RESUMEN
Childbirth is one of the main causes of hospitalisation for women, accounting for about 5% of hospital activity in most Organisation for Economic Co-operation and Development countries. We analysed the factors that explain variations in resource use for child delivery in ten European countries. We compared the performance of three models for explaining the variations in resource use (log cost or length of inpatient stay) at patient and hospital level. The first model used only the DRGs to which child deliveries were coded (M(D) ), the second used a set of 'patient-level' and delivery specific explanatory variables (M(P) ), and the third model combined both sets of variables (M(F) ). Countries vary both in the number of DRGs and the criteria used to classify cases of child delivery (range: 3-8) and in the percentage of deliveries classified as 'delivery without complication' (range: 53-90%). The capacity of DRGs and patient level variables to explain cost variation for child birth ranges from 48% in Sweden to over 70% in Spain. There is room for improving current DRG classification in most countries, but this does not necessary imply multiplying the groups and/or complicating criteria. Countries with a higher number of DRGs do not always perform better.