A comparison of psychiatric casemix in the UK and Italy

Abstract

  Background: The objective of this study was to compare casemix groups in Italy and the UK. Three hypotheses were tested. First, that length of stay would be shorter in Italy. Second, that diagnosis would not be an effective predictor of length of stay. Third, that diagnosis-based groups would be better predictors of length of stay in Italy due to shorter in-patient episodes. Methods: In-patient episodes in Pavia (Italy) and Camberwell (UK) for patients with psychosis were allocated to four Healthcare Resource Groups. Average lengths of stay, and coefficients of variation in the two areas were compared. A regression model was constructed to determine the impact of diagnosis groups, gender, age and area on length of stay. Results: Length of stay was significantly shorter in Pavia than in Camberwell for three of the four groups. The coefficient of variation was generally above 1, indicating a lack of within-group similarity. With outliers trimmed, the figures were below 1, except for one group in Camberwell. Pavia revealed greater group homogeneity than Camberwell, with one exception. The regression model explained 9% of variation when the two areas were both included, and around 5% when compared separately. Age had a non-linear impact on length of stay in Camberwell. Conclusions: We have again shown that diagnosis does not adequately predict length of stay. However, in Italy, where admissions are for shorter periods, diagnosis groups are more homogeneous. This may be because the initial part of a hospital stay is due to clinical factors, whereas prolonged stay is influenced by supply-side issues.