Format

Send to

Choose Destination
Conf Proc IEEE Eng Med Biol Soc. 2013;2013:4549-52. doi: 10.1109/EMBC.2013.6610559.

Statistical analysis of tracheal breath sounds during wakefulness for screening obstructive sleep apnea.

Abstract

Obstructive sleep apnea (OSA) is a prevalent disorder. The accepted method of diagnosis in widespread clinical practice, polysomnography (PSG), is costly and very time consuming; therefore, quick screening methods, especially when there is a need for quick diagnosis, is of great interest. Diagnostic methods which exploit subtle differences in breath sounds recorded during wakefulness, such as our group's Awake-OSA technology, have shown their capability to diagnose OSA at the research stage. Simplifying the breath sound recording procedure employed in the Awake-OSA diagnostic method would increase its efficiency when used in a clinical setting. In this study, we adopted breath sound data collected during wakefulness in two positions (sitting upright and supine) and two breathing maneuvers (nose and mouth breathing) from our previous study, and ran hypothesis tests on a wide variety of sound features to select the most significant features correlated with OSA. The goal was to investigate which combinations of patient position and breathing maneuver contribute the least to the significant features amongst groups of people with differing OSA severity, thus permitting simplification of the recording protocol. The results show that all signals recorded by a combination of the two breathing maneuvers and two positions result in features significantly correlated with OSA severity; this makes it impossible to confidently recommend that a combination be omitted from the recording protocol. Nevertheless, the results show that the majority of significant features originated from recordings made in the supine position. Therefore, as a step toward simplification of the Awake-OSA diagnostic algorithm, we may use breath sound signals recorded only in the supine position and further investigate the accuracy of the algorithm in distinguishing amongst groups with differing OSA severity.

PMID:
24110746
DOI:
10.1109/EMBC.2013.6610559
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for IEEE Engineering in Medicine and Biology Society
Loading ...
Support Center