Overview and Recommended Practice for Adjustment
In the PneuView system, “breath parsing” is the term used to encompass the time and frequency domain analysis techniques used to identify the various phases of the breath (e.g., baseline, inspiration, inspiratory hold, expiration). As with any waveform analysis system, there are tradeoffs between sensitivity to detail events and stability in identifying larger features. The Breath Parsing dialog box accessible from within the PneuView system using the key combination was originally designed for only internal use as we ran experiments to determine an optimal compromise between sensitivity and stability. It became clear, however, that there was the opportunity to have it both ways – keep sensitivity high and minimize the amount of filtering needed on the data stream, but allow the user to adjust sensitivity in special cases or for particular applications. Thus, the Breath Parsing dialog remains a ‘feature’, albeit a largely undocumented one, of the program.
The first thing to understand is that adjustments made to the breath parsing settings can affect the values calculated for the Measured Parameters during a test. This is not necessarily a bad thing, however, and long as the adjustments are appropriate for the type of real world waveform undergoing analysis. The settings determine when the system will start to recognize that the breath pattern may be entering a new phase. The three different rules sets essentially cast ‘votes’ for the breath phase each thinks the breath pattern is in. A vote occurs every 10ms.
For example, the “Flow Threshold” settings determine the level, in increments of 0.1LPM, the system considers either positive enough of negative enough to indicate the breath pattern is entering the inspiratory phase or expiratory phase, respectively. This means that the instantaneous flow rate calculated must achieve at least that threshold before the Flow Rule will cast more than half its votes for a change of state. For flow waveforms exhibiting a fast rise time (e.g., ‘square’ and ‘sinusoidal’ flow waveforms), these settings may be adjusted to large magnitudes without adversely affecting the calculation of Measured Parameters. Raising the Inspiratory Flow Threshold can help eliminate ‘false triggers’ caused by the lung top plate coming down hard against its rest in low airway resistance, low compliance conditions with no significant baseline pressure.
The Differential Pressure and Volume rules function similarly, except that the former looks at the difference in pressure across the simulated airway and the latter looks for a change in physical lung volume. At high airway resistances the Differential Pressure threshold settings may be adjusted quite high without adversely affecting the calculation of Measured Parameters. Such adjustments may be useful if the flow waveform is spiky or erratic because of ‘noisy’ valve actions or other mechanical or acoustic interference in the system being tested.
In cases where the breath is very shallow and the flow waveform is not steep, it can be helpful to turn the Flow Rules OFF and rely primarily on the Volume Rules for breath parsing. This can be especially important when airway resistance is low (e.g., Rp5), in which case it might help to turn the Differential Pressure Rules OFF as well.
It should be added that, unless you explicitly turn a rule OFF, it might cast votes for a certain breath state even if a corresponding threshold is not completely met. For example, if the Inspiratory Flow Threshold is set to 10 (1.0L/min) then it will start casting some of its votes for entering the Inspiratory phase when the calculated instantaneous flow rate reaches +0.5L/min. This way, even if no other rule picks up the slight but present beginnings of inspiration, the system can recognize the start of the Inspiratory phase more quickly than it otherwise might if it were required to rise high enough to exceed the threshold absolutely.
The best way to select breath parsing settings is to begin by considering the forgoing discussion to select which rule(s) should be adjusted. Make adjustments incrementally until the breath is being satisfactorily parsed. To ensure that the changes are not adversely affecting the calculation of Measure Parameters, note the measurements (or save a sample in a Data Table) and continue incrementing the settings in the same direction until you see a change in measurements. As long as the settings can be changed two or more clicks without significantly affecting the measurements, one can be reasonably assured that the breath parsing settings are valid and are not causing meaningful information to be filtered from the data stream. Having done this, system sensitivity and stability have also been optimized for the current test conditions.