Edge Intelligence in Healthcare: A Systematic Literature Review

Healthcare has gained a new traction to personalized care with the introduction of technological monitoring devices. These systems enable non-invasive recording of physiological parameters such as blood pressure, glucose levels, sleep patterns, heart rate, and water levels. Furthermore, with the innovation of biomaterials, it is possible now to measure bioelectric signals with high precision and accuracy, which, through machine learning algorithms, allows irregular patterns to be identified. However, this type of technology lacks data/information storage infrastructures, whose architecture has shortcomings such as response time and privacy. Hence, new paradigm has emerged based on peripheral computing technology and on-device artificial intelligence-edge Intelligence (EI) to resolve some of these issues. This type of architecture improves the efficiency-effectiveness ratio in healthcare provision. Through ultra-thin, skin-compatible, flexible, and high-precision devices and sensors, low-cost communication technology, on-device intelligence, edge intelligence, and edge computing technology enable a personalized application of services. To deepen the discussion, this article followed a Systematic Literature Review using the PRISMA protocol (Preferred Reporting Items for Systematic Review and Meta-Analysis Protocol). The Scopus database was selected to explain the concept and assess the benefits and challenges of EI in healthcare today. As a result, this article presents a conceptual theoretical model for EI in the health sector and discusses the most important topics such as its capabilities and limitations.