An evaluation of the variability of episiotomy rates across hospitals: the case of Sardinia

The aim of this paper is to assess the extent of variation in the use of epi- siotomy across hospitals and to evaluate how much of this variation can be explained by case-mix factors. Using o cial hospital abstracts on deliveries occurred in the hospitals of the Italian region of Sardinia during 2009, we implement a multilevel logistic model in order to predict the likelihood of an episiotomy from a set of covariates which includes both socio-demographic and clinical indices. Results suggest that, with the exception of education, socio-demographic variables were not signi cant in determining episiotomy while several clinical predictors were signi cant. Our main nding is that almost half of the variation in episiotomy rates remains unexplained after conditioning on clinical indicators and socio-demographic factors.