Previous Abstract | Next Abstract
Printable Version
October 23, 2022
10/23/2022 3:45:00 PM - 10/23/2022 4:45:00 PM
Room Virtual - Mobile App
The Accuracy Of 12 Low-cost Finger Pulse Oximeters
Isabella Auchus, B.A., Koa Gudelunas, B.S., John R. Feiner, M.D., Philip E. Bickler, M.D.,Ph.D., Michael Bernstein, B.S., Zainab Choudhery, Student, Sky Vanderburg, M.D.,M.P.H., Michael S. Lipnick, M.D.
University of California, San Francisco, San Francisco, California, United States
Introduction: Inexpensive pulse oximeters are widely available and commonly used to make clinical decisions, especially in resource-variable and homecare settings. However, the accuracy of these devices is known to be variable and has not been fully characterized.
Methods: We evaluated the accuracy of 12 low-cost ([ENTamp;#60;$30 per unit) finger pulse oximeters during stable arterial oxygen saturations (SaO2) between 70% and 100% in healthy subjects with a range of skin tones and perfusion indices (PI). Perfusion index was measured using Masimo Radical 7 Pulse Oximeter. Skin pigment was determined according to the Fitzpatrick scale. EighteenOximeters tested were the Bodymed BDMOXMTRBLK, Contec OLED CMS50NA, FaceLake FL-400, LPOW Pulse Oximeter A340, McKesson Medical MD300C21C, Medline Fingertip Pulse Oximeter, Roscoe POX-ROS, Santa Medical SM-519BR-BL, Santa Medical SM-1100S, SureLife FS10D, Walgreens Oxywatch C20, and Zacurate 500C. We monitored and adjusted inspired oxygen, nitrogen, and carbon dioxide partial pressures via a partial rebreathing circuit to achieve 10 to 12 stable SaO2 plateaus between 70% and 100% and PaCO2 levels of 35 to 45 mm Hg. We compared pulse oximeter readings (SpO2) with arterial SaO2 (by Radiometer ABL90) to calculate bias (SpO2 - SaO2) mean, precision (SD of the bias), and root mean square error (ARMS).
Results: Pulse oximeter readings from 59 subjects, corresponding to 2859 blood samples were analyzed. Eighteen subjects had Fitzpatrick skin type of 5 or 6, and 1915 data points had a PI>2, and 943 had a PI< 2. Seven of the 12 oximeters tested showed significant errors (up to -3.15% mean bias, precision 5.02%, 5.43 ARMS) in estimating saturation when SaO2 was between 70% and 100%. When SaO2 was reduced to<80%, 8 oximeters demonstrated large errors, and 7 of the oximeters demonstrated large errors when estimating saturations between 80% and 90%. Five of the pulse oximeters tested (McKesson Medical MD300C21C, Medline Fingertip Pulse Oximeter, Roscoe POX-ROS, Santa Medical SM-519BR-BL, and Santa Medical SM-1100S) demonstrated ARMS of≤3% at SaO2 between 70% and 100%, which meets performance criteria established by the FDA.
Conclusions: Similar to prior studies, we found a wide range of performance for low-cost pulse oximeters with only some meeting current FDA requirements. Some of the low-cost pulse oximeters tested demonstrated inaccuracies that are likely to impact clinical care decisions. However, the accuracy of some low-cost pulse oximeters tested here performed similarly to more expensive, FDA-cleared units when measuring hypoxia in healthy subjects. Donors, aid recipients, healthcare providers, and patients should be aware of the limitations of some inexpensive fingertip oximeters. Additionally, good performance in the laboratory setting does not guarantee acceptable or safe performance in the clinical setting. Additional testing is needed to fully characterize the impact of perfusion and skin pigment on oximeter performance to further assess the accuracy of these low-cost oximeters.

Copyright © 2018 American Society of Anesthesiologists