Background: Traditional pulse oximetry (SpO2) has been widely criticised for racial bias, yet the underlying mechanisms remain obscured by the use of “Race” as a categorical social proxy. This study deconstructs these disparities by isolating the impact of objective skin physics and clinical severity on diagnostic accuracy. Methods: Using the ENCoDE dataset (N= 68,413 synchronised SpO2 SaO2 pairs), we replaced self-reported race with the CIELAB L* (Lightness) scale. Participants were segmented into quartiles based on L* values (Q1: 42.1-56.9; Q4: 68.3-91.3). We quantified instrumental bias via Mean Absolute Error (MAE), assessed diagnostic failure risk through logistic regression, and developed a “Time-to-Event” algorithm to measure the “Lost Window”- the duration of unrecognised occult hypoxemia. Results: A significant linear gradient of error was observed (P < 0.0001, F = 1571.86), with the darkest quartile (Q1) exhibiting an MAE of 21.27% compared to 15.27% in the lightest (Q4). This “Double Jeopardy” effect, where darker skin tone and lower arterial saturation compound, resulted in a 28.8 Hour “Lost Window” for Q1 patients, nearly double that of lighter skinned cohorts. Logistic regression yielded an Odds Ratio of 0.9518 per unit of L*, indicating a 4.82% increase in the risk of occult hypoxemia for each 1-point decrease in skin lightness. Conclusion: Pulse oximetry bias is a predictable function of skin physics and clinical state. Our proposed “Skin-Aware” recalibration algorithm successfully reduced recognition delays by 7.5 hours in simulations. These findings advocate for a shift toward L* stratified device validation and software-based corrections to ensure physiological equity in acute care.

References:

1. Sjoding MW, Dickson RP, Iwashyna TJ, Gay SE, Valley TS. Racial bias in pulse oximetry. N Engl J Med. 2020;383(25):2477-2478. doi:10.1056/NEJMc2029161

2. Bickler PE, Feiner JR, Severinghaus JW. Effects of skin pigmentation on pulse oximeter accuracy at low saturation. Anesthesiology. 2005;102(4):715-719. doi:10.1097/00000542-200504000-00008

3. Alsan M, Arbel Y, Durvasula M, et al. Clinician behavior when skin-tone affects test results. NBER Working Paper No. 34168. Published February 2026.

4. 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(7):730-738. doi:10.1001/jamainternmed.2022.1906

5. Vyas DA, Eisenstein LG, Jones DS. Hidden in plain sight- Reconsidering the use of race correction in clinical algorithms. N Engl J Med. 2020;383(9):874-882. doi:10.1056/NEJMms2004740

6. Gottlieb ER, Ziegler J, Morley K, Rush B, Celi LA. Assessment of racial and ethnic differences in oxygen supplementation among patients in the intensive care unit. JAMA Intern Med. 2022;182(8):849-858. doi:10.1001/jamainternmed.2022.2587

7. Wong AK, Capraro A, Linscott R, et al. Developing the ENCoDE dataset: A multi-center approach to evaluating diagnostic errors in pulse oximetry. J Clin Monit Comput. 2024;38(2):455-467.