PULSE OXIMETRY; A PRIMER ON LIMITATIONS

The pulse oximeter has revolutionized clinical care enabling inexpensive, continuous, noninvasive oxygenation measurements, facilitating prompt detection of desaturation before visible cyanosis, while avoiding frequent phlebotomy (1–4). These devices have become standard care, ubiquitous in clinical arenas caring for critically ill children. Essential roles span diagnosis (including neonatal congenital heart disease screening), illness severity definition, triage for admission criteria, and guiding respiratory interventions. Contemporary pediatric acute respiratory distress syndrome definitions include saturation measurements as part of a paradigm shift away from invasive arterial monitoring (5). With expansion into homes, beyond clinical utility, Medicare reimbursements rely on saturation thresholds (6). As pulse oximeters have become ingrained in clinical practice, results are highly valued as “objective” and taken at face value. An understanding of the basic principles behind mechanism of function is essential to appreciate the true scope and associated limitations relating to perfusion, pigmentation, and external validity (1,2).

Pulse oximetry (Spo2) relies on differential absorption of red and infrared light by oxyhemoglobin and deoxyhemoglobin. A pulse oximeter determines only arterial Spo2 by measuring changes in absorbance over time, using cardiac-cycle fluctuations in light absorbance and subtracting constant, nonpulsatile components (1,2). Factors affecting light absorption such as melanin and hemoglobin type affect accuracy (1–3). Specifically, melanin is a strong absorber of light in the near-infrared range. While this premise underlies potential for inaccuracies in people of color, skin pigmentation and other surface light absorbers theoretically should not cause erroneous values as pigments technically absorb a constant fraction of incident light (7). Pulse oximeter microprocessors, however, also require empirical calibration (1,2). Despite broad applications in modern medicine, the oximetry instrument was originally designed for altitude travel, with results standardized against a calibration curve generated for light skinned individuals (6,8). Calibration based on unrepresentative cohorts that grossly under-sampled populations with diverse skin tones dramatically limits generalizability (7).

OXIMETRY BIAS AND RACE

Adult data from as early as the 1990s suggested that skin pigmentation was associated with both oximetry bias (discrepancy between Spo2 values for a given arterial oxygen saturation [Sao2]) and variance (wider range of Spo2 for a given Sao2) (9–12). These inaccuracies largely resulted in overestimation of Sao2 in patients with diverse racial and/or ethnic backgrounds, especially at lower saturations (3). Oximetry bias and occult hypoxemia according to race and/or ethnicity (unrecognized desaturation with Spo2 measured Spo2 > 92% while Sao2 < 88%) has been extensively identified across populations, including in preterm infants where oxygen titration within a narrow saturation range is crucial (13–24). In contrast, only one pediatric publication observed no statistical difference in accuracy by skin pigmentation (25). While limited data evaluates clinical implications of occult hypoxemia in children, all published adult studies observed associations with adverse outcomes, including receipt of less supplemental oxygen, delayed COVID-19 therapies or ICU transfer, and higher organ dysfunction as well as mortality (22,26–30).

In this issue of Pediatric Critical Care Medicine, Savorgnan et al (31) share results of their single center retrospective study of 2,713 children with COVID-19. Compared with White patients, oximetry bias was more common in Black patients, especially at lower oxygen saturations. Occult hypoxemia was three times more common in Black patients. No relationship was observed between oximetry bias or occult hypoxemia and hospital length of stay, adjusting for ICU support. Strengths include the large sample size and high proportion of Black patients (61.3%). While one of few pediatric studies addressing this question, there are numerous limitations. Race is self-reported in two categories rather than a spectrum of skin pigmentation. Spo2–Sao2 value pairs compute average Spo2 signal over the 10-minute interval preceding Sao2 measurement with no clarity around handling of low-quality measurements. Outcomes focus on length of stay (vs respiratory support). Patients on mechanical circulatory support were not excluded and limited clinical covariates were collected (including no vasoactive doses), making malperfusion a potential unmeasured confounder. Importantly, although a multivariate analysis examined the association between race and oximetry bias, included covariates were not provided, and the analyses of oximetry bias and outcomes were adjusted for race. Despite limitations constraining conclusions from retrospective studies about hazards of device interpretation, the signal demands attention.

