A comprehensive overview of the medical expert system CeliPert, the development and peculiarities of its knowledge base, its inference mechanism, its components and their implementation is given within the scope of this master thesis. CeliPert was developed in order to assist in the complex procedure of making not only a diagnosis but also providing therapy suggestions for celiac disease.

Celiac disease is a disease of the proximal small intestine characterized by permanent intolerance to gluten. Oral gluten intake results in an abnormal small intestinal mucosa, followed by malabsorption and clinical signs of disease. Removal of gluten from the diet leads to full clinical remission and restoration of the small intestinal mucosa to normal. Celiac disease is a life-long disorder that may present either in childhood or in adult life.

The knowledge base of the expert system was created both, in cooperation with the physicians of the Department of General Pediatrics of the University´s Children´s Hospital Vienna, and by review of the patient cases from the respective database. This knowledge base contains rules that make it possible to detect and to exclude celiac disease with different grade of probability. A graphical knowledge acquisition system serves for extension and modification of the knowledge base, which can be represented as a decision graph too.

After a first version of CeliPert had been completed, it was verified by means of 107 retrospective test cases of patients from the University´s Children´s Hospital. First results seemed very promising, but a modification of the knowledge base of CeliPert is necessary in the future in order to reduce its size and complexity and thus shorten CeliPert’s response time.

CeliPert has been developed by the Department of Medical Computer Sciences, and the Department of General Pediatrics, University of Vienna Medical School, Austria.

Core Unit for Medical Statistics and Informatics (MSI), Medical University of Vienna, Spitalg. 23, A-1090 Vienna, Phone: +431 40400 6668, Internet: http://www.meduniwien.ac.at/msi/mes.