Covid-19 made pulse oximeters an even more important tool for measuring the amount of oxygen in the bloodstream than they had been before. Widely used in hospitals and health systems, these small finger clips were also flying off the shelves in pharmacies and being ordered online. But pulse oximeters aren’t as accurate as many people — and doctors — believe.
A good example of their limitations was my patient, Mr. V, a Hispanic man in his 50s. He came to the emergency department one morning after days of having trouble breathing. We quickly admitted him to the intensive care unit for rapidly progressing Covid-19. That afternoon, his breathing became increasingly labored, even when he used a mask that pushed oxygen into his lungs. That wasn’t enough. I had to put a breathing tube in him. The pulse oximeter on his finger showed that his oxygen levels were in the mid-90s — which is normal. But he kept getting worse. Fast.
Was I missing something? I ordered a blood test called an arterial blood gas to directly measure the oxygen levels in Mr. V’s blood. While this test generally agrees with the pulse oximeter reading, I suspected they might differ in this case. I was right — Mr. V’s arterial blood gas test showed that his oxygen levels were in the mid-80s, well below the lowest “safe” limit of 88, starving his body of oxygen. We tried everything we could to help him, but he died that evening.
Mr. V’s death illustrates a basic health care inequity that must be addressed: Pulse oximeters do a poor job of measuring the blood oxygen level of people with darker skin, like Mr. V. Their skin has more melanin, a skin pigment that interferes with measuring the oxygen level in the blood.
A pulse oximeter works by shining light through the skin to measure the amount of oxygen in the blood. It has replaced the arterial blood gas test as a vital sign because it is fast, easy to use, painless, and provides continuous readings. But recent studies have revealed that pulse oximeters are not as accurate as once believed. In February 2021, the Food and Drug Administration acknowledged the problem in a safety communication, warning that pulse oximeters might be less accurate in people with dark skin pigmentation. A year later, the Critical Care Societies Collaborative petitioned the FDA to do more than that.
The FDA now recommends using pulse oximetry readings as “an estimate of blood oxygen saturation.” As an estimate, rather than a precise measurement, a pulse oximeter reading of 90% may represent an arterial blood oxygen saturation somewhere between 86% and 94%. That’s an enormous spread for readings as crucial as blood oxygen saturation, which can guide life-or-death interventions for critically ill patients. Simply put, pulse oximeters are not precise enough.
Several colleagues and I at Duke Health have studied pulse oximetry disparities in detail. We reviewed five electronic health record databases that included records from 215 hospitals over a 20-year period. We examined arterial blood gas measurements from 87,971 patients from these databases and compared them to the patients’ pulse oximeter readings. As we reported in JAMA Network Open, even after adjusting for confounders like age and how sick patients were, we identified discrepancies between arterial blood gas and pulse oximeter readings — called hidden hypoxemia — in all races. The problem is worse, however, among patients of color: For every five white patients with hidden hypoxemia, seven Black patients (38% more) were affected. Hidden hypoxemia was associated with 70% higher mortality and more severe organ dysfunction. We also found that patients of color were 40% to 70% less likely to have had an arterial blood gas test ordered to confirm their pulse oximeter readings.
People who suspect they have Covid-19 are currently advised to stay home and check their oxygen levels using a pulse oximeter, but the readings they get may falsely appear to be normal. This could cause some patients to stay home when they really should go to the hospital, delaying treatment, which happens more among people of racial and ethnic minority groups.
Health tech companies need to help solve this problem by introducing equity into their equipment testing to ensure that it works equally well on patients of all races and ethnicities. Current FDA guidelines require that, in calibrating pulse oximeters, testing must include only two “dark-skinned” patients of 10, or 15%, whichever is greater. That isn’t sufficient to ensure accurate readings and draw statistical conclusions about these discrepancies.
Our team is currently researching how to improve pulse oximetry accuracy, regardless of skin color. But until this technology is fixed, it’s important that everyone be made aware of the limitations of these devices: Doctors should order arterial blood gas tests more frequently to confirm safe oxygen levels. There aren’t yet official guidelines for at-home use, but I recommend that anyone using one of these devices — especially those with dark skin — call their doctor if their oxygen level declines or they feel short of breath even if a pulse oximeter indicates their blood oxygen level is in the normal range.
This simple, accessible test is ubiquitous — but it isn’t equitable. It needs to be improved so everyone has equal protection against low oxygen levels and the ability to receive the oxygen therapy they need.
A. Ian Wong is a critical care specialist and pulmonologist at Duke Health in Durham, North Carolina.