CHOP Researchers Investigating Skin Color Bias in Pulse Oximetry Measurements

By Lauren Ingeno

Pulse oximeters can help clinicians make critical decisions, ranging from deciding whether to admit a patient to the hospital to choosing the kinds of treatment they should receive. 

But the COVID-19 pandemic raised new questions about the accuracy of pulse oximeter readings — some Black and Brown patients were in fact much sicker than the device suggested

That’s because pulse oximeters work by measuring the amount of light the finger absorbs, which is proportional to the concentration of oxygenated hemoglobin in the blood. Past studies have shown the pulse oximeter can overestimate oxygen saturation in patients with darker skin, possibly because melanin can interfere with the measurement.

While pulse oximeters are used in nearly every pediatric hospitalization to monitor patients, few studies have investigated the relationship between skin color and pulse oximeter accuracy in children.

Now, Michael O’Byrne, MD, an interventional cardiologist at Children’s Hospital of Philadelphia who leads the Cardiac Center Clinical Research Core, and Halley Ruppel, PhD, RN, an assistant professor at CHOP and Penn Nursing, have received funding from the National Institutes of Health to lead a multisite, prospective study of 600 patients under 18-years-old who are undergoing cardiac catheterization. Dr. O’Byrne and Dr. Ruppel are both core faculty members of Clinical Futures, a CHOP Research Institute Center of Emphasis.

Their goal is to contribute knowledge about this issue to reduce disparities and inequities in healthcare, Dr. O’Byrne and Dr. Ruppel said. Long-term, that could mean changes to the pulse oximeter design or algorithm. In the meantime, the research team’s data could help inform individual clinicians’ decision-making processes. 

“Our findings might help clinicians to think differently when they’re evaluating a child who looks sicker than the number they’re seeing on the pulse oximeter,” Dr. O’Byrne said. “This could help them decide whether they should perform additional testing or interventions.”

 In a preliminary study led by Dr. Ruppel, researchers used retrospective electronic health record data from a pediatric cardiac catheterization laboratory to analyze the connection between race and pulse oximetry. The findings, published last year in JAMA Pediatrics, found that 12% of Black patients with true low oxygen levels had a “normal” pulse oximetry measurement, compared with 4% of white patients.

As an extension of that work, the research team will use a tool called a colorimeter in the new prospective study to obtain objective skin color measurements of patients, rather than relying on race reporting from patients’ electronic medical records. 

“Because of this presumed mechanism of the way the pulse oximeter works — the melanin absorbing light — we rthink it’s important that we measure skin color rather than relying on race reporting,” Dr. Ruppel said.

A second strength of the new study is the unique population that the researchers will be enrolling and evaluating in real-time. Children undergoing cardiac catheterization — which helps provide information about the heart and blood vessels — have a wide range of baseline arterial oxygen saturations, are continuously monitored with pulse oximeters, and their arterial blood oxygen is measured as part of routine care. 

In addition to CHOP, data will come from patients at Texas Children’s in Houston and NewYork-Presbyterian Morgan Stanley Children’s Hospital in Manhattan.