A closed-loop system for pervasive health

As life expectancy in the industrialized world increases, so does the number of elders with chronic health conditions such as diabetes and congestive heart failure that require complex self-management routines. The traditional model of episodic care in clinic and hospital-based settings is not optimal for improving chronic disease outcomes. Instead, solving these issues requires a low-cost longitudinal monitoring method of health and daily activities, as well as techniques for effective behavioral modifications. Fortunately, we are starting to acquire the tools that will help us counter those challenges. Supported by innovations in physiological sensing, wireless communication, and cloud computing, the marketplace of consumer medical devices is booming. We term such devices conducting continuous monitoring of people's everyday lives as pervasive health devices. Likewise, we define Web/mobile applications that analyze and visualize the data from pervasive health devices for users and caregivers are called as pervasive health applications. However, existing pervasive health devices and applications pose the following challenges to be effective in practice. First, they lack caregivers connection and feedback. Due to the absences, people tend to lose their interests in using devices quickly. More importantly, the existing pervasive systems are designed closed and vertically-integrated, thereby impeding the development of sophisticated pervasive health applications for complex, multi-factorial disease and health conditions. Furthermore, the closed systems complicate unified management of pervasive health devices. In this dissertation, we propose a closed-loop system to solve the challenges in pervasive healthcare. In the first part, we introduce the overall concept of closed-loop design, as well as the hardware and software components comprising the loop. In doing so, we also discuss what are the potential impacts that it can make on caregivers, device users, and application developers. In the second part, we focus on HealthOS: a platform to develop pervasive health applications and also a core system to integrate all other components in our close-loop. The final part of the dissertation introduces Trend Finder: a system to detect trends/events in data streams from devices.