IMPLICATIONS

The work by Savorgnan et al (31) adds to a growing body of literature highlighting unconscious racial bias in the design and implementation of Spo2 in diverse patients. With overwhelming data describing inequitable outcomes according to race and/or ethnicity, researchers and clinicians have been urged to comprehensively scrutinize all aspects of healthcare for contribution to disparities, and strive to address them. Structural racism is pervasive in medicine. We are learning that about many things we rely on. Measuring devices such as pulse oximeters and randomized control trials are no exception. Underrepresented races and/or ethnicities have been systematically excluded from clinical trials and testing processes have lacked diversity (32). For devices shaped by “discriminatory design,” even unintentional inequalities “produce patterned exclusions and inequitable outcomes” (6,33). Our heavy reliance on a measurement instrument that may promote inequitable allocation of resources or interventions demands a critical relook to better understand associations between skin pigmentation and oximeter accuracy, its role in disparate health outcomes according to race and/or ethnicity, and to consider solutions.

FUTURE DIRECTIONS

While no simple solutions exist, efforts can be directed toward design and calibration stages, regulatory approval, and bedside interpretation to acknowledge and mitigate overestimation of oxygen saturation in patients with diverse skin colors and ameliorate impacts on outcomes.

Design and Calibration

Pulse oximeter technology developers should focus attention on developing accurate devices across all skin tones. Areas under research include capability to automatically recognize skin pigmentation or built-in user-optional adjustments that allow operator selection (34,35). Such advances could permit setting individualized baselines for each wearer, enhancing accuracy and safety. While innovative designs may be “just around the corner,” this will require time to validate. In the meantime, it is imperative that devices are tested in diverse populations and that results are calibrated to a range of skin tones.

Stricter Regulatory Requirements

Regulatory agencies should spearhead efforts to understand how structural racism impacts development and use of medical devices (4). In the United Kingdom, an independent review is underway to evaluate potential bias in medical devices and its impact on patients from different ethnic groups (36). Similarly, in response to a congressional inquiry, the U.S. Food and Drug Administration (FDA) convened a meeting to discuss concerns about pulse oximeter inaccuracies with recommendations expected imminently (37,38). Current FDA regulations require that clinical study samples for pulse oximeters use a demographically representative sample of the U.S. population, including at least two darkly pigmented participants or 15% of the participant pool, whichever is larger (39). This established threshold seems surprisingly low in the context of a diverse U.S. population and evidence suggesting disparities in oximeter performance under the current regulations. Regulatory bodies must demand premarket testing be performed with a sample size large enough to reliably detect statistical differences in accuracy between demographic groups, and inclusion of a higher percentage of patients with diverse skin tones.

The Clinical Setting—Expanding Precision Medicine Beyond “Omics”

Data observing inaccurate oximetry readings led to proposals for higher Spo2 targets in diverse populations to safeguard against desaturation (11). Race is a social construct, so race-based clinical algorithms are not recommended as they perpetuate disparate outcomes for underrepresented races and/or ethnicities, and reinforce dangerous notions of biologic determinism (40–42). While “race” inadequately groups individuals with different lived experiences into socially derived categories, a patient’s skin pigmentation is a unique signature. A precision medicine paradigm is emerging with personalized therapies such as drugs tailored to pharmacogenomic variants, and chemotherapy targeted at the tumor-specific molecular level. Could skin tone be considered when designing algorithms utilizing Spo2 for triaging patients and guiding clinical management? Baseline uncertainty (an Spo2 reading may be ±4% for the Sao2) means the use of a threshold Spo2 in treatment algorithms is limited. While establishing different Spo2 thresholds according to skin pigmentation could lead to oversimplification, acknowledging limitations of Spo2 in clinical pathways may be a step in the right direction. At minimum, these data suggests that bedside providers and trainees should understand margins of error and carry an elevated index of suspicion of hypoxemia in darker skinned individuals when making decisions around respiratory support and care escalation (4,43).

Researchers

Research is vital in better understanding disparities associated with Spo2 inaccuracies. Funding agencies should encourage race-conscious research addressing medical device biases and study clinically meaningful outcomes (40,42). Publicly available databases used for device testing and outcome evaluation should achieve balanced representation of different populations, and the role of disaggregation for calibration needs to be thoughtfully considered (44). Finally, race categories are inappropriate when what we are measuring is a color-tone scale rather than a social construct. In this research, investigators should endeavor to use objective measures of skin pigmentation over subjective definitions of race and ethnicity (45).

