The interoperability of clinical information systems is particularly complicated due to the use of outdated technologies and the absence of consensus about standards. The literature applies standard-based approaches to achieve clinical data interoperability, but many systems do not adopt any standard, requiring a full redesigning process. Instead, we propose a generic computational approach that combines a hierarchical organization of mediator schemas to support the interoperability across distinct data sources. Second, our work takes advantage of knowledge bases to be linked to clinical data, and exploit these semantic linkages via queries. The paper shows case studies to validate our proposal.