In order to provide dependable healthcare services for the elderly, it is necessary to have a patient-centric healthcare architecture in which context-aware healthcare services can be provided at any time and anywhere. Such a service automation has the virtues to overcome the disadvantages arising from the disabilities that are inherent in the elderly population, physically challenged, and those who live in remote areas. In order that patients trust the healthcare services provided by the system, the creation of healthcare services must be founded on accurate personalized health model of patients, and must be delivered by experts through dependable medical devices and secure channels. Motivated by this goal, this thesis proposes a layered health model that can be personalized to meet the privacy requirements of a patient, and constructs a context-aware healthcare architecture in which healthcare services for each patient is specialized based on personalized health models, health contexts, and emerging health situations. A prototype implementation of the architecture is validated for Hypertension and Dementia case studies.