Financial incentives, healthcare providers and quality improvements

Within healthcare there has been a longstanding interest in how the type and amount of payment to healthcare organisations and practitioners affects the type and amount of services received by consumers and, ultimately, the costs of healthcare to individuals, employers, insurers and governments. There has been an equally longstanding interest in how the cost of health services to consumers affects the services they seek out and use, along with the implications of this for overall costs at various levels of aggregation. But there has been much less attention devoted by researchers to the impact of financial incentives on the quality of care. The Institute of Medicine in the USA has defined quality of care as ‘the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge’(p. 21).

This study reviews the healthcare literature that examines the effect of financial incentives on the behaviour of healthcare organisations and individuals with respect to the quality of care they deliver to consumers. Its purpose is to provide guidance to policy-makers in government and decision-makers in the private sector in their efforts to improve quality of care through payment reforms. In this review and synthesis we assess the quality of the evidence relating to the relationship between financial incentives for providers and quality improvement. Specifically, we address:

1. How effective are efforts to reward providers for improving the quality of care they provide or achieving benchmark levels of quality?
2. Whether, and to what degree, financial incentives intended to restrain costs and utilisation have had secondary effects on quality of care?

We also review studies where the main focus was on analysis of utilisation and costs, and find clear evidence that financial incentives can influence these two outcomes. Although this is outside the scope of our ‘quality-enhancing’ brief, we include these studies in an appendix in order to provide a sense of how the literature on the impact of financial incentives on quality has evolved over time, and because many countries have a growing interest in modifying use of health services as costs continue to escalate.

Financial incentives directed at improving quality

We reviewed 36 published articles that examined the impacts of financial incentives directed at improving the quality of care delivered by institutional providers and by healthcare practitioners, particularly physicians. The findings from this review can be summarised as follows:

• The findings from studies on the effect of payer initiatives that reward providers for quality improvements or the attainment of quality benchmarks are mixed. Relatively few significant impacts are reported, and it is often the case that payer programmes include quality improvement components in addition to incentive payments, making it difficult to assess the independent effect of the financial incentives.
• Very little research has been done on the impact of direct payments to hospitals to improve quality. The published research to date in this area is too limited to draw conclusions with confidence.
• Though relatively more attention has been paid to preventive services, there is limited evidence that targeted interventions employing financial incentives to improve the delivery of preventive services are effective. The few studies in this area with strong research designs find small, if any, effects of payments to providers that are intended to improve quality.
• The accumulated body of research described in this chapter is not yet sufficient to assess the relative significance of identified barriers to the effective design and implementation of payfor- performance initiatives.

There are large pay-for-performance programmes underway in the US and the UK with more evaluations likely to appear in the peer-reviewed literature in the near future. Because of the variation in the way these programmes have been designed and implemented, synthesising their findings to provide useful guidance for decision-makers will be challenging. It will be especially important to have comprehensive reporting of results in future studies (not limiting results to a subset of quality measures rewarded by payers), accompanied by complete descriptions of study context and possible confounding factors. In the meantime, policy-makers can support, and learn from, process evaluations of ongoing pay-forperformance efforts with particular attention to accurate documentation of costs as well as continued tracking of outcomes.

Secondary impacts on quality of financial incentives directed at reducing utilisation and costs

We reviewed 45 published articles that addressed, in some manner, financial incentives and their secondary impact on quality of care. We drew the following conclusions from that review:

• The evidence regarding the secondary impacts of financial incentives on quality of care is not compelling. There are several possible explanations. First, the incentives studied were designed, for the most part, to reduce utilisation of services. Generally, the hypothetical link between service reduction and quality in the studies is not clear, especially where utilisation may have been excessive prior to the introduction of different payment arrangements. Second, the literature reports results for a wide range of quality and outcome measures, making it difficult to detect patterns in the findings. The most commonly used outcome measure – mortality – may not be sensitive to the relatively modest changes in financial incentives found in many studies. Also, mortality can be influenced by a host of factors, many unrelated to medical care, making it difficult to isolate the marginal effects of financial incentives.
• How incentives are transmitted to the level at which decisions about treatment are actually made is not clear in most studies. Typically, information is lacking concerning other efforts to address quality via the health plan, hospital, physician practice or government agencies. It seems likely that these efforts would interact with financial incentives for providers to influence quality of care. Most studies do not control for these quality management efforts when drawing conclusions about the impact of financial incentives.
• The use of multiple quality of care and patient outcome indicators in a single research study enables a richer interpretation of findings. However, when results are conflicting in these situations, no clear overall picture of the impact of incentives emerges. Also, it is not clear in most of these studies if the authors adjusted their statistical tests to account for the multiple comparisons undertaken in their analyses.
• The exact nature of provider payment arrangements often is not clearly described in the studies. This is true in particular for comparisons of quality of care under different insurance arrangements. Because the relationships between payment arrangements and quality are likely to be more subtle than the links between payment and service utilisation, the absence of a description of provider payment incentives makes interpretation of findings even more difficult.
• Many of the studies were cross-sectional in design. There may have been considerable variability in provider quality of care, irrespective of financial incentives, that made it difficult for researchers to detect the influence of financial incentives on quality without access to adequate control variables.

Observations

The literature on the influence of financial incentives for healthcare providers on quality of care is under-developed, but that situation seems to be changing at a relatively rapid pace. The science of measuring quality in the healthcare arena is advancing at the same time that purchasers and funders are intensifying their efforts to measure and reward quality improvement. This is likely to generate a significant amount of new research on the topic in the near future. In addition to documenting the relationships between financial incentives and adherence to best practices or changes in patient outcomes, this research should also:

• contribute to a better understanding of the linkages between financial rewards for quality and practitioner behaviour
• assess the cost-effectiveness of pay-for-performance initiatives of different types
• thoroughly document unintended and unexpected effects of pay-for-performance on the healthcare system as a whole.

Definitions and methods

The use of financial incentives to influence behaviour is common in all areas of commerce. The design and impact of incentives has been examined in the research literature at many different levels:

• in theoretical economics, the principal–agent relationship has been studied with the goal of specifying optimal financial incentives in contracts under different assumptions (see Pratt and Zeckhauser, 1985; Eisenhardt, 1989; Sappington, 1991; Milgrom and Roberts, 1992; Prendergast, 1999)
• literature on employee compensation examines use of different payment approaches to encourage desired behaviours on the part of workers (for example, Gerhart and Rynes, 2003; Rynes, Gerhart and Parks, 2005; Sliwka, 2007)
• marketing literature addresses consumer responses to targeted incentive programmes. Further, an empirical, applied literature on the impact of financial incentives on behaviour can be found pertaining to each sector of the economy, including healthcare, and related conceptual issues are addressed in several academic disciplines, such as psychology and decision-making, in addition to economics (see, for instance, Rynes, Gerhart and Parks, 2005).

Interest in the impact of financial incentives on provider behaviour has, until the last decade, focused on an imperative in publicly funded systems to improve efficiency and, in market-based systems, a desire on the part of purchasers, private and public, to moderate the growth in healthcare costs. It is only recently that attention has broadened to include the specific relationship between financial incentives aimed at providers and quality of care. This new focus has come about, to a large degree, because of a growing body of research and suggests that:

• there is wide variation in the services delivered to patients, unexplained by clinical variables, in response to specific diagnoses (see Wennberg and Gittelsohn, 1973)
• the evidence basis for determining which services to provide in treating many clinical conditions remains weak (see Eddy, 2005)
• for decisions where there is clear evidence that a particular treatment approach embodies superior quality of care, patients frequently do not receive this course of treatment, whether for preventive, acute or chronic cases (see Schoen et al, 2005; Schoen et al, 2004; Hussey et al, 2004; McGlynn et al, 2003).

