Abstract 14058: Racial Bias in Pulse Oximetric in Pediatric COVID Patients

Abstract

INTRODUCTION Pulse oximetric (Sp02) may overestimate patient oxygen saturation. This bias is possible more pronounced in patients of black ethnicity. The current study aimed to assess whether SpO2 accurately predicted laboratory oxygen saturation (SaO2) in symptomatic hospitalized pediatric patients with covid 10. And whether disparities between black and white race ethnicity existed. Methods: This retrospective study evaluates covid 19 patients between May 2020 and December 2021 admitted to Texas Children's Hospital (TCH). We used multivariate logistic regression to examine the association between race and bias in Sp02. The bias is defined as the difference in SaO2 and SpO2. All analyses were performed using Python Statsmodels v0.12.2. Results: There were 4300 patients with covid with complete physiological data in the Sickbay and electronic chart system. Figure 1 shows a paired SpO2 and SaO2 comparison between the white and black hospitalized pediatric patients with covid 19. Figure 2 depicts the paired bias comparison between the black and white hospitalized pediatric patients with covid 19. Multivariate logistic regression showed that the black race-ethnicity is associated with worse bias (OR (95%CI) = -1.94 (-3.22, - 0.65), p<0.003) in the non-invasive oxygen saturation. Discussion: We found a bias in the measurement of Sp02 about Sa02 in symptomatic hospitalized pediatric patients with the diagnosis of covid 19; furthermore, the race-ethnicity is related to an increased bias in the reading of Sp02 about Sa02. Like the previous studies, we consistently found more bias in reading Sp02 about Sa02 in the black population.