Master’s thesis by Sophie Wagner: Advancing Clinical Patient Monitoring: Robust RR Interval Extraction and Heart Rate Variability Analysis from Continuous Photoplethysmography Data in COVID-19 Patients
This thesis showcases the viability of obtaining dependable long-term HRV data using a cosinuss° in-ear PPG sensor. A thorough examination of the DiAssCo data set was carried out, which includes 145 patients who were hospitalized with COVID-19 at the Großhadern Clinic in Munich in the first half of 2021. All patients were monitored continuously with the cosinuss° Two, covering a wide range of disease severities and outcomes.
As HRV integration in long-term recordings is not trivial, the research contributes to understanding HRV’s potential in healthcare and emphasizes the importance of diligent data cleaning methodologies. To gain insights into COVID-19 disease progression, a machine learning algorithm was successfully employed to predict health outcomes, independent of the subjects’ age.
As a scientific consensus regarding the most descriptive HRV parameters in the context of COVID-19 has yet to be established, this thesis presents feature rankings to indicate parameters that are associated with adverse outcomes.
