Visual analytics in clinical medicine

This chapter describes how visual analytics may be employed in clinical applications. This requires to dig into how patient data is typically represented, which information is available, and how the quality of that information can be assessed and improved. To visualize the patient history at a glance and to enable selective filtering of diagnostic results, treatments and symptoms are essential goals. The patient history is characterized by many different types of data, among which event-typed data are essential, e.g., the begin and end of a treatment or a hospital stay. Therefore the visual exploration of event-type data is discussed intensely. Whereas one complex patient history is already challenging to visualize, it is often useful to analyze a cohort of patients to identify frequent patterns during a treatment process. Thus simplification of patient histories and alignment of patient histories to make them comparable are essential topics. All the available clinical data require interactive visualization to be able to make sense of their underlying complex patterns. Here, two important applications are predictive analytics (answering the question Can we predict a severe course of a disease early?) and clinical decision-making support.