Abstract
Due to the increasing pressure on resource limitations and the need for efficiency improvements, effective healthcare service planning should analyse trends in citizens' demand for healthcare services, as well as patients' choices of healthcare providers, which are determined not only by prestige, but also by physical accessibility and availability. The main objective of the present study is to explore the benefits of adopting the social network analysis (SNA) approach to investigate the determinants of a patient's choice of healthcare provider as a way to support better resource allocation decisions in healthcare systems. For the purpose of the analysis, administrative data was used to track the flow of patients within orthopaedic departments of the Lombardy region (Italy) in 2014. A network was developed with two types of nodes: those of orthopaedic departments in the regional hospitals (114 nodes), and those of municipalities (5092 nodes). Using the Louvain algorithm, the communities and sub-communities were determined on the basis of patients' choices, without any prior geographical considerations or planned catchment areas. Traditional SNA measures and other novel indicators, specifically developed for this study, were applied: in particular, attractiveness (i.e. the preference that patients assign to a certain department against the others available in the same reference area) and confinement (of demand) (i.e. the degree of the unique use of a certain department). A decision support matrix based on these indicators was defined for decision makers to use in evaluating department profiles, and optimizing the location of services and allocation of resources, while preserving accessibility to care and patients' preferences. The decision support matrix was also evaluated on the basis of real practice and decisions made by regional healthcare managers.