Doctor Behaviour Under a New Pay for Performance Contract

Abstract

The UK National Health Service has introduced an extensive system of payment for performance against 146 quality indicators for its general practitioners. GPs can increase their income by meeting a clinical indicator for more patients, or by increasing the number of patients they exception report as ineligible for an indicator, or by varying the number they report as having a relevant condition. We develop a model of GP behaviour and derive tests for gaming. Using data from 916 Scottish practices in 2004/5 and 2005/6, we find evidence of gaming of prevalence. Prevalence reports are also associated with practice characteristics, such as whether the practice was fundholder, suggesting that the QOF prevalence reports may not be a reliable epidemiological resource. We also find strong evidence of gaming of exception reporting in 2005/6 by practices for indicators in which they failed to achieve the maximum reward in 2004/5.

Keywords: Quality, Incentives, Gaming, Pay for performance

JEL Classification: H51, I18