Abstract
Pulse oximeters are unique among patient monitors in that they cannot be calibrated nor can calibration be verified. The purpose of this study was to determine whether an inexpensive commercial device simulating a human finger could produce measurements of oxygen saturation (SpO2) within the error specifications supplied by pulse oximeter manufacturers. METHOD: Five brands of pulse oximeters were evaluated. We used the Nonin Finger Phantom to simulate SpO2 values of 80%, 90%, and 97%. Pulse rate was simulated by manually compressing the device at rates of 120, 84, and 60 beats/min as paced by a metronome. For each saturation level, 8 measurements (different probes) were made at each pulse rate (n = 24). Bias and imprecision of measurements were evaluated with t and x2 tests. Inaccuracy intervals were constructed to include 95% of future measurements at the 99% confidence level. RESULTS: Simulated saturations showed less imprecision but more bias than manufacturers’ specifications. Total measurement error (expressed as an inaccuracy interval) was within manufacturers’ specifications for all cases except for the Novametrix and the BCI at the 80% saturation level. Individual measurements may occasionally fall outside specified values by chance for some well-functioning oximeters at some saturation levels. CONCLUSIONS: Spot checks consisting of single measurements with the Finger Phantom are probably adequate for evaluating the performance of all devices studied. Spot checks using the mean values of repeated measurements could reduce the false-positive rate. [Respir Care 1996;41(2):100-104]
Get full access to this article
View all access options for this article.