Project manager: Martin Hassel
Project participants: Hideyuki Tanushi, Hercules Dalianis, Mia Kvist
Funding: Internal, 3 months.
Project period: October - December 2011

Comorbidity is a term that describes how diseases cooccur and are grouped within the population. One way to identify such groups is to find recurring patterns in large clinical data sets, such as patient records. We have previously made an initial study of comorbidity in a large number of patient records from the Stockholm County Council. This study resulted in a scientific article that focuses on pairwise groups of diseases, where we studied major endemic diseases such as diabetes mellitus and chronic obstructive pulmonary disease (COPD) and their sequelae.

Comorbidity-View aims to construct a demonstrator that presents all of the relationships between diseases that are co-morbid. These will be presented as nodes in a network with an interactive graphical interface that enables the exploration of these complex relationships. For this, we will make use of diagnosis codes (ICD-10) that are encoded in the records but also extract diagnoses from the free text notes in the medical records.

Demonstrator

Comorbidity-View: view co-occurring ICD-10 diagnoses as an interactive graph.