Healthcare has played a main role in the Semantic Web (SW) field given the knowledge representation possibilities that SW is capable of addressing. Nowadays there are a large number of ontologies which can be used for several domains of healthcare (genetics, proteins, cellular components, anatomy, and specific diseases among others). However, in some cases, the definition and population of these ontologies are not enough to be used in concrete domains. In this paper we provide the design of a set of ontologies for their direct use in diagnostic decision support systems. We have designed an ontology modular architecture where main (root) ontology is created to define the main relations which can be found in the aforementioned domain. A set of subsumed ontologies has also been designed following some principles of OBO-Foundry and using SNOMED-CT terminology as the main interoperability component. These ontologies have been also designed trying to create them as light as possible. The evaluation of the designed ontology is based on a set of quantitative aspects which aims to show the main principles which should be followed in the process of design ontologies for the domain of differential diagnosis.