method, the type and severity of the conditions targeted through incentives, the volume of activity and the location and type of organisation.[19]
Akey concern that recurs in the literature is whether financial incentives generate dysfunctional physician behaviour[22] or negatively affect motivation,[23] particularly in the light of well-established inverse care patterns at primary care level.[24] Some commentators have argued that there is a risk of neglecting the resources of emotion, morality and trust which are said to be a key part of a physician’s professional repertoire.[25] Research with GPs in the UK revealed that they are also anxious that ‘biomedical’ targets might undermine continuity of care of the ‘whole person’ and might mitigate against developing relationships with patients as treatment is increasingly divided up among a larger team of health practitioners14 as practices seek to set up more efficient disease management systems and GPs offload routine tasks to nursing staff. There is also some evidence that physicians who work within incentive systems that are designed to reduce secondary care referral rates are anxious that their ability to deliver quality care for their patients could be compromised as they experience pressure to reduce referral rates.[21]
A further concern about the impact of externally structured incentives such as financial inducements is that they might ‘crowd out’ professional self-esteem and a sense of self-determination. This might have implications for the quality of care offered by practitioners. However, it has been noted that there is an equal chance of a ‘crowding in’ effect if practitioners feel like they have some ownership of incentives.[26] Indeed, one study in the UK found that the QOF as an externally imposed system of incentives did not appear to damage the internal motivation of GPs.[23] The authors attributed this to the fact that the indicators within the QOF aligned with what GPs themselves considered good clinical care objectives. This could be because performance indicators in the QOF were negotiatedwith representatives of the profession itself, which ensured a degree of alignment of objectives and reduced the potential for decreased internal motivation. 14 Nonetheless, there is some evidence that GPs remain anxious about the impact of external incentives on professional (internal) motivation; although a recent study suggested that GPs feel that whilst professional autonomy has decreased and workload increased, they are paid more, their job satisfaction levels continue to improve and job pressures to decrease under the QOF. GPs also report that they feel the QOF has had a more positive impact on quality of care than they had initially thought it would.[15]
Another potential problem created by external financial incentive schemes is that they could lead to the neglect of those non-incentivised areas of care which will continue to rely on the professionalism or moral motivation of GPs. There is some evidence of concern amongst GPs that non-incentivised areas like acute care, preventive care, care for specific groups such as children or older people and care for patients with multiple comorbidities would suffer as GPs chased targets. Indeed, a recent study found that whilst quality of care for QOF-incentivised conditions improved substantially between 2003 and 2005, there was little or no improvement in non-incentivised quality indicators. [27] However, it has also been argued that this could be positively interpreted as GPs maintaining standards of care in these areas in spite of the lack of incentives and the time required to focus on QOF targets.[28]
Research conducted in the USA has found that the size and structure of incentives seems to be important in incentivising effective physician activity. Incentives have to be large enough to influence behaviour[29] and designed in such a way that they cannot be played off so as to reward both process and improved outcomes.[30] However, the size of incentive has also been found to be less of a factor in the use of care management processes for patients with chronic illnesses by physician organisations (POs) than are schemes that give public recognition for scoring well on quality of care measures, schemes which require POs to provide quality of care or outcomes data to outside organisations or those that reward high-quality scores with better contracts that assist in developing better organised quality provision.[31]
Questions continue to circulate about the likely individual and population health gain from P4P schemes. Evidence of physician activity is not always a measurement of better health outcomes.[20] The evidence of a relationship between incentive payment and likely health gain appears to be weak or mixed.[21] Furthermore, it is difficult to detect patterns from the diverse range of definitions of quality and the outcome measures used by researchers. The most common measure – mortality – may be unreliable because it is affected by wide range of factors and, as with other outcome measures, may be difficult to achieve or may be beyond the control of the physician or provider.[21] Some studies have called for a combination of process and outcome measures when structuring incentives.[30] It is technically challenging to connect performance targets with health gain and most P4P schemes adopt a pragmatic approach and focus on processes (such as measuring blood pressure) and intermediate outcomes (controlled blood pressure) for which there is either evidence or professional consensus and which can be easily measured and rewarded. This means that treatment and secondary prevention is favoured over primary prevention and can lead to the marginalisation of some conditions.
In the UK, the evidence of whether the QOF rewards outputs that are expected to lead to good outcomes is contradictory, demonstrating both that meeting certain QOF indicators might improve health outcomes in some areas16 and a weak causal relationship between key clinical indicators and outcomes.[32] There is some concern that the QOF may lead to an exacerbation of health inequalities by allowing GPs to use the exception reporting system to exclude highrisk patients, or by not sufficiently rewarding the extra work required in delivering equal treatment to disadvantaged populations, maintaining inverse care patterns.[33] Inequities between population groups remain, as it has been found that rates of statin prescribing in practices serving deprived populations is higher but prescribing volume in practices with higher proportions of older people and minority ethnic groups is lower.[34]
In the space of this article it has not been possible to examine all of the rapidly growing literature on P4P. However, we have provided a significant flavour of the current debates. Our intention has been to provide a review of the key themes and raise concerns about the relationship between P4P and how we understand quality in primary care. This brief review of the evidence related to P4P schemes suggests that the use of financial targets is effective in changing the behaviour and activities of practitioners (doctors and others). Such schemes, by setting targets, also have an impact on the range of activities undertaken by practitioners. Generally, the adoption of P4P schemes demonstrates that financial payments are a key incentive for adopting new processes such as blood pressure measurement, cholesterol screening, statin prescribing and the measuring of blood sugar levels and body mass index. However, the implications for organisational aspects and patient care are less clear. There has been criticism of P4P schemes for not adequately addressing health outcomes and aspects of patient perceived quality are generally not included in these schemes. While in the UK, waiting times and access are measured and incorporated in the QOF these are not necessarily priority criteria of quality from a patient perspective. In fact the literature on P4P tends to equate quality of primary care only with clinical processes, despite a substantial literature identifying other aspects of primary care as being important constituents of quality.4–6
As we have noted, incentive payments may skew physician activity towards high-reward labour-intensive activities with relatively low health benefits, thereby marginalising non-incentivised areas. This potential for ‘gaming’may create a conflict of interest for physicians betweenmaximising revenue and ensuring good quality care. Financial incentives may also distort care by encouraging a focus on individual measures for care management instead of a more integrated approach which might be appropriate, particularly in areas of comorbidity. In addition, the use of targets and financial incentives can have unintended consequences on practitioner behaviour, such as goal displacement and rule following, leading to the ‘crowding out’ of and reduction in focus on non-incentivised tasks. Thus areas of clinical activity not included within P4P schemes become less important. Studies have also found that financial reward is not necessarily the main incentive for practitioners to engage in quality improvement and while targets clearly deliver changes in behaviour, they can lead to goal misplacement in which rule following becomes a means to an end. It would seem there is a need for intelligent and vigilant structuring of P4P schemes lest they lead to a narrowing of definitions of quality in primary care and restrict our focus to clinical process at a time when richer meanings of quality should be gaining currency.
We would like to thank the editors for their constructive feedback
