An Ontology-Driven Application to Improve the Prescription of Educational Resources to Parents of Premature Infants

CST’s Baby CareLink provides a ‘collaborative healthware’ environment for parents of premature infants that incorporates just-in-time learning as one means of knowledge exploration and patient empowerment [1], [2], [3]. As the Baby CareLink content base has continued to grow, it has become increasingly difficult for content prescribers to identify all relevant resources for parents at a given point in time in their child’s course of care. In addition, the growing content base has become increasingly difficult to maintain without a rich indexing system. In order to address these issues, we have developed an ontology-driven application that supports the indexing and retrieval of educational materials according to rich descriptions of premature infants. We have developed an initial OWL-DL-based [4] ontology describing relevant concepts in the domain of neonatology, including clinical conditions, diagnostic testing, therapies, durable medical equipment, and the infants themselves. We have developed an initial terminologic model describing typical clinical problems and therapies that occur over the clinical course of these premature infants. Indexers tag educational resources through a web based client application that allows them to create rich descriptions of educational resources based on the reference ontology. Clinical end-users interact with a client application that identifies educational resources appropriate to clinical scenarios occurring over the course of a typical premature infant’s development based on the description generated from records of existing infants in the Baby CareLink system, or according to a user-created description. Network Inference’s tools were used to develop the neonatology ontology and implement the run-time system. ConstructTM, a Visio-based ontology modeling tool was used to develop the reference ontology. This tool allowed the representation of the domain’s concepts, subsumption relationships, properties, instances, and axioms diagrammatically. Cerebra ServerTM, an OWL-DL-based inferencing platform was used to provide logical consistency-checking at modeling time directly from ConstructTM. Cerebra ServerTM was integrated with Baby CareLink and to the indexing and retrieving tools through a .Net connector that provides an API for 1) extending the ontology at indexing time with newly-classified educational resources, 2) dynamically creating instances of individual premature babies using data from Baby CareLink and 3) querying the ontology at run-time. Howard Goldberg et al. This knowledge centric approach to the identification and recommendation of educational resources is anticipated to better individualize information for parents of infants managed through Baby CareLink and increase the efficacy of the information prescription process. The approach is expected to increase the return on investment provided by Baby CareLink through reduction in the time required by nurses and care managers to interact with the system.