Wheeze detection in the pediatric intensive care unit: comparison among physician, nurses, respiratory therapists, and a computerized respiratory sound monitor.

Objective: To correlate wheeze detection in the pediatric intensive care unit among staff members (a physician, nurses, and respiratory therapists [RTs]) and digital recordings from a computerized respiratory sound monitor (PulmoTrack).

Methods: We prospectively studied 11 patients in the pediatric intensive care unit. A physician, nurses, and RTs auscultated the patients and recorded their opinions about the presence of wheeze at baseline and then every hour for 6 hours. The clinician auscultated while the PulmoTrack recorded the lung sounds. The data were analyzed by a technician trained in interpretation of acoustic data and by a panel of experts blinded to the source of the recorded data, who scored all tracks for the presence or absence of wheeze. The degree of correlation among the expert panel, the staff, and the PulmoTrack was evaluated with the Kappa coefficient and McNemar's test. The determinations of the expert panel were taken as the true state (accepted standard).

Results: The PulmoTrack and expert panel were in agreement on detection of wheeze during inspiration, expiration, and the whole breath cycle; in all cases the Kappa coefficients were 0.54, 0.42, and 0.50 respectively. The PulmoTrack was significantly more sensitive than the physician (P = .002), nurses (P < .001), or RTs (P = .001). However, the specificity of the PulmoTrack was not significantly different from that of the physician, nurses, or RTs.

Conclusions: Between the physician, RTs, and nurses there was agreement about the presence of wheeze in critically ill patients in the pediatric intensive care unit. Compared to the objective acoustic measurements from the PulmoTrack, the intensive care unit staff was similar in their ability to detect the absence of wheeze. The PulmoTrack was better than the staff in detecting wheeze.