CONCLUSIONS

Despite being considered an objective measure, robust to implicit biases, pulse oximeters add to a mounting list of chronic racial biases in medicine that may inadvertently promote inequitable outcomes. As structural racism is pervasive, we must seek to critically evaluate all factors that potentially contribute to healthcare disparities. The simple pulse oximeter has numerous limitations. A multitiered approach to improving device accuracy is crucial to advancing healthcare for all but is especially important in improving care of patients who already experience significant health disparities due to their race and/or ethnicity.

REFERENCES 1. Chan ED, Chan MM, Chan MM: Pulse oximetry: Understanding its basic principles facilitates appreciation of its limitations. Respir Med. 2013; 107:789–799 2. Patterson S, Sandercock N, Verhovsek M: Understanding pulse oximetry in hematology patients: Hemoglobinopathies, racial differences, and beyond. Am J Hematol. 2022; 97:1659–1663 3. Shi C, Goodall M, Dumville J, et al.: The accuracy of pulse oximetry in measuring oxygen saturation by levels of skin pigmentation: A systematic review and meta-analysis. BMC Med. 2022; 20:267 4. Jamali H, Castillo LT, Morgan CC, et al.: Racial disparity in oxygen saturation measurements by pulse oximetry: Evidence and implications. Ann Am Thorac Soc. 2022; 19:1951–1964 5. Pediatric Acute Lung Injury Consensus Conference Group: Pediatric acute respiratory distress syndrome: Consensus recommendations from the Pediatric Acute Lung Injury Consensus Conference. Pediatr Crit Care Med. 2015; 16:428–439 6. Moran-Thomas A: How a Popular Medical Device Encodes Racial Bias. Boston Review. 2020. Available at: https://www.bostonreview.net/articles/amy-moran-thomas-pulse-oximeter/. Accessed January 30, 2023 7. Ralston AC, Webb RK, Runciman WB: Potential errors in pulse oximetry: I. Pulse oximeter evaluation. Anaesthesia. 1991; 46:202–206 8. McFarling UL: Pulse Oximeters and Their Inaccuracies Will Get FDA Scrutiny. What Took So Long? Stat. 2022 Available at: https://www.statnews.com/2022/11/01/pulse-oximeters-inaccuracies-fda-scrutiny/. Accessed January 30, 2023 9. Seitz KP, Wang L, Casey JD, et al.: Pulse oximetry and race in critically ill adults. Crit Care Explor. 2022; 4:e0758 10. Bickler PE, Feiner JR, Severinghaus JW: Effects of skin pigmentation on pulse oximeter accuracy at low saturation. Anesthesiology. 2005; 102:715–719 11. Jubran A, Tobin MJ: Reliability of pulse oximetry in titrating supplemental oxygen therapy in ventilator-dependent patients. Chest. 1990; 97:1420–1425 12. Ries AL, Prewitt LM, Johnson JJ: Skin color and ear oximetry. Chest. 1989; 96:287–290 13. Lee KH, Hui KP, Tan WC, et al.: Factors influencing pulse oximetry as compared to functional arterial saturation in multi-ethnic Singapore. Singapore Med J. 1993; 34:385–387 14. Sjoding MW, Dickson RP, Iwashyna TJ, et al.: Racial bias in pulse oximetry measurement. N Engl J Med. 2020; 383:2477–2478 15. Gadrey SM, Mohanty P, Haughey SP, et al.: Overt and occult hypoxemia in patients hospitalized with COVID-19. Crit Care Explor. 2023; 5:e0825 16. Groomes CL, Watson NL, Schulz EV: Increased risk of pulse oximeter bias in Black versus White preterm infants. J Perinatol. 2022 Sep 30. [online ahead of print] 17. Andrist E, Nuppnau M, Barbaro RP, et al.: Association of race with pulse oximetry accuracy in hospitalized children. JAMA Netw Open. 2022; 5:e224584 18. Vesoulis Z, Tims A, Lodhi H, et al.: Racial discrepancy in pulse oximeter accuracy in preterm infants. J Perinatol. 2022; 42:79–85 19. Pritchett CM, O’Halloran CP, Lay AS, et al.: Pulse oximetry error in patients with single-ventricle palliation. Catheter Cardiovasc Interv. 2022; 1:100394 20. Ebmeier SJ, Barker M, Bacon M, et al.: A two centre observational study of simultaneous pulse oximetry and arterial oxygen saturation recordings in intensive care unit patients. Anaesth Intensive Care. 2018; 46:297–303 21. 