Article Preview
Top1. Introduction
The problem of sub-optimal quality is common to health systems around the world (Fisher & Wennberg, 2003; Hussey et al., 2004; Wennberg, Fisher, Goodman, & Skinner, 2010). The clinical performance spectrum has been characterised as a bell curve, with a few providers delivering markedly sub-standard care, some others achieving markedly superior outcomes, and a large majority clustered around the median performance level (Gawande, 2007). In light of these deficiencies, health systems have pursued an array of quality improvement strategies.
The perceived key determinants of health system performance may vary according to one’s disciplinary perspective. On the one hand, clinicians are instilled with altruistic professional values and generally motivated to deliver high quality care, therefore the most important determinants of provider behaviour are ethics, professionalism, and clinical expertise. Yet tensions may exist between the ethical imperative to deliver high-quality care and the need to generate income, as clinician livelihoods and the viability of provider organisations depend on a steady income stream (Reinhardt, 2012). Accordingly an economic interpretation of care patterns is defensible, and the economic incentives instilled by provider payment mechanisms are recognised as another determinant of performance (Roberts, Hsiao, Berman, & Reich, 2004). Motivated in part by the modest impact of quality improvement initiatives such as public performance reporting (Ryan, 2011), much policy attention now focuses on the use of financial incentives to improve quality.
Pay-for-performance (P4P) is a prominent tactic for quality improvement that modulates provider income in line with measured performance on specified indicators. Ideally, this sharpens incentives for high quality and compensates providers for the cost of investing in quality improvement. P4P can enjoy broad stakeholder support among payers, consumer groups, and political actors (Tanenbaum, 2009). Its popularity has grown, and a 2006 survey found that 126 out of 252 commercial HMOs in the United States (USA) were incorporating P4P into provider payment mechanisms (Rosenthal, Landon, Normand, Frank, & Epstein, 2006), while substantial investment in P4P is directed to low- and middle-income countries through the World Bank (Morgan, 2009).
But despite this burgeoning popularity, the evidence base for P4P’s clinical and cost-effectiveness is scant (Mullen, Frank, & Rosenthal, 2010; Van Herck et al., 2010; Scott et al., 2011). I briefly examine this evidence base before proceeding to the main focus of the paper, an analysis of P4P’s key design features. This literature review draws on four databases: PubMed, Embase, Google Scholar, and the Cochrane Library, until September 2012. Search terms included combinations of “pay for performance”, “financial incentives”, “quality and outcomes framework”, “performance based payment”, “health care”, and quality.
Relatively few P4P programmes have been robustly evaluated, and evaluations often uncovered no evidence of benefits. Some studies have found negative effects on the incentivized performance measures, implying performance inferior to a comparator group rather than deterioration in quality. In addition, valid inferences are often elusive due to the concurrent implementation of other quality improvement strategies (Mullen et al., 2010). A systematic review of P4P’s effects in low- and middle-income settings uncovered only nine studies, including non-randomized trials and interrupted time series studies. Results were mixed and interpretation difficult due to the high risk of bias (Witter, Fretheim, Kessey, & Lindahl, 2012).