Open Conference Systems, 50th Scientific meeting of the Italian Statistical Society

Font Size: 
Data mining and analysis of comorbidity networks from practitioner prescriptions
Giancarlo Ragozini

Last modified: 2018-05-16


For the present paper, the administrative databases of general practitioners were mined for healthcare systems analysis. A sample of 14,958 patients along with their 1,728,736 prescriptions were considered over a time span of eleven years. From this database, we derived a set of comorbidity networks by considering pathologies as nodes and the co-occurrences of two pathologies to the same prescription as links. The aim is to mine the complexity of this information by using network analysis techniques. Specifically, Islands algorithm method is well-suited to extract the most relevant and connected parts of large and dense networks, as in the present case. Main comorbidity patterns are discussed, along with future lines of research.

Full Text: PDF