TY - JOUR T1 - Measuring performance quality in general practice: is international harmonization desirable? JF - British Journal of General Practice JO - Br J Gen Pract SP - 691 LP - 692 VL - 57 IS - 542 AU - Richard Grol AU - Michel Wensing Y1 - 2007/09/01 UR - http://bjgp.org/content/57/542/691.abstract N2 - The call for transparency about quality of care provided to patients has become stronger in recent years in most developed countries. Society demands that care providers, GPs included, account for their performance, and is increasingly prepared to pay for better care. The development in accountability faces critical debate, such as questions about its impact on improving practice, and claims that it is a threat to professionalism, that it promotes strategic behaviour and even fraud, and increases bureaucracy and costs of health care.1–3As public reporting of performance and pay for performance are becoming more of a reality in health care, it is important to observe and evaluate these new ‘accountability approaches’.Seen from the other side of the North Sea, the pay-for-performance initiative in general practice in the UK (Quality and Outcomes Framework [QOF] and GMS contract) is one of the most interesting quality improvement experiments for GPs in the world today. We in the rest of Europe learn from this experiment. However, we sometimes do things differently, and also ask ourselves why the UK does not take on board some of the relevant experiences gained within Europe on measuring performance in general practice.One of the crucial issues related to any performance assessment, whatever its objectives, is the validity and reliability of the indicators and measurement instruments used. In particular, when indicators are used for comparison, incentives, or certification, they should be ‘correct and fair’ and avoid unjustified harm to practitioners.4 Indicators should meet the highest standards for quality, their features should be tested, the scores should be corrected for case-mix differences. Those assessed should have the possibility to check data before use by … ER -