Remote Monitoring of Cardiorespiratory Parameters in COVID-19 Patients Following Hospital Discharge

Ganesh Raam Kumarasamy1, Hélène De Cannière2, Julie Vranken2, Michiel Jacobs3, Peter Karsmakers3, Pieter Vandervoort2
1Hasselt University, 2Mobile Health Unit, Hasselt University, 3Declarative Languages and Artificial Intelligence - Advanced Integrated Sensing, KU Leuven


Abstract

Aim: The onset of COVID-19 created a need for monitoring patients outside the hospital to avoid overburdening of healthcare resources. Remote monitoring proved to be an effective solution to this problem. We hypothesised that studying the progression of cardiorespiratory parameters in COVID-19 patients after hospital discharge via remote monitoring can provide better insights into disease status. In this study, changes in parameters including HR, temperature and SpO2 were analysed.

Methods: 24 patients who were admitted to a hospital due to COVID-19 complications were provided with a pulse oximeter, thermometer and chest-based wearable patch on discharge. Discrete data was collected 3 times a day and comprised SpO2, HR, temperature and questionnaire information on general well-being and dyspnea. Continuous data from the patch comprised single-lead ECG and 3-axis accelerometer signals. Upon exceeding predefined parameter thresholds, alerts were generated and the corresponding caregiver was notified. Patients collected data for at least 5 days and were categorised into three groups for analysis – re-hospitalized, alert-generating and recovered groups. The progression of discrete data within these groups was studied using exploratory data analysis.

Results: HR, SpO2 and dyspnea score showed the highest difference between patient groups. Increase in median resting HR was highest among re-hospitalized patients, increasing by 6 bpm each day. The dyspnea score of re-hospitalized patients remained constant at a value of 5 on a scale of 1-8 while the corresponding value for other groups decreased during the study. Though SpO2 values of re-hospitalized patients increased by 1% each day, the end-study values (93%) were lower than the baseline value of the other two groups (94%).

Conclusion: The parameters that differed most between the patients were SpO2 and HR. While these changes were evident in the discrete data, more insights will be obtained from statistical analysis and processing of the continuous sensor data.