Pilot study of model-based estimation of inspiratory driving pressure in CPAP ventilation

Abstract Models of lung elastance, airway resistance, and patient work of breathing have been successfully applied to invasive mechanical ventilation data. Non-invasive mechanical ventilation data, including continuous positive airway pressure (CPAP), has presented challenges in predicting inspiratory driving pressure due to the combination of patient and device work. The model applied in this paper utilizes second order b-splines to describe inspiratory driving pressure. The model provided an accurate fit to the data, with an average root-mean-squared (RMS) error in model fit of 0.6 [0.425 – 0.675] cmH2O (median [lower quartile (LQ), upper quartile (UQ)]). Subject fit expiratory elastances were between 3.1 and 10.2 cmH2O/L and showed no correlation to set positive end-expiratory pressure (PEEP). Inspiratory driving pressure profiles approximated literature and work of breathing was shown to remain consistent between PEEP levels. Outlying data is hypothesized to be caused by subjects’ expiratory effort which was assumed negligible in the model. Further application of this model alongside validation data would provide more information on this and provide more evidence of model accuracy.