Ontology-Driven Search and Triage: Design of a Web-Based Visual Interface for MEDLINE

Area of Science:

• Information Science
• Medical Informatics
• Human-Computer Interaction

Background:

• Users struggle to articulate information needs and manage large document sets when searching health literature.
• Existing search interfaces like PubMed offer limited support for exploratory search tasks.

Purpose of the Study:

• To enhance exploratory search by integrating a formal ontology for vocabulary support and multi-stage triaging for information overload.
• To develop an interactive visual interface addressing key challenges in health literature retrieval.

Main Methods:

• Developed Ontology-Driven Visual Search and Triage Interface for MEDLINE (OVERT-MED).
• Utilized the Human Phenotype Ontology (HPO) to address vocabulary challenges.
• Implemented multi-stage triaging with interactive visualizations for managing large datasets.

Main Results:

• Formative evaluation indicates OVERT-MED effectively addresses vocabulary and information overload issues.
• Formal ontology use aids users in articulating information needs with precise terminology.
• Multi-stage triaging and interactive visualizations show promise in mitigating information overload.

Conclusions:

• The developed strategies are valuable for improving exploratory search in health and medical literature.
• The approach, tested with OVERT-MED, is anticipated to be transferable to other contexts and document collections.