RayaTabib, a new Persian medical advisor application based on artificial intelligence

Background and aims: In recent years, the use of self-service healthcare systems, that allow peopleto manage their medical concerns, has become a global trend. Mobile applications are widelyused and play an important role in this regard.Method: We have developed an innovative artificial intelligence-based symptom-checker application,called “RayaTabib”. First, a team of ۱۰ medical students, extracted background scientificdata, including symptom-disorder lists, as well as epidemiologic data, common clinical findingsand risk factors of disorders from several medical textbooks and online databases in the field ofdifferential diagnosis. We revised the collected data according to Iranian population epidemiology,unified the terminology and translated the entire textual material into Farsi. Then, a team of ۳software engineers organized the server database, designed the graphic materials and developedboth server and client sides of the system. The output product is a web application that can be usedon any device regardless of its operating system type.Results: The current version is beta version in Farsi language, and covers more than ۲۰۰ symptomsand ۳۰۰۰ disorders in all medical fields. RayaTabib, is a decision-support system based onartificial intelligence. The user enters age, gender and one symptom, then the system asks specificquestions about other symptoms and underlying risk factors, and then analyzes the answers onthe server and displays the result within seconds. Database on the server, includes all lists andcorrelations between symptoms, disorders and risk factors. A software on the server extracts thelist of disorders common between all selected symptoms, and computes an index for each disorderaccording to the information entered by the user. The final list of possible disorders is beingsorted according to indices and is completely unique for each patient. RayaTabib, also providesrecommendations about whether the patient needs urgent hospital care, family physician visit orspecialist visit. This feature is rule-based and is compatible with Iranian healthcare referral system.We have not presented RayaTabib for public use and we first need to evaluate its usabilityand accuracy, and define our business model.Conclusion: RayaTabib as a virtual primary care provider, can facilitate access to various healthcareservices, reduce unnecessary visits or costs, and ultimately improve Iranian society health.We have conducted several steps of systematic test and user experience test and early results arepromising. According to the experience of similar projects in developed countries, RayaTabibseems to be usable in Iran as the first Iranian symptom-checker and medical advisor application.We will conduct a study to evaluate the accuracy of the application regarding its clinical diagnosisand recommendations.