A Sensor Network Utilizing Consumer Wearables for Telerehabilitation of Post-Acute COVID-19 Patients

Abstract

A considerable number of patients with COVID-19 suffer from respiratory problems in the post-acute phase of the disease (the second-third month after disease onset). Individual telerehabilitation and telecoaching are viable, effective options for treating these patients. To treat patients individually, medical staff must have detailed knowledge of their physical activity and condition. A sensor network that utilizes medical-grade devices can be created to collect these data, but the price and availability of these devices might limit such a network's scalability to larger groups of patients. Hence, the use of low-cost commercial fitness wearables is an option worth exploring. This article presents the concept and technical infrastructure of such a telerehabilitation program that started in April 2021 in the Czech Republic. A pilot controlled study with 14 patients with COVID-19 indicated the program's potential to improve patients' physical activity, (85.7% of patients in telerehabilitation vs. 41.9% educational group) and exercise tolerance (71.4% of patients in telerehabilitation vs 42.8% of the educational group). Regarding the accuracy of collected data, the used commercial wristband was compared with the medical-grade device in a separate test. Evaluating zscores of the intensity of participants' physical activity in this test, the difference in data is not statistically significant at level p=0.05. Hence, the used infrastructure can be considered sufficiently accurate for the telerehabilitation program examined in this study. The technical and medical aspects of the problem are discussed, as well as the technical details of the solution and the lessons learned regarding using this approach to treat COVID-19 patients in the post-acute phase.