Multivariate data analysis techniques for healthcare organizational efficiency improvement

The main way to reduce costs ensuring good healthcare standards and improving the benefit-cost ratio, in Italy as well as in other countries, is connected to organizational choices: by example, the organizational pertinence in hospitalization typology (“ordinary admission” vs “day hospital/day surgery”). This paper aims to investigate such type of efficiency in healthcare facilities by using multivariate methods of data mining, precisely logit regression, segmentation analysis, and neural networks, in order to assess the organizational appropriateness, evaluating the incidence of the day hospital and day surgery procedures and analysing their relevance in the health system, as well as their pertinence level. Starting by a set of hospital administrative data (deriving from Hospital Discharge Datasheet provided by all Apulian healthcare facilities), this study provides interesting results about the decisional mechanism of the Healthcare management, as well as the ranking of organizational efficiency in the health Apulian network. Further analyses could clarify how (and how much) these results can be extended to other territorial systems.