Accuracy of Smartphone Pulse Oximeters in Patients Visiting an Outpatient Pulmonary Function Lab for a 6-Minute Walk Test

Abstract

Background: Pulse oximeters are commonly utilized medical devices used to measure blood oxygen saturation (SpO₂). These devices are either stand-alone or integrated into physiologic monitoring systems, using 2 wavelengths of lights to determine SpO₂. With recent advances in technology, some smartphones now have pulse oximetry capabilities. Since it is possible that some patients could utilize this technology, we sought to assess the accuracy, reliability and usability of smartphone pulse oximeters. Methods: This was an IRB-approved prospective, observational study which involved noninvasive measurements SpO₂ by three devices: 1) Masimo Radical-7 pulse oximeter, 2) Kenek Edge Pulse Oximeter with the Apple iPhone 6S, and 3) Samsung S8 smartphone. Ambulatory patients above 18 years of age visiting the pulmonary function lab for a 6-minute walk test were eligible to participate in the study. Each participant’s SpO₂ was measured using all three devices, simultaneously. Pre-test and post-test results were obtained for every participant. All results were calculated by Pearson correlation coefficient and Bland-Altman plots were used to compare smartphone values to the pulse oximeter used in the lab. When devices failed to provide a SpO₂ result, it was recorded as Unable to Obtain (UTO). Results: 48 participants enrolled in the study. Of the 48 participants, 77.1% were females and 22.9% were males. For SpO₂, there was a moderate, positive correlation between pre-test Masimo and Apple devices (n = 47, r = .519, P < .001) and Masimo and Samsung (n = 35, r = .552, P < .001). There was a moderate, positive correlation between post-test Masimo and Apple devices (n = 44, r = .675, P = < .001) and Masimo and Samsung (n = 31, r = .672, P = < .001). UTO pre-test SpO₂ results: Samsung: 13/48 (27%), Apple: 1/48 (2%); UTO in post-test SpO₂ results: Samsung: 17/48 (35%), Apple: 4/48 (8%). The Masimo device was able to obtain a SpO₂ result in each study participant (50/50, 100%). Pre- and post-test Bland-Altman plots show inconsistent agreement between devices. Conclusions: Smartphone pulse oximetry technology is inconsistent when compared to a device designed to measure oxygen saturation. More research is needed to assess differences in smartphone pulse oximetry applications and to better understand if this technology can be utilized in patients with varying degrees of hypoxemia.