FDA panel: Pulse oximeters less accurate for dark-skinned people
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Megan Robinson/Axios
Outside advisers to the Food and Drug Administration on Tuesday found "a clear signal" that pulse oximeters are less accurate for people with darker skin after scrutinizing evidence during a daylong meeting.
Why it matters: The expert panel's consensus that inaccuracies pose a distinct clinical risk for dark-skinned patients sets the stage for the FDA to further scrutinize manufacturing standards for the devices, which measure oxygen levels in the blood.
Driving the news: Wednesday's meeting followed years of longstanding concerns that were brought to the fore during the COVID-19 pandemic, when oximeters were widely used at home and in clinical settings.
- A 2020 article in the New England Journal of Medicine stated Black patients were nearly three times more likely than white patients to have low oxygen levels that go undetected.
- A September analysis from Sutter Health found Black COVID patients may have faced up to five hours in treatment delays due to inaccurate readings.
Yes, but: The panel split over recommendations for how to limit these inaccuracies in the future.
- Unresolved issues include how to categorize and report skin pigmentation data, what the accuracy percentage of pulse oximeters should be and how to best label warnings for consumers and clinicians.
What they're saying: "There needs to be transparency around it. How much inaccuracy is permissible? I'm not sure that there's a right answer," said panel chair Steven Nathan, the medical director of Inova's Advanced Lung Disease and Lung Transplant Program.
- Murad Alam, a panel member and Northwestern's vice chair of dermatology, said changes need to cover over-the-counter oximeters, which consumers use as an at-home medical alert.
- Alam urged the FDA to regulate them "with a level of vigor that would be sufficient to actually make some differences" without risking a spike in cost.
