An Expert System for Differential Diagnosis Based on ICD-10

This paper presents the prototype of an expert system for differential diagnosis based on the symptoms, signs and lab findings of a patient. The system can be used either for consultation or for educational purposes. Users of the system may be patients, physicians and students of a Medical School. Initially, the user submits to the system the symptoms of a patient. Then, the system makes a set of questions to the patient. Finally, the system performs differential diagnosis based on the given data. The system follows a clinical approach and the international classification of diseases (ICD-10). It has been implemented in Prolog. The initial diagnostic performance of the system is encouraging and promising. The system performance is 90% of the cases for a tested set of 120 cases of patients. Our system made successful differential diagnosis in 108 cases of patients. In 12 cases it made false negative diagnosis.