Smart City initiatives aim to enhance the effective management of resources while providing quality services to citizens. Central to these initiatives is the use of large-scale datasets that enable intelligent analytics and reasoning components in support of resource optimisation and service provision. Recently, there has been a growing interest in aspects of smart living, particularly due to the increasing adoption and use of Electronic Health Records (EHR).   A Smart City can introduce intelligent systems to support the usage of EHR to improve emergency response services. For instance, data derived from EHR is used in primary emergency care, as a component of emergency decision support systems and for monitoring public health. However, the delivery of healthcare information to emergency bodies must be balanced against the concerns related to citizens’ privacy. Besides, emergency services face challenges in interpreting this data; the heterogeneity of sources and the large amount of available information represents a significant barrier.   This thesis investigates the use of EHR for deriving useful information about people requiring assistance during an emergency, focusing on making rich data accessible to emergency services while minimising the amount of exchanged information. To perform this task, an intelligent system needs to estimate the probability that a potentially relevant condition mentioned in a health record is still valid at the time of the emergency. During our research work, we followed a knowledge engineering approach and developed the required knowledge components to support the intelligent delivery of relevant health information about people involved in an  emergency situation. These components, which include a knowledge component for representation and reasoning, and a novel knowledge base modelling the evolution of a large number of health conditions, form the basis of CONRAD, a system which is able to support effectively decision-making in an emergency scenario.