Development of a domain ontology for decision support in the treatment of gait-related diseases

Abstract

The growing complexity in diagnosing and treating gait-related diseases necessitates the existence of a domain ontology that can facilitate intelligent decision support on gait analysis. The study aimed to develop a domain ontology that can support decision-making on the treatment of gait-related diseases. The motivation for this study is deeply rooted in the multifaceted and interrelated challenges presented by gait-related diseases in both the diagnosis and treatment contexts. A coherent and accessible knowledge base stands as a tool to facilitate informed decision-making. This will ensure that gait experts are adeptly navigated through the complexities of numerous gait disorders, enhancing their ability to make precise and consistent clinical decisions. The domain ontology aimed to bridge existing knowledge gaps, streamline the retrieval and application of critical information, and ultimately, enhance the precision and consistency of clinical decisions by providing a unified and comprehensive view of the domain. The study adopted the Ontology Development 101 methodology to ensure a systematic, structured, and replicable approach to ontology creation. The Ontology 101 methodology guided the research through pivotal stages, including identifying the domain and scope, enumerating crucial terms in the domain, defining classes and properties, establishing class hierarchies, and ensuring detailed documentation. The Protégé ontology editor, a standard open-source ontology development tool, was used to create the gait analysis domain ontology (GADO). The developed domain ontology was evaluated using a combination of ontology verification and ontology validation procedures. For ontology verification, the study employed the Framework for Ontology Conformance Analysis (FOCA) evaluation methodology, analysed domain task fit using competency questions, and assessed content richness. The ontology verification underscored the GADO's proficiency in domain task fit and content richness, substantiating its potential as a viable tool for supporting clinical decisions in the domain of gait-related diseases. The validation of the GADO was done by utilising ontology reasoners, HermiT, and Pellet to determine structural and logical consistency of its components and its correctness. Description Logic (DL) queries and SPARQL queries were also used to assess the ontology’s aptitude in representing domain knowledge accurately and its ability to answer domain-specific queries. This was accomplished by trying to address some of the competency questions using SPARQL queries to evaluate the GADO's capacity to fetch pertinent ontological instances (individuals). The results from the ontology validation process indicate that the GADO effectively supports the retrieval of domain-specific knowledge. This ontology, therefore, stands poised to significantly impact the field by enhancing decision-making related to gait-related diseases and paving the way for future advancements in an AI-driven context that can facilitate clinical decision support.