Ontology for assessment studies of human–computer-interaction in surgery

• We developed an investigation model for assessment studies of HCI in surgery.
• We compared four top-level ontologies for model implementation: OCHRE, BFO, GFO, DOLCE.
• GFO was selected for investigation model implementation in the form of a new ontology.
• The investigation model and its ontology provide a modular study guideline.
• Investigation model ontology is a tool to create knowledge bases of HCI assessment.

ObjectiveNew technologies improve modern medicine, but may result in unwanted consequences. Some occur due to inadequate human–computer-interactions (HCI). To assess these consequences, an investigation model was developed to facilitate the planning, implementation and documentation of studies for HCI in surgery.Methods and materialThe investigation model was formalized in Unified Modeling Language and implemented as an ontology. Four different top-level ontologies were compared: Object-Centered High-level Reference, Basic Formal Ontology, General Formal Ontology (GFO) and Descriptive Ontology for Linguistic and Cognitive Engineering, according to the three major requirements of the investigation model: the domain-specific view, the experimental scenario and the representation of fundamental relations. Furthermore, this article emphasizes the distinction of “information model” and “model of meaning” and shows the advantages of implementing the model in an ontology rather than in a database.ResultsThe results of the comparison show that GFO fits the defined requirements adequately: the domain-specific view and the fundamental relations can be implemented directly, only the representation of the experimental scenario requires minor extensions. The other candidates require wide-ranging extensions, concerning at least one of the major implementation requirements. Therefore, the GFO was selected to realize an appropriate implementation of the developed investigation model. The ensuing development considered the concrete implementation of further model aspects and entities: sub-domains, space and time, processes, properties, relations and functions.ConclusionsThe investigation model and its ontological implementation provide a modular guideline for study planning, implementation and documentation within the area of HCI research in surgery. This guideline helps to navigate through the whole study process in the form of a kind of standard or good clinical practice, based on the involved foundational frameworks. Furthermore, it allows to acquire the structured description of the applied assessment methods within a certain surgical domain and to consider this information for own study design or to perform a comparison of different studies. The investigation model and the corresponding ontology can be used further to create new knowledge bases of HCI assessment in surgery.