Comprehensive Geriatric Assessment (CGA) is an integrated clinical process to evaluate frail elderly people in order to provide them with customized therapy plans. The whole process includes the completion of standardized questionnaires or specific movements, which are performed by the patient and do not necessarily require the presence of a medical expert. With the aim to automate CGA tests in as much as possible, we have designed and developed CLARC: a mobile robot aimed at helping physicians to capture and manage data during the CGA procedures, mainly by autonomously conducting a set of predefined tests. The design of CLARC has required dealing with both functional (robot’s skills and tasks) and non-functional aspects (e.g. performance, safety, or user satisfaction, among others). This paper describes a novel model-based approach aimed at helping designers (1) to specify the contextual information available to the robot; the Non-Functional Properties (NFP considered relevant for a given application; and how (and to what extent) changes in the context may affect these properties; and, from these models (2) to generate the runtime infrastructure allowing the robot to monitor its execution context and estimate high-level QoS metrics to know how well it is performing in terms of the selected NFPs.