Self-reported Race/Ethnicity and Occult Hypoxemia: Reply

We thank Dr. Rosenbloom for her comment on our work on self-reported race/ethnicity and intraoperative occult hypoxemia.^1,2  While we recognize race and ethnicity as social constructs, due to the nature of this large retrospective study, it was the most appropriate means to investigate this important topic. In any retrospective study, methodologic trade-offs are made, and our study is no different. We note the two references Dr. Rosenbloom provided demonstrating the nonequivalence of race/ethnicity and skin pigmentation^3,4  but also would like to reference a study demonstrating that although there is variation in skin pigmentation within racial and ethnic groups, there is a significant correlation between ancestry/ethnicity and skin pigmentation.⁵  Furthermore, self-reported race has been shown to correlate with the Fitzpatrick skin phototype classification system.⁶  Because of the discrepancies mentioned above and the nature of our retrospective study, we listed this as a limitation of to our study, but we believe this approach was reasonable and likely valid.

Similarly, the potential for varied pulse oximeter location and the use of nail polish/treatments were unknown variables within our data set. We therefore also highlighted these as limitations of our study. Additionally, we do not have data to support the concept that location of pulse oximeters or the use of nail polish/treatment would be different for any one group within our study. Therefore, we believe that a large data set such as ours allows us to capture the most likely and common practices and patient presentations, and it is likely valid to assume this means placement of pulse oximetry on a patient’s finger and the presence or absence of interfering nail polish colors would not differ between groups. Indeed, there is no real reason to argue otherwise.

In regard to the use of the Fitzpatrick skin phototype classification, we agree that an objective measurement of skin pigmentation would be beneficial in a prospective study. This classification system was originally created for differentiating skin pigmentation in White patients⁷  and was later amended to include more darkly pigmented skin,⁸  but concerns regarding the functionality in patients of color exist even in this more objective approach.⁹  Alternatives to the Fitzpatrick skin phototype classification have been posed, including a simple color bar tool¹⁰  or more advanced techniques using spectrophotometry or colorimetry.¹¹  Either way, a prospective study of this topic would be best, although it must be noted based on the incidences measured in our study that the sample size needed is likely prohibitively large, making this a difficult study to conduct.

We hope our article sheds light on this important topic and results in additional research to improve the care of Black and Hispanic patients in the future, and we thank Dr. Rosenbloom for the measured criticism.

The authors declare no competing interests.