Impact of Two Lung Elastance Identification Methods on Pulmonary Mechanics Prediction

Abstract Positive-end-expiratory-pressure (PEEP) have proved effective in recruiting lung volume and keeping alveoli open. However, there is no standard means to find an optimal patient-specific PEEP, creating variability care outcomes. There thus a need for personalized approaches the best PEEP optimise care. This research extends well-validated virtual patient model with newly proposed function predict distension, while impact on outcome of two different elastance identification strategies are discussed compared. A prior studied exponential basis set used as general model, identified using overlapped separate methods, respectively. In this approach, distension yields absolute median peak inspiratory pressure (PIP) prediction error 1.50cmH2O 623 cases. Comparison between clinically measurement PIP R2=0.90 across predictions total, R2=0.87 identification. Furthermore, both methods improvement compared without (R2=0.82). Validation fulfilled 18 controlled ventilation patients respiratory data at 7 baseline levels (0-12cmH2O) maximal interval 12cmH2O. Overall, results demonstrate well potential significance accurately capturing mechanics, which providing guidance clinical insights predictive mechanics modelling.