Uncertainty evaluation of a calibration method for metabolic analyzer in mechanical ventilation

A calibration methodology for metabolic analyzer to be used with mechanically ventilated patients is described. The associated uncertainty is evaluated by means of both standard and Monte Carlo methods, the effects of pulmonary ventilator bias flow are considered. An experimental set-up composed of a burning unit, for the combustion of ethanol, a mechanical ventilator and a test lung has been used to generate respiratory patterns, with known CO₂ production and respiratory quotient. Trials with and without bias flow at different oxygen fractions and positive end expiratory pressure levels have been carried out with two mechanical ventilators. An average uncertainty of 7% has been estimated for the CO₂ volume produced, whereas an average uncertainty of about 4% for the respiratory quotient has been obtained. Furthermore, the results obtained by application of Monte Carlo method are in agreement with standard method of uncertainty evaluation.