Fig. 4

Graph representation of comorbidities used for clustering. In the graph a node represents a disease and its size is proportional to the node degree. An edge represents a co-occurrence of two diseases and its width is proportional to the prevalence of the co-occurrence in the dataset. Using this visualisation, it is possible to obtain extra insights into the relationship between comorbidities, as it is possible to observe which diseases tend to co-occur more frequently. The thickness of the edges makes it possible to verify that HT and Anaemia, and HT and AFIB occur frequently together. CKD and ICM also show a high co-occurrence with HT and Anaemia. ICM–Ischaemic Cardiomyopathy, HT–Hypertension, AFIB–Atrial Fibrillation, CVD–Cerebrovascular Disease, VD-Valvular Disease, CKD–Chronic Kidney Disease, COPD–Chronic Obstructive Pulmonary Disease, CM–Other Cardiomyopathies