Do guidelines create uniformity in medical practice?

This article aimed to test the general hypothesis that guidelines create uniformity, or reduce variation, in medical practice. Medical practice variation has policy interest and is one of the reasons for developing guidelines. The development and implementation of guidelines was considered in the broader context of processes of rationalization. We focused on the influence of voluntary guidelines developed by the professional organization for family physicians in the Netherlands on variation in drug prescription. Data were used from the First and Second Dutch National Survey of General Practice (DNSGP1 and DNSGP2), collected in 1987 and 2001 respectively. DNSGP1 consisted of 103 practices and 161 GPs serving 335.000 patients. DNSGP2 consisted of 104 practices and 195 GPs serving 390.000 patients. Two groups of diagnoses were created, one containing all diagnoses for which guidelines were introduced and one containing all other diagnoses. For both groups a measure of concentration, Herfindahl-Hirschman Index (HHI), was used to represent variation. This measure of concentration was compared between both groups using multilevel analysis. Results showed that although there was an overall increase in variation (a significantly lower HHI) in prescription, the increase was less in the cases of diagnoses for which guidelines were introduced. Guidelines, primarily, had an effect on variations in single-handed practices. The overall conclusion is that the introduction of guidelines, although it probably tempered the increase in variation, did not reduce variation.