Comparative provocation test of respiratory monitoring methods

Objective. The aim of this study was to compare clinically relevant performance of: 1) a prototype respiratory sensor based on capnometry with two alternative signal receptor fixations, 2) a fiberoptic humidity sensor and 3) human visual observation. Comparative provocation tests were performed on volunteers at the Post-Anesthesia Care Unit at Västerås Central Hospital. Methods. The experimental tests involved 10 healthy, voluntary test subjects, instructed to intersperse normal breathing with protocol provocations of breath holding, limb and head movements, and nasal oxygen supplement. The signal outputs from the three respiratory monitoring methods were recorded on a personal computer. The signal analysis included visual categorising of the signals and counting breath events. Recognising that none of the methods could act as reference, events were classified as "unanimous," "majority" or "minority" events depending on how many of the three methods that detected a breath. Results. The average total recording time was 37 minutes per subject. The respiratory rates varied from 6.5 to 19 breaths per minute, with a mean value of 11.4 breaths/minute. The breath hold duration ranged from 18 to 50 seconds. Discrepancies between the three methods were found in more than 20% of the marked events. The most frequent majority events were due to events not recorded by the observer who, on the other hand, contributed the least to minority events. The provocations made by the subjects during the measurement did not increase the rates of majority and minority events, compared to periods of no provocation. The fiberoptic device exhibited a large count of minority events but a smaller contribution to majority events than the capnometry prototype. Conclusions. The capnometry and fiberoptic sensors exhibit differences in responses that may be understood from basic principles. The importance of the physical application of the sensor to the patient was clearly observed. The optimum design remains to be found.