Implementing a web-based application for analysis and evaluation of heart rate variability using serverless architecture

Abstract

This article is devoted to the development of a web-based application for analysis and evaluation of Heart Rate Variability (HRV) using serverless architecture. Advancements in information algorithms and computing technologies have been playing an increasingly important role in cardiology, as continuous monitoring of patients' health can be vital to their well-being. One physiological parameter that can be easily measured and that can provide indispensable insight into the state of the human body is the HRV. HRV analysis can assess not only the physiological state of the body but also provide the capability to monitor its dynamics and predict future diseases. As the research in the sphere of cardiology is constantly growing there is a multitude of new ways to assess the physiological state of patients and provide an early indicator to pathological conditions. Therefore, there is a need to bring these advances to a growing number of end-users (health-care professionals and patients) in the shortest possible time. To address this problem, this study proposes the development of a web-based application for analysis and evaluation of HRV by applying linear and nonlinear mathematical methods. The application is created using a serverless architectural approach, which allows for fast development time, as there is no need to manage server infrastructure, and for automatic scaling to dynamically match the number of requests. The developer can instead focus on implementing the logic for the HRV analysis algorithms and deliver new improvements at a faster rate. The proposed web application can be accessed by any device that is connected to the Internet and is optimized to handle both an intermittent and a consistent volume of requests. The algorithms implemented in the web application have been validated by examining two groups of subjects (young adults and older adults) using linear and non-linear models. The obtained results from the two groups can be compared with a set of reference values (only for the linear methods) and an assessment can be made whether each studied parameter is within the normal range or outside it (its value is too high or too low). To aid the assessment for HRV, the results obtained by the linear and nonlinear analysis are presented using a set of both graphs and tables.