Overview
Vital sign monitoring is important for patients, especially those with underlying cardiovascular diseases. Traditional monitoring including SaO2, blood pressure, pulse, rhythm monitoring can be cumbersome, with wires and cuff attached to a patient. This can lead to reduced compliance, especially in uncooperative patients or young kids. The use of photoplethysmography (PPG) is simple and is widely used clinically to achieve continuous monitoring of SaO2, respiratory rate and pulse rate. PPG also contains important physiological signals that can correlate with blood pressure.
Remote PPG (rPPG) with can be collected by cameras. This is being developed to achieve contactless monitoring. Theoretically, by analyzing video recording, the fluctuation of intensity of lights reflected from patients' body surface can be measured. Physiological parameters such as heart rate, SaO2, and blood pressure can then be calculated. However, because of environmental signal contamination, motion artefacts and other limitations, rPPG remains inaccurate for routine clinical use. To improve accuracy of PPG recordings, machine learning approach has been attempted. This study aims to gather physiological data from patients to train these models.
Description
This is an observational study. The physiological signal collected includes blood pressure, pulse rate, respiratory rate, SaO2, cardiac rhythm. PPG signals will be extracted from dedicated devices such as commercially available SaO2 monitors. Remote PPG signals will be collected if patient consent for video recording.
Patient demographic data, including age, sex, past medical history, medication history, body weight and height, etc, may be collected to validate training of this model.
After collection of these data, machine learning approach will be used to analyze these recordings and correlate them with PPG and rPPG data.
After their enrollment, during the duration of the study, patient's electronic record may be accessed in the next 5 years to explore any association of PPG or rPPG signals and long term outcome.
Eligibility
Inclusion Criteria:
- Patients > 18 years old
- Has a known medical disease
- Able to comply with recording of physiological signals
Exclusion Criteria:
- Patients who are not able to provide informed consent