While it’s well known that darker skin tone affects the accuracy of pulse oximeter readings¹, it’s not known how far off those readings are from concurrent blood oxygen values or how often that discrepancy results in inadequate treatment.

To better understand the incidence and potential implications of inaccurate pulse oximeter oxygen saturation readings, we studied 13,483 patients hospitalized between January 1, 2016, and November 9, 2023, who had a documented pulse oximeter reading (SpO2) followed by a blood gas oxygen saturation result (SaO2) within five minutes. We stratified patients by race and compared the rates of inconsistent SpO2 and SaO2 readings.

The FDA currently states that pulse oximeters are considered meeting accuracy requirements if SpO2 readings are within 2-4% of SaO2 readings.² We found that 24.7% of non-Hispanic Black patients had SpO2 readings that were at least five percentage points greater than their SaO2 readings compared to 19.0% of non-Hispanic White patients. Additionally, 10.2% of non-Hispanic Black patients had SpO2 readings at least 15 percentage points greater than their SaO2 readings compared to 7.3% of non-Hispanic White patients.

One of the primary concerns of inaccurate SpO2 readings is the potential for a delay in treatment. Providers often order supplemental oxygen or other treatment when oxygen saturation drops below 88%. If an SpO2 reading shows a value of 92% or greater but does not accurately reflect the patient’s actual oxygen saturation, there may be a delay in potentially life-saving treatment, which could result in significant risk to the patient. For this study, we considered patients to have occult hypoxemia when the patient had an SaO2 value less than or equal to 88%, but an SpO2 value of 92% or greater. We compared the incidence of occult hypoxemia for non-Hispanic Black and White patients and found that Black patients are 31.9% more likely to experience occult hypoxemia than White patients, as shown in Figure 2.

These results support the need for further investigation of pulse oximeter accuracy to account for differences in patient skin pigmentation, race, and ethnicity as suggested by the FDA.¹

These data come from Cosmos, a collaboration of Epic health systems representing more than 238 million patient records from 1,345 hospitals and more than 28,000 clinics from all 50 states and Lebanon. This study was completed by two teams that worked independently and was completed in collaboration with researchers from the University of Maryland Medical System, University of Maryland School of Medicine, and Federation of American Scientists. The two teams came to similar conclusions. Graphics by Kayla Monnette. 

References

1. Shuren J. CDRH takes steps to advance further discussions on pulse oximeters. U.S. Food and Drug Administration. Published November 17, 2023.  https://www.fda.gov/medical-devices/medical-devices-news-and-events/cdrh-takes-steps-advance-further-discussions-pulse-oximeters. Accessed January 9, 2024.
2. Pulse oximetry. Yale Medicine. Published January 3, 2023.  https://www.yalemedicine.org/conditions/pulse-oximetry. Accessed January 9, 2024.