In reviewing the literature relating to our first question (see introduction above) a significant challenge is establishing an up-to-date picture of findings. Study results are appearing in a variety of academic journals at an accelerating pace, reflecting the increasing number of interventions and pilot programs being launched by purchasers that reward providers for quality improvement or the achievement of quality benchmarks. To address these challenges, we have taken a relatively inclusive approach when searching the literature.

A challenge in reviewing the literature relating to our second question (see introduction above) is identifying the relevant research. This is because the impact of financial incentives frequently is addressed as a sub-question – sometimes an afterthought – in the context of studies designed to address the impact of incentives on other outcomes, especially service utilisation and costs.

Electronic searches were performed by the Centre for Reviews and Dissemination (CRD) at the University of York, with supplementary searches undertaken by the research team. We conducted electronic searches of MEDLINE®, EMBASE, Cochrane Database of Systematic Reviews (CDSR), Database of Reviews of Effects (DARE) and EconLit. We also drew on work from the Agency for Healthcare Research and Quality (AHRQ), Organisation for Economic Co-operation and Development (OECD) and World Health Organization (WHO). Search terms are shown in Appendices 2 and 3.

We aimed to bring together the best available evidence on incentives as used in healthcare contexts. In line with the search strategy adopted across the Health Foundation’s quality enhancing inititiatives (QEI) series, we used a ‘best evidence’ approach to conduct our review, initially searching for high quality systematic reviews. If reviews were sparse or out of date, we then searched for individual studies within the following hierarchy of evidence:

• randomised controlled trial
• quasi-experimental study
  
• controlled observational study (for example, cohort or case-control study)
  
• observational study without control group (for example, cross-sectional study, before-and after study, or case series).

Broad inclusion criteria were adopted because of the methodological challenges inherent in assessing the impact of incentives on outcomes and processes of healthcare. The search strategy comprised two phases: the first focused on retrieving systematic reviews, using a wide range of search terms (see Appendix 2). The second used a more limited number of search terms but included a wider range of research designs (see Appendix 3).

Article titles and abstracts for 1564 articles were reviewed by two report authors during an initial search, which covered the period from January 1988 to August 2006. We updated our search results in June 2007, adding newly published articles as well as those that came to our attention through further scans of reference lists in related articles. There is a substantial research literature prior to our starting date, particularly with respect to the impact of financial incentives on utilisation and costs. We rely on the results of published research syntheses to characterise the findings from this early literature (see Appendix 1).

We selected articles that were empirical and that focused on the effect of financial incentives on some aspect of provider behaviour. At this stage, we did not require that the studies address impacts on quality specifically. However, we did eliminate studies from consideration that focused on providers that delivered only mental health and substance abuse services. There are a number of strong studies in this area, but we chose to focus specifically on providers of medical services. We also eliminated studies of the impact of financial incentives on provider location decisions, choosing to focus only on how incentives affect treatment decisions more directly. Again, there is a substantial literature in this area, largely consisting of evaluations of government programmes designed to reward physicians for locating in under-served areas.

We retrieved and reviewed the full articles in every case, writing an article summary using a standard abstract format. The majority of the articles that we retrieved through our search process, and therefore included in this review, analysed payment reforms and initiatives set in the US. This no doubt reflects the multiplicity of payers in the US, creating more opportunities to evaluate different payment initiatives, as well as the relatively large number of academics conducting research on health services there. With respect to studies involving statistical analyses, there was wide variation in the sophistication of the analytic approaches used. Not surprisingly, studies where large datasets were employed, and where findings were published in academic research journals, used the most sophisticated statistical techniques. In contrast, those published in journals aimed more at practitioners or policy-makers and that sought to offer ‘early findings’ regarding specific interventions, were more likely to present simple comparisons of mean values, sometimes without statistical tests for significance of observed differences.

 Read Martin Roland's commentary on the report >