40+ years later, we're still selling and relying on this very basic, unreliable medical reading
And we wonder why people are skeptical of medicine--and why Black women in particular are angry. If we can't trust medicine to fix something this obvious, what can we trust from our healthcare?
Excuse me for a rant here. I won’t do it often, but this one breaks my heart. I’ve been a provider; the thought of doing harm without knowing it is terrible. The fact that our health system has knowingly ignored the unreliability of this very simple measurement of wellbeing every. single. day. for almost five decades is mortifying.
Until recently, the documented unreliability of pulse oximetry readings was largely ignored and even denied by both manufacturers and regulators, and not well understood by clinicians, allowing for the widespread proliferation of countless biased devices. Worse, clinicians and medical students have largely been unaware of pulse oximeter shortcomings, so we sure can’t expect the aide taking the measurement to welcome expressions of doubt about reliability.
You can buy a pulse oximeter for $10 just about anywhere. Over 3,000 people bought this one alone. Particularly since COVID—when the world grabbed them up by the millions—the market has been flooded with cheap pulse ox devices. You probably have one somewhere. Any time a healthcare provider takes your blood pressure, they’re probably taking a pulse ox reading, too. Hundreds of medical-grade pulse oximeters are on the market, and countless consumer-grade models are available.
What they’re not guaranteed to be? Accurate. And we’ve known it for decades.
Well, that’s not completely true. If you’re White, your odds are pretty good: studies say pulse oximeters overestimate your oxygen level “only” 13% of the time. (Yes, one out of ten times—which is bad enough. Would you fly an airline that crashed 13% of the time?) But if you’re Black or Brown, for one out of five readings, you might as wave your finger in the air. For you, an inaccurate reading may say you’re better oxygenated than you are, potentially depriving you of oxygen, delaying treatment, damaging you further, or hastening death from multiple diseases. You’re right to ask pointed questions about that reading, particularly if it’s already on the low side.
And we’re not talking missing a minor cold. We’re talking COPD and acute respiratory distress syndrome (ARDS)—people gasping for air. People on ventilators. Transplants and heart failure.
We’re also not taking about differences between, say, a reading of 10 or 90. We’re talking a four percentage point inaccuracy that can mean the difference between being sent home, unrecognized as very sick and at risk for worse, or timely, effective treatment that saves lives, and in the long run, medical costs for all of us.
A healthy oxygen saturation level for a healthy person at sea level is between 95% to 100%. A reading below 92% may indicate the hypoxia, a condition where the body's tissues aren't getting enough oxygen—which can cause something as ‘simple’ as a confusion-related accident when you’re driving. If your reading is 88% or lower, you’re supposed to seek immediate medical attention.
And yet, we know these devices overestimate oxygen saturation one out of five times in darker skinned people by as much as four percentage points—for example, read 92% when it’s actually that dangerous 88%. An ICU study with the sickest patients found that 12% of the time, when Black patients had a pulse oximeter reading in the “safe” range—92% to 96%, their actual saturation on a much more reliable arterial oxygen (ABG) test was below 88%. Yet decisions whether to give more aggressive care continue to be made on the unreliable pulse ox reading, even more so when staff are stressed and busy. If it were your family member, you’d be furious.
If you have the stomach for it, it’s easy to uncover too many more instances of unintended, but existing, racial bias in health care today.1 This is just one, but it’s one to which millions of people are exposed daily, and not something we just discovered recently. We’ve been gathering data on pulse oximeters since the 1980s, with the first definitive studies published in 1990. As this Hopkins review says, it’s a “problem hiding in plain sight.”
In 2013, 23 years after published research reporting the issue, the FDA recommended that manufacturers test pulse oximeters on participants with a range of skin pigmentations. So much for that: we know that, today, only a quarter of medical manufacturers follow that recommendation—yes, those are the devices being sold to hospitals and doctors’ offices. Yet, the devices are cleared for use—and those are the ones that go through the FDA approval process. It’s hard to believe the flood of cheap ones bother; those manufacturers are here today and gone tomorrow. Talk about a toothless recommendation. There’s now updated FDA guidance, but it’s still on the recommendation level for what’s described as a “wicked problem with no easy solutions.” That horse is well out of the barn.
The damage? Studies during COVID showed inaccurate pulse ox readings on one out of five Black or Brown patients. Today, Black people in the hospital are 32% more likely than White patients to have pulse ox readings that overestimate their oxygen levels by at least 4 percentage points. This isn’t about needing a little fresh air. We’re talking about delays in life-saving treatments—ultimately much more costly care and re-hospitalizations that show up in eventually in everyone’s healthcare costs.
This affects all of us. The US continues to rank 11th out of 11 peer nations in healthcare outcomes and costs, driven by data exactly like this. These willfully-ignored mistakes bear costs that are ours as a nation, and for those of us in healthcare, the reasons it’s hard to tell people we have the best healthcare in the world while keeping a straight face. This isn’t about people of color; it’s about who we are as a nation and what we believe in. You should be as angry about this—and asking as many questions—as I am.
This webpage is for informational purposes only and does not constitute medical advice or diagnosis. Consult a licensed clinician for medical advice applicable to your specific needs.
There’s plenty more about our racial biases in medicine, from the unbelievable but very real Tuskegee syphilis ‘study’ that every Black person knows about, to the 14,000 Black kidney transplant patients moved up on the transplant waitlist just last April when a lab test used for decades was found to calculate kidney function differently for Black patients. Or Black patients whose heart attack symptoms—different from Whites—continue to be ignored in emergency rooms. Perimenopausal Black women, for instance, who today are a third less likely today to receive guideline-based treatment such as a stent to open a blocked artery when having a heart attack than Whites. Or Black diabetics who are half as likely as Whites to get limb-saving procedures—tripling the odds of amputation. The downstream cost affects all of us: There are few studies about how amputation affects work, but one study showed a third of amputees with prior work experience were unable to work after amputation, and the delay between amputation and return to work was over two years. Those are societal costs that eventually get passed down to us in multiple ways, like unemployment and our own healthcare costs. Yes, much of the bias is ignorance; very little if any was intended at the start. But don’t tell that to the families who bear the burden of loss. Their mistrust and anger are not misplaced.