Mining the sociome for Health Informatics: Analysis of therapeutic lifestyle adherence of diabetic patients in Twitter

Abstract In recent years, the number of active users in social media has grown exponentially. Despite the thematic diversity of the messages, social media have become an important vehicle to disseminate health information as well as to gather insights about patients’ experiences and emotional intelligence. Therefore, the present work proposes a new methodology of analysis to identify and interpret the behaviour, perceptions and appreciations of patients and close relatives towards a health condition through their social interactions. At the core of this methodology are techniques of natural language processing and machine learning as well as the reconstruction of knowledge graphs, and further graph mining. The case study is the diabetes community, and more specifically, the patients communicating about type 1 diabetes (T1D) and type 2 diabetes (T2D). The results produced in this study show the effectiveness of the proposed method to discover useful and non-trivial knowledge about patient perceptions of disease. Such knowledge may be used in the context of Health Informatics to promote healthy lifestyles in more efficient ways as well as to improve communication with the patients.