The diagnosis of epilepsy in the interactive mode using computer technology

: The paper describes an EDITA programming subsystem for computer diagnostic technology for epilepsy. The EDITA is based on a classifier, a system of a patient's individual signs. The classifier has been developed in accordance with the epilepsy classification developed by the International Epilepsy Control League (Kyoto, 1982) and involves many-year experience in diagnosing and treating epilepsy in Udmurtia. The crucial principle for a computer to make a diagnosis is not only medical experience, but a formalized system of weight ratios, which is an integral part of the computer classifier. The validity of weight ratios were assessed in a representative taught sample (260 patients with wellknown types of epilepsy). A statistic method based on the Bayes' approach was applied by employing the same taught sample when the classifier and the reliability of computer-aided diagnosis were checked. Computer-aided diagnosis was made in three steps: registrations or selection of a patient, collection of history data and diagnosis, prescription of antiepileptic drugs. According to these three steps, three sections (or program operating modes) were identified. These included a patient's medical card, a classifier, and drugs.