Rejection of Erroneous Saturation Data in Optical Pulse Oximetry in Newborn Patients

Pulse oximetry (PO) is extensively used in intensive care unit (ICU); this is mainly due to the fact that it is a non‐invasive and real‐time monitoring method. PO allows to measure arterial oxygen saturation $(SaO2)$ and in particular hemoglobin oxygenation. Optical PO is typically realized by the use of a clip (to be applied on the ear or on the finger top) containing a couple of monochromatic LED sources and a photodiode. The main drawback with the use of PO is the presence of movement artifacts or disturbance due to optical sources and skin, causing erroneous saturation data. The aim of this work is to present the measurement procedure based on a specially developed algorithm able to reject erroneous oxygen saturation data during long lasting monitoring of patients in ICU and to compare measurement data with reference data provided by EGA.

We have collected $SaO2$ data from a standard PO and used an intensive care unit monitor to collect data. This device was connected to our acquisition system and heart rate (HR) and $SaO2$ data were acquired and processed by our specially developed algorithm and directly reproduced on the PC screen for use by the clinicians. The algorithm here used for the individuation and rejection of erroneous saturation data is based on the assessment of the difference between the Heart Rate (HR) measured by respectively by the ECG and PO. We have used an emogasanalyzer (EGA) for comparison of the measured data. The study was carried out in a neonatal intensive care unit (NICU), using 817 data coming from 24 patients and the observation time was of about 10000 hours.

Results show a reduction in the maximum difference between the $SaO2$ data measured, simultaneously, on the same patient by the EGA and by the proposed method of 14.20% and of the 4.76% in average over the 817 samples. The measurement method proposed is therefore able to individuate and eliminate the erroneous saturation data due to motion artifacts and reported by the pulse oxymeter. Specifically in neonatal ICU, it allows to come to more efficient individuation of apnea, hypoxia and hyperoxia events and consequently to operate more adequate and efficient therapeutic actions.