A Study on the Subject and Location Specificity in Reflectance based SpO2 Estimation using R-value based Calibration Curve

Design and development of an accurate and reliable pulse oximeter depend on various critical components ranging from sensors, cabling and electronics, signal processing, value estimation and informing the results to the caretaker in a timely manner. Starting from the optical interface between the sensor and tissues, optical, mechanical and electrical design of the sensor and sensor calibration serves as the building block of pulse oximetry technology. In this paper, we present the analysis of reflectance Photoplethysmogram (PPG) signals acquired using a custom-built data acquisition platform in a clinical setting on 28 subjects suffering from various pulmonary disorders. The paper mainly focuses on the feasibility of using R-value for blood oxygen saturation level (SpO
2
) measurement using reflectance PPG signals, which is a widely accepted method. All analysis in this study is done using a NellcorTM transmittance probe as the reference. R-value obtained from reflectance PPG signals shows a correlation coefficient of 0.9 and above against the R-value computed from transmittance PPG signals for individual subjects on a specific location. The reflectance PPG based R-value is analyzed over the 28 subjects to determine the trend in variations over the entire subject population and measurement sites. R-value is subject-specific and is also heavily dependent on the measurement site. The analysis carried out shows that the usage of a generic R-value based calibration model for computing SpO2 using reflectance PPG signals from various measurement sites is an erroneous approach.