Development A Guideline-based Decision Support System to Diagnosis of Primary Immunodeficiency Diseases

Abstract

Introduction: Primary immunodeficiency diseases (PID) are generally rare genetic disorders affecting the immune system. Overlapping PIDs symptoms and signs is a challenge to diagnosis and treatment. On the other hand, remembering of all diagnosis criteria is difficult for practitioners. The purpose of this research is developing guideline-based clinical decision support system for diagnosis of primary immune deficiency diseases, to assist practitioner in order to diagnose of disease in early stage and to minimize complications of such diseases. Material and Methods: To provide data a checklist was used and most important demographic information, symptoms, family history, physical findings and laboratory findings to diagnose eight common PIDs extracted from guidelines and literature under specialists opinion. The diagnosis inference model design and develop in Protege (version 3.4.8) frame based ontology modeling using "Noy and McGuinness" method. Then the mobile based inference model in Eclipse (SDK version 3.7.1) software has been developed and clinical decision support system of primary immunodeficiency has been created. Results: To design the diagnosis inference model in Protege software, data were classified in 5 main classes and 24 subclasses as hierarchical. Then, specific properties of each class, and determine the value of each property. Then define Instances of each class and initialized instance properties. Then use this model to develop CDSS based on mobile in Eclipse software. At the end, the inference model and the CDSS test with 110 patient’s record data and both of them recognized all 110 patient correctly such as specialist recognition. Conclusion:Guideline-based decision support systems help to detect diseases correctly, quickly and early. Guideline-based decision support systems are very reliable to practitioner, because guidelines are accepted to their. These systems reduce the forget probability of diagnosis stages and percentage error of diagnosis by practitioner and increase the accuracy of diagnosis.