A62-18 Racial Bias in Pulse Oximetry: Are Non-White Infants on Home Oxygen Therapy Being Weaned off Oxygen Too Quickly Using a Saturation Target Algorithm?

Background/Rationale

Pulse oximetry is central in the management of home oxygen therapy (HOT) in infants with bronchopulmonary dysplasia (BPD). The Recorded Home Oximetry (RHO) Program, a standardized remote monitoring program, provides objective oxygen saturation data to guide outpatient titration of oxygen flow rates between clinic visits. RHO has been shown to facilitate weaning off home oxygen therapy. Infants discharged from the neonatal intensive care unit (NICU) on HOT rely on pulse oximetry data to guide oxygen titration and ultimate discontinuation1. Literature suggests that pulse oximeters may overestimate peripheral oxygen saturation (SpO2) in individuals with darker skin pigmentation, potentially leading to occult hypoxemia1. If skin color affects oximetry accuracy2,3, this may influence the perceived readiness for oxygen weaning and contribute to inappropriately shorter HOT duration in darker-skinned infants compared to their white counterparts. The goal of this study is to evaluate whether differences in HOT duration by maternal race were observed among infants enrolled in the RHO implementation trial.

Methods

This was a retrospective analysis of preterm infants discharged on HOT from 14 centers participating in the multi-center implementation of the RHO program. Infants were provided a WIFI-enabled pulse oximeter that wirelessly transmitted recorded oximetry to an assigned data coordinating center. Infants’ oximetry was downloaded twice weekly, which produced a recommendation to decrease, maintain, or increase the infants flow rate based on the program’s algorithm. Race was categorized as black, white, or other/multirace based on maternal race identification reported at birth. Duration of HOT was defined from NICU discharge date to successful discontinuation. Duration was then compared in white and non-white infants.

Results

398 infants were followed by the RHO Program during the Implementation trial period from June 2021 to July 2024; 171 babies were successfully weaned utilizing the RHO Program. Of weaned infants, median duration of HOT did not differ by maternal race. Median (IQR) duration of oxygen was 45.5 days (30.3-83.0) among black infants, 50.0 days (27.0-83.8) among white infants, and 41.0 days (29-79.0) among infants of other or multiracial backgrounds (p = 0.94) (Figure 1). Median gestational age and BPD severity were similar across groups.

Conclusions

In this multicenter cohort of infants, duration of HOT did not differ by maternal race. These findings are hypothesis-generating and suggest that structural home monitoring protocols may mitigate potential inequities related to pulse oximetry bias4. Larger, adequately powered studies are needed to confirm these findings and assess for confounding variables.

This abstract is funded by: None