Implementation of a virtual patient model for older people and extraction of clinical profiles through clustering analysis

• Κύριος συγγραφέας: Καλογιάννης, Σπυρίδων
• Άλλοι συγγραφείς: Μεγαλοοικονόμου, Βασίλειος
• Μορφή: Thesis
• Γλώσσα: English
• Έκδοση: 2019
• Θέματα: Electronic health records; Semantic interoperability; OpenEHR; HBase; Frailty; Clustering; Manifold learning; Locally linear embedding; Non-linear dimensionality reduction; Ηλεκτρονικοί φάκελοι υγείας; Σημασιολογική διαλειτουργικότητα; Συσταδοποίηση; Μη-γραμμική μείωση διαστατικότητας; Ευθραυστότητα; 610.285

In this work we exploit the openEHR framework for the representation of frailty in ageing population in order to attain semantic interoperability, and we present the methodology for adoption or development of archetypes. Frailty is a common clinical syndrome in ageing population that carries an increased risk for adverse health outcomes including falls, hospitalization, disability, and mortality. As these outcomes a ect the health and social care planning, during the last years there is a tendency of investing in monitoring and preventing strategies. Although a number of electronic health record (EHR) systems have been developed, including personalized virtual patient models, there are limited ageing population oriented systems. We also propose a framework for a one-to-one mapping between openEHR archetypes and a column-family NoSQL database aiming at the integration of existing and newly developed archetypes into our HBase storage system. As a result a detailed personalized virtual patient model (VPM) of the health pro le has been designed and constructed by unobtrusively monitoring the older person's everyday life through a variety of embedded and wireless smart indoors and outdoors sensors, as well as social interaction, clinical assessment and self-evaluation tests. Focusing on the aging population as a case study, in this study we also present a methodology for extracting and analyzing multi-sensor measurements and use them to identify data clusters that might be linked to clinical pro les. This was achieved by fusing information of the VPM from multiple devices, such as a wearable device for physiological and kinetic monitoring, a dynamometer for strength measurement, a game platform and a global positioning system (GPS) for outdoor monitoring. Principal component analysis (PCA) was applied to remove correlations in the extracted features, followed by locally linear embedding (LLE) to embed the data in a lower dimensional space where unsupervised clustering is feasible. Exploration of the identi ed clusters revealed patterns of frailty manifestation that were in high accordance with geriatric indices from several domains (medical, cognitive, psychological, etc). Our results highlight the potential of the applied data mining methodology for the extraction of novel frailty indicators.