Ontologies, or data dictionaries, dictate the ease of use and success of self-service research query applications, such as Informatics for Integrating Biology and the Bedside (i2b2). Users in the CTSA community have expressed a need for greater usability and usefulness in the ACT Ontology. The ACT Ontology supports queries across multiple research institutions in the ACT Network (SHRINE).
Usability in this context refers to the ability to navigate hierarchies naturally and intuitively, understanding the meanings of terms, and finding classes of terms that are relevant to research retrievals. Usable ontologies include classes that are findable and reduce the need for users to assemble long lists of relevant terms to cover the topic of interest. Usefulness means that queries work as expected, such that proper data are retrieved, but much work is needed to rectify deficiencies and discrepancies in the healthcare coding system. We have been developing the UAB Foundational Ontology (UFO), which has a secondary organizational structure of standard terminologies that improves usability and usefulness. We aim to expand these efforts through collaboration to gain insights from and improve querying from other CTSA hubs and the wider CTSA community.
We have been discussing collaborative development of i2b2 ontologies with the CTSA community for the past three years. Colleagues from several CTSA hubs have expressed that they are supportive and interested in participating in our effort. We will continue to solicit members through various outreach methods (education, training, and publications) to have broad and diverse representation from the CTSA community, including developers and maintainers of the official ACT Ontology. The project’s success will be assessed through usability testing, interviews, and surveys. Qualitative measures include issues and difficulty developing enhancements, how the enhanced ontologies affect queries, the overall user experience, and feasibility. Quantitative measures that will compare standard and enhanced ontologies will include survey ratings, such as satisfaction, ease of use, likelihood of finding correct terms, ratings of usefulness, usage statistics, and time needed to find appropriate terms and build queries. Additional measures include time needed to create enhancements. These measures will drive iterative improvements.