903 adults on ventilators in 24 intensive care units in England. Patients with darker skin were deliberately over-sampled. Skin tone was measured with a handheld spectrophotometer, not guessed from ethnicity or a colour chart. Each fingertip reading was paired with an arterial blood sample, the reference test. That gave 11,018 pairs.

All five models read higher on darker skin. At the same blood oxygen level, readings on the typical dark skin tone averaged 0.6 to 1.5 percentage points higher than on the typical light one, depending on the model. Every model tended to read too high when blood oxygen was low and too low when it was high. So at high oxygen levels, the extra points on darker skin meant less error. At low levels they meant more.

NHS Home Oximeters and Skin Tone: BMJ Study

Q: What can't it tell you?

Home use. These were critically ill patients. Poor circulation can make the skin tone error larger. Other devices. The authors say the results may not apply to other oximeter models. No ring or overnight recorder was tested. Low oxygen. Only 16% of blood samples were at 92% or below.

A 2026 BMJ study tested five fingertip pulse oximeters the NHS gave out for home use. All five read 0.6 to 1.5 points higher on darker skin, and missed low oxygen more often.

PulseOxTracker Team9 October 2026

Infographic. NHS home oximeters and skin tone. 0.6 to 1.5 points higher on darker skin: the average SpO2 reading on dark against light skin tone, at the same blood oxygen level. All five models read higher. Who was studied: 903 adults in 24 intensive care units in England, 11,018 fingertip readings paired with arterial blood samples, skin tone measured with a spectrophotometer, 2022 to 2024. What the study found. Low oxygen missed more often on darker skin: with blood at 92% or below, the oximeter read above 94% for 7.6% to 62.2% of readings on dark skin and 1.2% to 26.9% on light skin, depending on the model, 2.3 to 7.1 times as often. Fewer false alarms: the false positive rate fell with darker skin tone. All five models: tended to read too high at low oxygen, too low at high. Standard accuracy score (ARMS): no clear, consistent pattern across skin tones. What it cannot show: these were intensive care patients, not home users. Poor circulation in critical illness may have added error. Five fingertip models, and other models may differ. Only 16% of blood samples were at 92% or below. — A summary of the 2026 BMJ study: five NHS home pulse oximeters read 0.6 to 1.5 points higher on darker skin, and missed low oxygen more often.

During COVID, the NHS in England gave people fingertip pulse oximeters to check their oxygen at home. It's been known for years that oximeters can read high on darker skin. A BMJ study published in January 2026 tested five of those NHS models against blood samples to measure by how much.

Who was studied?

903 adults on ventilators in 24 intensive care units in England.
Patients with darker skin were deliberately over-sampled.
Skin tone was measured with a handheld spectrophotometer, not guessed from ethnicity or a colour chart.
Each fingertip reading was paired with an arterial blood sample, the reference test. That gave 11,018 pairs.

What did it find?

All five models read higher on darker skin.
At the same blood oxygen level, readings on the typical dark skin tone averaged 0.6 to 1.5 percentage points higher than on the typical light one, depending on the model.
Every model tended to read too high when blood oxygen was low and too low when it was high.
So at high oxygen levels, the extra points on darker skin meant less error. At low levels they meant more.

Does one point matter?

At a cut-off, yes. The home scheme told people to seek medical help at 94% or below.

For blood samples at 92% or below, the researchers estimated how often the oximeter still read above 94%, comparing the typical dark and typical light skin tone in the study:

Light skin: 1.2% to 26.9% of readings, depending on the model.
Dark skin: 7.6% to 62.2%.
That's 2.3 to 7.1 times as often.

False alarms went the other way and were less common on darker skin.

ARMS, the single accuracy score the international standard relies on, showed no clear pattern by skin tone.

What can't it tell you?

Home use. These were critically ill patients. Poor circulation can make the skin tone error larger.
Other devices. The authors say the results may not apply to other oximeter models. No ring or overnight recorder was tested.
Low oxygen. Only 16% of blood samples were at 92% or below.

What can you do with this?

The authors say trends in SpO2 should carry more weight than single readings, particularly for people with darker skin.

Compare your own nights with each other, on the same device and the same finger.
Be careful with fixed lines like 90%, which is what T90 counts against.
If you feel short of breath, get medical help whatever the number says.
See what else affects pulse oximeter accuracy.

Not a medical device. Does not diagnose, treat or prevent any condition. Discuss results and changes with a clinician.

References

Martin DS, Doidge JC, Gould D, Shahid T, Cowden A, Charles WN, Francis Johnson A, Garrett R, Mbema C, Olusanya O, Healy E, Rowan K, Mouncey P, Harrison DA; EXAKT Study Investigators. The impact of skin tone on performance of pulse oximeters used by NHS England COVID Oximetry @home scheme: measurement and diagnostic accuracy study. BMJ. 2026;392:e085535. doi:10.1136/bmj-2025-085535