Accuracy of Displayed Airway Resistance in an ICU Ventilator: A Bench Study

Abstract

Background:

The Hamilton G5 Ventilator (Hamilton Medical, Switzerland) displays automated measures of respiratory mechanics such as static compliance (C_RS) and airway resistance (R_aw) using the least squares fitting method. The manual calculation for constant flow resistance is commonly recognized, while the manual calculation for decelerating flow is not readily known. This bench study was conducted to determine the reliability and accuracy of the displayed values for R_awwith a decelerating flow pattern.

Methods:

A bench test was conducted using the ASL 5000 Breathing Simulator (IngMar Medical, Pennsylvania) to compare the displayed calculations with manual calculations in passive conditions. Two clinical scenarios were created using compliance and resistance values (C_RS 50 mL/cm H₂O, R_aw 8 cm H₂O/L/s) and (C_RS 40 mL/cm H₂O, R_aw 15 cm H₂O/L/s). Two different flow patterns were assessed: constant and 50% decelerating. The tests were done using synchronized controlled mandatory ventilation, (S)CMV. Statistics were calculated in RStudio (Posit, MA) using the Bland-Altman analysis. The mean bias, limits of agreement (LoA), and concordance coefficient (CCC) are reported. The automated calculations of C_RS and R_aw were then compared based on the clinical scenarios.

Results:

The R_aw mean bias -7.26 (95% CI -8.01 to -6.52), LoA -9.83 and -4.70 (95% CI -10.67 to -8.98 and -5.54 to -3.86) with a CCC of 0.13 (95% CI 0.07 to 0.19) for difference between calculated and displayed values for decelerating flow were not favorable. Using the ASL 5000 as a reference value compared to displayed values of the Hamilton G5 yielded a bias of -0.73 (95% CI -0.90 to -0.55), LoA -1.31 and -0.16 (95% CI -1.59 to -1.12 and -0.35 to 0.03) with a CCC 0.97 (95% CI 0.94 to 0.98).

Conclusions:

The automated calculations from the Hamilton G5 ventilator based on the least squares fitting model were dependable for both C_RS and R_aw measurements. As there are no defined calculations for bedside calculation of R_aw for a decelerating flow pattern, using the manual calculation for the constant flow pattern was not accurate. Using the ASL 5000 values as the reference measurement, the Hamilton G5 R_aw was reliable and accurate for flow pattern irrespective of the type described by low bias, Bland-Altman LoA, and CCC. The LoA show that the values cannot be used interchangeably for a decelerating flow pattern. These were in passive conditions and from previous studies would not be as reliable with a patient actively breathing.