Supporting health equity with bias-free pulse oximetry

点击翻译按钮获取中文摘要

Project Summary Pulse oximeters are essential for physicians’ diagnosis and monitoring of respiratory anomalies in patients. During the COVID-19 pandemic, their importance has grown because pulse-oximeter measurements of hypoxemia have become the major indication for hospitalizing patients. Clinical studies have shown that commercially available pulse oximeter measurements (SpO2) systematically overestimate true arterial oxygen saturation measurements (SaO2) for persons with dark skin pigmentation at low concentrations of O2 in the blood. This bias results in respiratory compromised persons with dark skin not meeting criteria for hospitalization or initiation of ventilator support, thereby putting specific populations (Black, Latinx, and Native American) at disproportionately greater risk for higher mortality or morbidity than those with light skin. No published explanations exist for this well-documented observation and, indeed, the available literature often contrarily states that pulse oximeter measurements are not affected by skin pigmentation. Our theoretical analysis and pilot research, however, demonstrate that the bias is due to present-day devices’ use of red light-emitting diode light sources, whose broad spectral bandwidth interacts with the spectral absorption of melanin concentration in skin to systematically shift the devices’ calibration. This shift causes artificially high values of SpO2 at low blood concentrations of O2 for patients with dark skin. The 3 proposed aims will extend our efforts to date, providing a scientific foundation for eliminating this bias and to foster development and promotion of simple, inexpensive, and bias-free pulse oximeters. In Aim 1, we will: (a) determine if there are other spectrally-dependent constituents in the finger that change with each pulse; (b) determine how light-source bandwidth interacts with melanin, including whether there are other pulse-dependent changes in spectral transmission through the fingers, and how sensitive SpO2 measurements are to light source bandwidth; (c) specify the practical peak wavelength and spectral bandwidth needed for bias-free pulse oximetry; and (d) fabricate an optimized light source that provides bias-free pulse oximeter measurements for testing in Aim 2. In Aim 2, we will demonstrate that the finger probe developed in Aim 1d provides bias-free pulse oximeter measurements (SpO2) that do not overestimate true arterial oxygen saturation measurements (SaO2) for persons with dark skin pigmentation at low blood concentrations of O2. In Aim 3, which will not employ human subjects as in Aims 1 and 2, we will translate our findings (i.e., print, in-person, and social media) to physicians, hospitals, and health care facilities commonly serving underrepresented minority populations. As part of Aim 3, in collaboration with a Stakeholder Board convened by the Institute for Health Equity Research, we will target the medical and affected communities to mitigate the impact of current devices through educational outreach and introduce them to bias-free pulse oximeters once they are available. Finally, we will license the technology to major manufacturers of pulse oximeters, giving preference to those who supply hospitals and clinics serving disadvantaged populations.