Within-breath analysis of respiratory mechanics – resistance (Rrs) and reactance (Xrs) – provides important information of the health of the respiratory system, but little work has been done estimating the accuracy of tracking the temporal changes in impedance. The accuracy of current oscillometry techniques can only be assessed via computational models.
Here we modeled the respiratory system as a single-compartment lung model with time-varying Rrs and Xrs for obstructed asthma according to measurements of mechanics and breathing noise from 7 children (4m/3f, ages 7 to 12) with asthma specifically chosen to span a range of breathing frequencies (0.25 to 0.5 Hz) using an airwave oscillometry device (tremoFlo™). We also extended the theory to include the temporal changes in the model, which helps compute an estimate of the temporal tracking error (TTE).
Accuracy for mean Rrs and Xrs exceeded 99% for all conditions. TTE increased from 3.1 ± 1.2 % to 9.9 ± 2.3% with increasing breathing rate up to 0.43 Hz independent from noise amplitude, and only exceeded 10% of the mean Rrs, at the highest breathing frequency (15.3% ± 3.1% at 0.5 Hz). Coherence remained at 0.92 ± 0.02 at all frequencies.
Results indicate that coherence is not a good measure of data quality for within-breath tracking of impedance. Moreover, while mean breathing frequency can increase errors, this does not greatly affect mean Rrs, thus the results further recommend training subjects to reduce breathing rate to accurately track temporal variation in respiratory mechanics.