Supervisors info:
Αναστασία Κριθαρά, Μεταδιδακτορική Ερευνήτρια, Εργαστήριο Τεχνολογίας Γνώσεων και Λογισμικού, Ινστιτούτο Πληροφορικής και Τηλεπικοινωνιών, ΕΚΕΦΕ «Δημόκριτος»
Summary:
Chronic diseases management is one of the greatest challenges of modern healthcare systems. Given the fact that non-communicable diseases are responsible for more than 70% of deaths worldwide [1], the constant monitoring of a patient’s health condition has become vital need and, hence, the era of mobile health starts to rise. At the same time, the idea of self-managing personal aspects of life, and not only, through the prism of new technologies, the so-called Quantified Self, gains ground rapidly. Nowadays, sensors constitute an integral part of the daily life and monitor almost every aspect of it, gathering enormous quantities of data. The challenge is how to control the data that derive from the combination of electronic health services with wearable sensor technologies and broaden the horizons of scientific research [2]. At this point, data analytics assumes its decisive role. Patients using such technologies gain the capability to record and process their vital signs, fitness activities, everyday habits, or even feelings [3]. The resulting data constitute the gemstone for statistical and machine learning techniques to be performed so that knowledge discovery can take place and, as a consequence, identify the risk factors in patients’ health and provide personalized medical follow-up and immediate feedback to avoid emergent situations.
This graduate thesis proposes a data analytics solution that will examine patients’ consistency in their measurement schedule and study the interaction among the different daily measurements, with the scope of determining how these factors can influence the monitoring of their health. Studies generalized on a demographic level, including sex, age and geolocation, will also take place so that statistical significant differences can be identified in the medical values and, thus, appropriate recommendations can be derived per population group. Aiming at improving and personalizing the medical monitoring of chronic health conditions, the proposed solution can circumvent the challenges of electronic health systems and provide benefits for the involved patients, such as enhancement of their welfare, early detection of dangerous situations, assumption of further targeted monitoring, motivation to engage in self-caring activities and follow treatment and, last, modeling of their behavior to improve self-care and enjoy a better quality of life.
Keywords:
chronic diseases management, quantified self, mobile health, statistical data analysis, machine learning
