Scholarlink

<p>In today’s data-driven world, integrating diverse healthcare data sources into a unified framework is essential. The COVID-19 pandemic has underscored the critical need for extracting meaningful insights from fragmented clinical data, particularly in areas such as treatment efficacy, risk factor identification, and drug interactions. To address these challenges, we propose the COVID-19 Drug and Risk Ontology (COViDRO)-a formally developed OWL-DL ontology designed to model and integrate COVID-19 treatment options aligned with the “PRADiCT” framework (Patient Risk factors, Adverse effects, Drug interaction, Clinical findings, and Treatment procedure). We hypothesize that this ontology will assist healthcare professionals in discovering and recommending COVID-19 therapeutics tailored to individual patients by considering risk factors, underlying health conditions, ongoing medications, potential drug interactions, and adverse effects. To validate its reliability and effectiveness, COViDRO underwent a multi-tier evaluation process: (1) Quality-based assessment using the Ontology Pitfall Scanner (OOPS!) to detect and resolve modeling errors, benchmarking COViDRO against related ontologies based on structural, functional, and usability dimensions; (2) Structural and logical validation using OntoDebug for structural integrity checks and the Pellet reasoner for logical consistency verification; (3) Quantitative evaluation using the OntoMetrics framework to assess ontology metrics such as attribute richness, relation richness, and knowledge base complexity, comparing with related ontologies; and (4) Query-based evaluation using SPARQL to assess the ontology’s reasoning and retrieval capacity. The evaluation results confirm that COViDRO effectively organizes 135 classes, 32 object properties, and 15 data properties, enabling structured clinical reasoning. SPARQL queries successfully demonstrate its ability to retrieve patient-specific therapeutic recommendations, assess risk factors, and generate drug interaction alerts, validating its practical utility in healthcare settings. As a formal, DL-enabled ontology, COViDRO can contribute to automated inference of treatment options, thereby enhancing decision support systems and knowledge-based applications. Its structured and extensible approach makes it a valuable resource not only for COVID-19 but also for future pandemics and infectious disease management, reinforcing its significance in healthcare informatics.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Ontological approach towards discovering and recommending COVID-19 therapeutics, risk factors, and drug interactions

  • 期刊论文

  • 发表日期：

    2025年5月7日

查看全文
  • Debanjali Bain,
  • Biswanath Dutta

摘要

In today’s data-driven world, integrating diverse healthcare data sources into a unified framework is essential. The COVID-19 pandemic has underscored the critical need for extracting meaningful insights from fragmented clinical data, particularly in areas such as treatment efficacy, risk factor identification, and drug interactions. To address these challenges, we propose the COVID-19 Drug and Risk Ontology (COViDRO)-a formally developed OWL-DL ontology designed to model and integrate COVID-19 treatment options aligned with the “PRADiCT” framework (Patient Risk factors, Adverse effects, Drug interaction, Clinical findings, and Treatment procedure). We hypothesize that this ontology will assist healthcare professionals in discovering and recommending COVID-19 therapeutics tailored to individual patients by considering risk factors, underlying health conditions, ongoing medications, potential drug interactions, and adverse effects. To validate its reliability and effectiveness, COViDRO underwent a multi-tier evaluation process: (1) Quality-based assessment using the Ontology Pitfall Scanner (OOPS!) to detect and resolve modeling errors, benchmarking COViDRO against related ontologies based on structural, functional, and usability dimensions; (2) Structural and logical validation using OntoDebug for structural integrity checks and the Pellet reasoner for logical consistency verification; (3) Quantitative evaluation using the OntoMetrics framework to assess ontology metrics such as attribute richness, relation richness, and knowledge base complexity, comparing with related ontologies; and (4) Query-based evaluation using SPARQL to assess the ontology’s reasoning and retrieval capacity. The evaluation results confirm that COViDRO effectively organizes 135 classes, 32 object properties, and 15 data properties, enabling structured clinical reasoning. SPARQL queries successfully demonstrate its ability to retrieve patient-specific therapeutic recommendations, assess risk factors, and generate drug interaction alerts, validating its practical utility in healthcare settings. As a formal, DL-enabled ontology, COViDRO can contribute to automated inference of treatment options, thereby enhancing decision support systems and knowledge-based applications. Its structured and extensible approach makes it a valuable resource not only for COVID-19 but also for future pandemics and infectious disease management, reinforcing its significance in healthcare informatics.