Evaluation of an Intelligent Electronic Prescription Ontology Using Open Source Tools: A Botswana Case Study

Abstract

Healthcare interoperability in Botswana is hindered by fragmented information exchange across public and private providers, resulting in prescription conflicts, duplicated treatment, and delayed care. In prior work, we developed the Intelligent Electronic Prescription Ontology (IEPO), an open-source OWL 2 ontology for semantically standardized prescription data exchange. Although IEPO was previously validated using competency question-driven metrics (recall, precision, and F1-score), those measures did not assess ontology structural quality or schema-level expressiveness. This paper presents a post-development evaluation of IEPO using two opensource ontology assessment tools: The Ontology Pitfall Scanner (OOPS) and OntoMetrics, with the Prescription of Drug Ontology (PDRO) used as a benchmark. OOPS identified initial IEPO modelling issues in consistency (60%) and completeness (40%), which were resolved through iterative refinement in Protégé, resulting in 0% residual detection error under the evaluated OOPS criteria. In contrast, PDRO retained 100% detection error across the reported OOPS categories in the baseline comparison. OntoMetrics further showed that IEPO outperformed PDRO in class richness (1.000 vs. 0.001), attribute richness (4.429 vs. 0.012), relationship richness (0.839 vs. 0.175), and axiom/class ratio (63.93 vs. 10.62), indicating stronger structural expressiveness and richer schema representation. These findings show that IEPO is not only functionally queryable but also structurally robust and semantically well formed according to open-source rule-based and quantitative evaluation criteria, strengthening its readiness for future standards-aligned e-prescription interoperability in Botswana’s healthcare ecosystem.