Financial incentives are increasingly adopted to improve allocative efficiency and quality in primary care. Although it has been recognised that incentive-based remuneration schemes can have an impact on GP behaviour, there is still weak empirical evidence on the extent to which such programs influence healthcare outcomes and on the degree of physicians' responsiveness to their introduction. This problem reflects the lack of adequate empirical data but also the complexity of general practice systems where many confounding and institutional factors are likely to influence physician behaviour. Given this background, we investigate the impact on quality of care of the introduction of pay for participation incentives in primary care contracts in the Italian region Emilia Romagna. We concentrate on patients affected by diabetes mellitus type 2, for which the assumption of responsibility and the adoption of clinical guidelines are specifically rewarded. We test the hypothesis that, other things equal, patients under the responsibility of GPs receiving a higher share of their income through these programs are less likely to experience hospitalisation for hyperglycaemic emergencies. To this end, we examined the combined influence of physician, organisational and patient factors through the use of multilevel modelling. Data were obtained form a large dataset made available by the Regional Agency for Health Care Services of Emilia Romagna. This dataset covers patients and GPs of the whole region and provides detailed information on healthcare consumption of the population, on the different components of GP remunerations, on morbidity levels of large groups of patients. Estimations are obtained for the year 2003.