Development of knowledge base structure for gout diagnosis

Abstract

Gout belongs to the group of microcrystalline arthritides and is a systemic tophus manifested by infl ammation of various tissues caused by deposition of monosodium urate crystals. Verifi cation of gout diagnosis in Russia takes on average four to eight years from the beginning of the disease. This period is suffi cient for the development of various complications (for example, gouty nephropathy) and disability of patient, and, therefore, it is necessary to revise the existing strategy for the diagnosis of this disease. One of the options for improving the provision of medical care can be attributed to the creation of a clinical decision support system (CDSS), which is based on the knowledge of experts, formalized as a knowledge base (KB).Aim. To develop a KB structure for CDSS for gout diagnosis.Material and Methods. Clinical information for gout diagnosis, presented as a nomenclature of medical concepts and logical schemes, which were collected on the basis of federal clinical recommendations, various literature sources and expert knowledge were used as materials. The ontological method was used as a method of knowledge structuring. Combined semantic network and frames were used as the methods of representation.Results. While developing KB structure, a combination of two methods of knowledge representation including semantic network and frames was used. An ontological approach was used in terms of knowledge structuring. The structure was built on the clinical knowledge collected in cooperation with experts in gout diagnosis. Compared with similar developments of diagnosing diseases based on knowledge engineering methods, the main feature of developed KB structure was the use of a separate type “Syndrome” concept as an aggregator accepted in medicine, signifi cantly reducing the volume of KB for diseases.Conclusion. The KB structure was developed comprising the use of seven types of concepts and 11 types of relationships. The structure involved the use of ontological approach and combination of two models of knowledge representation, namely: a semantic network and a frame model.

Hybrid clinical decision support for cardiology: architectural foundations for integrations

“…However, this approach is limited by simplified knowledge models that may not always reflect the complexity of real clinical scenarios. Such CDSS form explanations of the decision-making process according to their simplified knowledge model [34], but advanced knowledge models corresponding to real-world representations in this field are required for intelligent physician support.…”

Prototype Expert System for Gout Diagnosis on an Outpatient Basis

“…Prototype expert system for gout diagnosis contains two components: knowledge base (KB), including 1144 medical concepts and 5 640 522 links, and computer program incorporating problem solver. The development of the KB started with primary formalization of the subject area by nomenclature of medical terms and logical schemas for gout diagnosis [5]. These information objects were used to increase effective interaction between knowledge engineer and expert group allowing to exclude various technical issues.…”

“…KB architecture was designed using combination of two methods of knowledge representation (semantic nets and frames), ontology approach and elements of fuzzy logic [5]. Architecture describes 7 types of concepts bound by 11 types of links.…”