Abstract
The application of ontologies in the medical field has become very popular in the last decade. These ontologies are developed to meet the needs of understanding, sharing and integrating knowledge. The development of a new medical ontology is a serious challenge since existing methodologies do not provide much guidance on how the steps should be carried out. This challenge becomes more difficult when there is a multitude of sources of medical knowledge and ontologies covering the field, each having a different representation of the same concept. Moreover, ontology development is a multidisciplinary work involving domain experts and knowledge engineers. Bringing together such a team is not easy, and therefore often, ontologies are created with limited expertise either in the medical domain or in ontology engineering domain.
In this context, our reflection focused, on the one hand, on the construction of a new detailed methodology for creating a medical ontology with limited involvement of domain experts. On the other hand, on a proof of concept that consists of applying the methodology on the development of an ontology covering the pneumonia diagnosis. The methodology uses clinical practice guidelines (CPG) to build a preliminary ontology and then, reuse existing ontologies that are relevant to the domain of interest. The methodology guides the ontologist to resolve inconsistencies and conflicts between concepts having different representations in different ontologies. It is only in the domain definition and the evaluation phases when the domain experts are consulted. To enhance interoperability and reusability with other biomedical ontologies, the methodology follows the OBO Foundry principles.
The methodology has been tested for building a pneumonia diagnosis ontology (PNADO). The definition of the PNADO domain as well as its evaluation involved several physicians. The results of the evaluation validated our methodology and showed its effectiveness.
PNADO is the first reported pneumonia diagnosis ontology developed to represent different diagnosis aspects in a formal logical format. It covers symptoms, clinical signs, types of pneumonia, history, pathogens, differential diagnoses, complications, laboratory tests, imaging, and results. It will be used by a clinical decision support system that will be integrated into the SEKMED (Software for the Evolution of Knowledge in MEDicine) platform.
Although our methodology was intended for medical ontology building, it can be very well applied to the building of ontologies covering the domains of other fields, by using documents that synthesize the knowledge of the domain, such as CPG.
In this context, our reflection focused, on the one hand, on the construction of a new detailed methodology for creating a medical ontology with limited involvement of domain experts. On the other hand, on a proof of concept that consists of applying the methodology on the development of an ontology covering the pneumonia diagnosis. The methodology uses clinical practice guidelines (CPG) to build a preliminary ontology and then, reuse existing ontologies that are relevant to the domain of interest. The methodology guides the ontologist to resolve inconsistencies and conflicts between concepts having different representations in different ontologies. It is only in the domain definition and the evaluation phases when the domain experts are consulted. To enhance interoperability and reusability with other biomedical ontologies, the methodology follows the OBO Foundry principles.
The methodology has been tested for building a pneumonia diagnosis ontology (PNADO). The definition of the PNADO domain as well as its evaluation involved several physicians. The results of the evaluation validated our methodology and showed its effectiveness.
PNADO is the first reported pneumonia diagnosis ontology developed to represent different diagnosis aspects in a formal logical format. It covers symptoms, clinical signs, types of pneumonia, history, pathogens, differential diagnoses, complications, laboratory tests, imaging, and results. It will be used by a clinical decision support system that will be integrated into the SEKMED (Software for the Evolution of Knowledge in MEDicine) platform.
Although our methodology was intended for medical ontology building, it can be very well applied to the building of ontologies covering the domains of other fields, by using documents that synthesize the knowledge of the domain, such as CPG.
| Translated title of the contribution | New methodology for building medical ontologies: case study: pneumonia diagnosis |
|---|---|
| Original language | French |
| Qualification | Doctor of Philosophy |
| Awarding Institution |
|
| Award date | 2 Mar 2021 |
| Place of Publication | Gatineau, Quebec |
| Publisher | |
| Publication status | Published - 5 May 2022 |
| Externally published | Yes |