Modeling of FiO2 delivered by self-inflating ventilation device: Modelagem de FiO2 entregue por bolsa auto inflável

During the COVID-19 pandemic, professionals in the medical field have been using bag resuscitators, also called bag valve masks, to provide respiratory support for patients for long periods. A relevant issue arising because of such use is the little information regarding the FiO2 delivered by these devices, leading to toxic effects of O2 and its waste. This paper presents a model that relates volume, respiratory frequency, and flow in an O2 line with the FiO2 delivered by bag resuscitators. The model was obtained based on a formal procedure and data from a 1500 ml Hsiner bag resuscitator collected in the literature, both presented in this paper. Then, we compared through Bland-Altman analysis the FiO2 values estimated by our model with the ones from data considering different bag resuscitators to evaluate the amplitude of the errors it can generate. The average error for Hsiner and Laerdal bag resuscitators was less than 0.035 (or 3.5%) and less than 0.063 (or 6.3%) for Ambu devices. The maximum error found for Ambu bag resuscitators was high. The obtained model presented satisfactory errors, compatible with the uncertainty levels associated with FiO2 measuring instruments and blenders, meaning the model can be used to estimate the delivered FiO2 accurately. Additionally, the procedure presented in this paper can be used to obtain specific models for bag resuscitators of other companies, equipping professionals with tools that can accurately estimate FiO2.