Racial bias is automated into various technologies such as facial recognition.¹ Racial bias is also ingrained in our health care system. However, the intersection of medical technology, health care, and racial bias is often overlooked. Lately, pulse oximetry has become an especially important example of that. This technology has played a big role during the COVID-19 pandemic—which is why we need to understand its strengths and weaknesses. 

Few studies have evaluated the differences in the accuracy of pulse oximetry readings depending on race. Recently published research in the New England Journal of Medicine demonstrates how commonly the administration of supplemental oxygen depends on readings supplied by the pulse oximeter.² In the study, researchers evaluated adults who received supplemental oxygen. Pulse oximetry measurements were paired with arterial oxygen saturations collected within 10 minutes of each other. The study population included inpatients from the University of Michigan Hospital from January to July 2020 and patients in intensive care units in 178 hospitals from 2014 to 2015. The study evaluated occult hypoxemia in patients who self-identified their races as either white or Black. Adjustments were made based on age, sex, and cardiovascular health using the Sequential Organ Failure Assessment (SOFA). From the University of Michigan site, 10,789 paired measurements were obtained from 276 Black patients and 1,333 white patients, and from the multicenter sites, 37,308 paired measurements were obtained from 1,050 Black patients and 7,352 white patients. The findings were impressive. “Black patients had nearly three times the frequency of occult hypoxemia that was not detected by pulse oximetry as white patients,” the authors write. 

An older study published in 2007 in Anesthesia and Analgesia evaluated a smaller study population of 36 individuals and classified their skin tones as either light, intermediate, or dark.³ That study noted that pulse oximeters were more likely to overestimate oxygen saturation in hypoxic patients with darker skin. The findings also varied by pulse oximeter type, with adhesive sensors introducing greater bias when compared to clip-on sensors.

COVID-19 brings urgency to highlighting these differences. As we all know, a great deal of triaging occurs based on patients’ initial vital signs as well as those recorded during evaluation and treatment in the emergency department. During the pandemic, peripheral oxygen saturation levels have arguably become the most important vital sign. In fact, whether patients should receive dexamethasone, which is the only therapy with a proven mortality benefit in treating COVID-19, hinges on oxygen levels; hypoxic patients should receive dexamethasone while normoxic patients should not. (There was a signal of harm in the RECOVERY trial among normoxic patients who received the steroid.)⁴

The reliance on pulse oximetry to triage Black patients in particular places these patients at an increased risk for delayed interventions and therefore an increased risk for hypoxemia and its downstream effects. Special attention should be paid to our discharge recommendations for patients with a known or suspected COVID-19 diagnosis. As reported by Dr. Levitan and Dr. Chang here, there are benefits to having patients monitor their oxygen saturation with at-home pulse oximeters.⁵ However, efforts should be made to educate patients, especially those with darker skin tones, about the inaccuracy of these measurements and to place more emphasis on their symptoms rather than the measurements from these devices. Further study may help discover whether the threshold for what is considered a “hypoxic” reading should differ among people depending on their skin tone.

Dr. Otugo is a health policy research and translation fellow at Brigham and Women’s Hospital in Boston.

References

1. Harwell D. Federal study confirms racial bias of many facial-recognition systems, casts doubt on their expanding use. Washington Post website. Accessed Jan. 28, 2021.
2. Sjoding MW, Dickson RP, Iwashyna TJ, et al. Racial bias in pulse oximetry measurement. N Engl J Med. 2020;383(25):2477-2478.
3. Feiner JR, Severinghaus JW, Bickler PE. Dark skin decreases the accuracy of pulse oximeters at low oxygen saturation: the effects of oximeter probe type and gender. Anesth Analg. 2007;105(6 Suppl):S18-23.
4. RECOVERY Collaborative Group; Horby P, Lim WS, Emberson JR, et al. Dexamethasone in hospitalized patients with Covid-19 – preliminary report [published online ahead of print July 17, 2020]. N Engl J Med. doi: 10.1056/NEJMoa2021436.
5. TorjesenI. Covid-19: patients to use pulse oximetry at home to spot deterioration. BMJ. 2020;371:m4151.