Continuous Non-Invasive Measurement of Core Vital Signs in the Hospital Ward
No Thumbnail Available
Author
Bonnier, Guillaume
Adam, Karen
Proust, Yara-Maria
Renevey, Philippe
Braun, Fabian
Proença, Martin
Theurillat, Patrick
Vasireddy, Rakesh
Degiorgis, Yan
Fuligno-Schmidt, Christian
DOI
10.1515/bmt-2025-1001
Abstract
Methods Data was collected from various protocols using the vital+ device developed by Rheo AG. The SpO2+HR dataset involved 29 participants undergoing controlled hypoxia protocols, the RR dataset included 25 participants with RR protocols, and the BP dataset comprised 107 ICU patients. The SpO2 and HR estimates were estimated using red, infrared, and green PPG, over 10 second windows. The RR algorithm processed PPG signals to estimate RR from frequency spectra of IBIs and pulse amplitudes. BP estimates were obtained using a proprietary pulse wave analysis algorithm, mapping PPG waveform features into BP estimates over 180second windows. HR accuracy was assessed using RMSE between PPG- and ECG-derived HR, as per FDA and ISO standards. SpO2 accuracy was evaluated using acceptance rate and RMSE using blood sample references, as recommended by ISO 806012-61:2017. RR was evaluated using RMSE between estimated RR and reference flow sensor values. BP accuracy, precision and trending ability were evaluated against an invasive arterial line, using the ISO 81060-2:2018 criteria as agreement limits.
Results The device showed promising results: SpO2 RMSE of 2.06%, RR RMSE of 0.88 cycles/ minute, systolic BP error of -0.1±7.8 mmHg, diastolic BP error of 0.2±4.3 mmHg, and HR RMSE of 2.08 beats/ minute.
Conclusion The study demonstrates the feasibility of measuring various vital signs with a single non-invasive wearable device, meeting medical requirements. This device has the potential to improve patient monitoring in hospital wards, enhancing patient care and reducing workload medical staff workload.
Publication Reference
BMT 2025, Muttenz (Switzerland)
Year
2025-09-11