A new retrospective study from Johns Hopkins Hospital adds to concerns about pulse oximeter reliability. Reviewing five years of intensive care unit data, researchers found that the devices gradually shifted from overestimating blood oxygen levels to underestimating them at higher oxygen levels. The inaccuracies were worse in patients of color, but the directional change appeared across all patients, suggesting the problem is not simply a fixed racial bias.
The findings complicate an already murky picture. Earlier work, including a landmark 2020 study, showed pulse oximeters overestimate oxygen in darker skin tones, potentially masking dangerous hypoxemia. But a recent FDA-funded prospective study called EquiOx found underestimation, and the Hopkins data show year-to-year inconsistency. As one pulmonologist put it, a broken clock that is always five minutes slow can be adjusted for, but a clock that sometimes runs fast and sometimes slow is far harder to work around.
Masimo, which manufactured the oximeters used in the study, said it is unaware of any sensor, software, or manufacturing changes during the study period that would explain the results. The company also pointed to two studies it supported that found no skin-tone differences. The Hopkins authors acknowledge the underlying cause is unknown; one called the devices a "black box." The findings underscore how difficult it is for clinicians to interpret pulse oximeter readings when the direction and size of error can change over time.