Uncertainty Estimation of Biomedical Oximeters: Proposal for an Experimental Test-Bench

Abstract

Paramagnetic oximeters are frequently used in clinical practice to measure and titrate the fraction of inspired oxygen (FiO2) administered to a hypoxic patient. Newborns are prone to several diseases leading to hypoxia but also to oxidative damage, therefore a precise measurement of administered oxygen is of paramount importance. In this study, we evaluated and compared the performance of four commonly used gas analyzers at a Neonatal Unit. A paramagnetic gas analyzer was considered the standard for reference. Gas mixing performed by blenders was also evaluated. We first assessed the stability and composition of O2 and medical air. A settling time study was conducted to estimate a steady-state. Results show how tested blenders underestimated the real measures when FiO2 was <50% and overestimated them when FiO2 ≥50%. In all the tested oximeters the FiO2 was underestimated by 1.18%, but the overall agreement with the gold standard was very good.