Autonomous Self-Diagnosis System

With the latest technological advancements and the unprecedented number of Internet connected devices, digital health products have become an attractive topic. However, when it comes to self-diagnosis systems, even if the state-of-the-art applications are very popular, their functioning is undisclosed. There is no transparency in terms of dataset, medical contributions and symptom assessment algorithms. Therefore, this paper proposes an open-data, open-source and community-driven data aggregation system that can receive and validate medical contributions from around the world. The resulting dataset enables the development of heuristic-driven self-diagnosis systems that can provide the statistical likelihood of having a particular condition or disease. Our solution for obtaining this dataset aims to promote transparency and trust among these autonomous diagnosis systems. The aggregation pipeline is designed with guidance from medical specialists and the collected dataset will be validated, anonymized and made publicly available. A self-diagnosis system was designed as a proof-of-concept for the open-data platform. A symptom-disease knowledge database was used as dataset and the system was deployed on a cloud-native environment so it can be validated by doctors and users. The idea of having a publicly available and community-driven medical dataset was validated by 27 (93.1%) out of 29 interviewed doctors. The proof-of-concept was assessed as “correct” by 17 (58.6%) specialists out of the same 29 doctors, and as “satisfactory” by 23 (69.7%) users out of a total of 33 interviewed users.