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:S18–S23 22. Wong AKI, Charpignon M, Kim H, et al.: Analysis of discrepancies between pulse oximetry and arterial oxygen saturation measurements by race and ethnicity and association with organ dysfunction and mortality. JAMA Netw Open. 2021; 4:e2131674 23. Bangash MN, Hodson J, Evison F, et al.: Impact of ethnicity on the accuracy of measurements of oxygen saturations: A retrospective observational cohort study. EClinicalMedicine. 2022; 48:101428 24. Burnett GW, Stannard B, Wax DB, et al.: Self-reported race/ethnicity and intraoperative occult hypoxemia: A retrospective cohort study. Anesthesiology. 2022; 136:688–696 25. Foglia EE, Whyte RK, Chaudhary A, et al.: The effect of skin pigmentation on the accuracy of pulse oximetry in infants with hypoxemia. J Pediatr. 2017; 182:375–377.e2 26. Gottlieb ER, Ziegler J, Morley K, et al.: Assessment of racial and ethnic differences in oxygen supplementation among patients in the intensive care Unit. JAMA Intern Med. 2022; 182:849–858 27. Fawzy A, Wu TD, Wang K, et al.: Racial and ethnic discrepancy in pulse oximetry and delayed identification of treatment eligibility among patients with COVID-19. JAMA Intern Med. 2022; 182:730–738 28. Sudat SEK, Wesson P, Rhoads KF, et al.: Racial disparities in pulse oximeter device inaccuracy and estimated clinical impact on COVID-19 treatment course. Am J Epidemiol. 2022 Sep 29. [online ahead of print] 29. Crooks CJ, West J, Morling JR, et al.: Differential pulse oximetry readings between ethnic groups and delayed transfer to intensive care units. QJM. 2023; 116:63–67 30. Henry NR, Hanson AC, Schulte PJ, et al.: Disparities in hypoxemia detection by pulse oximetry across self-identified racial groups and associations with clinical outcomes. Crit Care Med. 2022; 50:204–211 31. Savorgnan F, Hassan A, Borges N, et al.: Pulse Oximetry and Arterial Saturation Difference in Pediatric COVID-19 Patients: Retrospective Analysis by Race. Pediatr Crit Care Med. 2023; 24:458–462 32. Ma MA, Gutiérrez DE, Frausto JM, et al.: Minority representation in clinical trials in the United States: Trends over the past 25 years. Mayo Clin Proc. 2021; 96:264–266 33. Benjamin R: Race After Technology: Abolitionist Tools for the New Jim Code. Oxford, United Kingdom, Polity, 2019 34. Gokhale SG, Daggubati V, Alexandrakis G: Innovative technology to eliminate the racial bias in non-invasive, point-of-care (POC) haemoglobin and pulse oximetry measurements. BMJ Innov. 2023; 9:73–77 35. Kyriacou PA, Charlton PH, Al-Halawani R, et al.: Inaccuracy of pulse oximetry with dark skin pigmentation: Clinical implications and need for improvement. Br J Anaesth. 2023; 130:e33–e36 36. Department of Care, Social Health: Review Launched Into the Health Impact of Potential Bias in Medical Devices. 2021. Available at: www.gov.uk/government/news/review-launched-into-the-health-impact-of-potential-bias-in-medical-devices. Accessed January 30, 2023 38. U.S. Food and Drug Administration: Pulse Oximeter Accuracy and Limitations: Safety Communication. 2022. Available at: www.fda.gov/medical-devices/safety-communications/pulse-oximeter-accuracy-and-limitations-fda-safety-communication. Accessed January 30, 2023 40. Kuban KCK, Boynton-Jarrett R, Heeren T, et al.: A consideration of racism in pediatric epidemiologic studies. J Pediatr. 2021; 239:225–227 41. Williams JC, Anderson N, Mathis M, et al.: Colorblind algorithms: Racism in the era of COVID-19. J Natl Med Assoc. 2020; 112:550–552 42. Cerdeña JP, Plaisime MV, Tsai J: From race-based to race-conscious medicine: How anti-racist uprisings call us to act. Lancet. 2020; 396:1125–1128 43. Hidalgo DC, Olusanya O, Harlan E: Critical care trainees call for pulse oximetry reform. Lancet Respir Med. 2021; 9:e37 44. Sinaki FY, Ward R, Abbott D, et al.: Ethnic disparities in publicly-available pulse oximetry databases. Commun Med. 2022; 2:59 45. Okunlola OE, Lipnick MS, Batchelder PB, et al.: Pulse oximeter performance, racial inequity, and the work ahead. Respir Care. 2022; 67:252–257