A prototype device for standardized calibration of pulse oximeters II.

Objective: There is no commonly accepted in vivo calibration method for pulse oximeters available up to now. On the basis of a prototype device for the calibration of pulse oximeters which was introduced recently, a second approach based on the same concept was tackled in order to design a reliable method for standardized calibration of pulse oximeters.

Methods: An extensive clinical database of time-resolved optical transmission spectra of patient fingers is used to simulate the behavior of patients. A device which is capable of playing back these spectroscopic data to pulse oximeters, and a database where the oxygen status measured with the reference method (Co-Oximetry) is stored, are the main parts of the concept. The playback device has an artificial finger as interface to the pulse oximeters and serves to collect light from the pulse oximeter for analysis and to playback simulated light to the pulse oximeter. The light intensity emitted by two LEDs which illuminates the pulse oximeter detector is controlled via a computer in such a way that it is the same as if the pulse oximeter light had passed the finger. The pulse oximeter display during the data playback can thus be compared to the true SaO2 of the patient. The device is tested with 4 pulse oximeters based on 100 patient spectra.

Results: For the four pulse oximeters used in this investigation, an Agilent Technologies CMS monitor (formerly Hewlett-Packard), an Ivy 2000 with Masimo Set technology and Nellcor N-3000 and N-395, there is good correlation between SPO2 and SaO2, and mean and standard deviation of in vivo SpO2-SaO2 and playback SpO2-SaO2 are in good agreement. For two instruments, Nellcor N3000 and Agilent CMS Monitor, a quantitative comparison between the in vivo and in vitro SpO, results was derived. A mean of the deviation playback vs. in vivo SpO2 is less than 0.5% SpO2. The error limits are comparable with the calibration error of the conventional calibration routine. The device is also capable of data playback even in situations with rapid desaturation changes, as displayed in Figure 2. For the other tested pulse oximeters the results are comparable.

Conclusions: Compared to the first prototype the current version is simpler and less expensive in production. Many of previously existing problems are solved and the applicability to a large variety of pulse oximeters and sensors is given. The novel concept for the calibration of pulse oximeters is a tool for assessing the performance of pulse oximeters.