In pervasive medical systems with context-aware computing facility, the quality ofrndecisions made by medical professionals is influenced by the quality of context informationrnsupplied by the context management sub-system. Central to context management is contextrninformation refinement aimed at deriving context information that can assist applications inrnmaking valuable decisions about what to deliver to users. In this thesis, we identified thernshortcomings of existing works in relation to context information refinement in pervasivernmedical systems. The shortcomings are lack of adequate consideration for: qualityrnparameters of context information, relevance of context information and particularrnrequirements of the pervasive healthcare domain.rnIn order to overcome these shortcomings, we proposed a context information refinementrnarchitecture that facilitates and coordinates the refinement procedure starting fromrnacquisition of context information up until the refined context information is delivered tornthe target application (user) in a pervasive medical system by addressing the abovementionedrnshortcomings. The architecture is composed of the client mobile device end andrnthe context refinement server. The client mobile device end consists of componentsrnresponsible for tracking the device context (E.g. location), listening to incoming eventsrnfrom the server, providing interface for specifying service constraints and local cachingrnservices. On the context refinement server side, the major components are the reasoningrnand decision engine that performs ontology-supported context reasoning, the servicernparameters manager that maintains up-to-date list of service constraints, the contextrnontology that models the concepts and relationships between concepts in the pervasivernhealthcare domain and the context acquisition component that collects and aggregatesrncontext data from potential context data sources like sensors.rnTo demonstrate the validity of the proposed architecture, we developed a prototype thatrnimplements the core components of the proposed architecture. The implementation hasrnbeen evaluated with a real-life pervasive healthcare scenario and encouraging initial resultsrnhave been obtained as an indication to the usability of the proposed architecture in a reallifernsetting.