Building an Electronic Health Record Knowledge Graph for the Mexican Health System

An EHR collects health data from a person who considers the different aspects of medical care. This data must be more organized, updated, and accessible for information systems to process. Processing systems and representation structures are necessary to facilitate their management. These systems must integrate dependable medical information sources from national standards and medical regulations. Employing knowledge graphs to represent and manage EHR-required information provides several benefits. It enables detailed characterization of all concepts or resources within the EHR, such as patient data, pathological and non-pathological personal history, physical exams, vital signs, laboratory results, diagnoses, and therapies. Graph representation allows relationships between represented resources, thereby achieving semantic expressiveness. Thus, this paper outlines the design and development of a model using knowledge graphs to integrate all the necessary information sources for EHR management in the Mexican Health System. It provides a structured representation of EHR to make it computer-processable. The proposed knowledge graph-based model aims to support the overall care process performed by a doctor, from the initial patient interview to diagnosis and the determination of pharmacological treatment.