PROJECT SUMMARY
While pulse oximetry is a standard of care that guides management of critically ill patients, there is growing
evidence that racial bias in pulse oximeter accuracy may overestimate oxygenation in hypoxic patients with
pigmented skin and cause harm by impeding escalation of respiratory support. Pulse oximeters noninvasively
estimate blood oxygenation (SpO2) and peripheral perfusion (perfusion index, PI) by detecting tissue absorption
of red and infrared LEDs. SpO2 is based on calibration via the modulation ratio R, the ratio of pulsatile and non-
pulsatile absorption from one LED divided by the ratio from the second LED, regressed against arterial blood
oxygenation (SaO2). Oxy- (HbO2) and deoxy-hemoglobin (Hb) have distinct absorption at these wavelengths
and because, unlike melanin, they exhibit pulsatile variation, it was assumed that SpO2 estimates based on the
pulsatile signal would be pigmentation-independent. Alarmingly this assumption appears to be invalid as recent
work has found that oximeters are more likely to provide falsely normoxemic SpO2 in hypoxemic patients if they
are darkly pigmented. Differences in melanin-related absorption and scattering, particularly at red wavelengths,
are thought to underly pigmentation-related errors in estimates of oxygenation, and likely also impact estimates
of perfusion. However, studies demonstrating pigmentation-related device error have not controlled for other
confounding physiologic factors that may impact pulse oximetry, and it is therefore not known if health
disparities faced disproportionately by people with pigmented skin may independently drive pulse oximeter
errors. Overcoming pigmentation-related pulse oximeter bias and decoupling it from other sources of oximeter
error is critically important to ensuring that darkly pigmented patients, who already face inequities in the health
care system, are not denied life-saving care due to devices that fail to meet the needs of a diverse population. To
solve these problems, we propose (i) utilizing novel experimental systems that allow pigmentation to be varied
while other physiologic parameters remain constant, and (ii) shift pulse oximeter wavelengths from red (with
high melanin absorption and scattering) to short-wave infrared (SWIR, 900-1700 nm) where melanin absorption
and scattering is minimized. Successful completion of the proposed studies will ensure that life-saving pulse
oximeter technology serves the needs of patients across the full spectrum of pigmentation and no longer fails an
already marginalized community.
