A pulse oximeter looks simple: clip it on a finger, wait a few seconds, read the number. But the device is doing something more delicate than it appears — shining two wavelengths of light through your fingertip and calculating oxygen saturation from how much of each wavelength gets absorbed. Anything that interferes with that light path, or with the blood flow underneath it, can shift the number on the screen. If you or a family member relies on a home oximeter to track a chest infection, COPD, or a post-viral recovery, knowing what genuinely affects pulse oximeter accuracy — and what doesn’t — matters more than the reading itself.
How a pulse oximeter actually measures oxygen
The device sits on a fingertip and passes red and infrared light through the tissue to a sensor on the other side. Oxygenated and deoxygenated blood absorb these wavelengths differently, and the device uses that difference, timed to your pulse, to estimate the percentage of haemoglobin carrying oxygen — your SpO2. It is an estimate, not a direct laboratory measurement, and it depends on a clean, steady light path and normal blood flow at the sensor site. Disrupt either of those and the estimate drifts, sometimes by more than most users would expect.
Nine things that can throw off your reading
Some of these are well known; several are not. All of them can move a genuine 96% reading down into a range that looks concerning, or up into a range that hides a real problem.
| Factor | Effect on the reading | What to do |
|---|---|---|
| Nail polish, gel or acrylic nails | Dark polish especially can block or scatter the light path | Remove polish, or use a different finger |
| Henna dye or tattoos on the fingertip | Pigment interferes with light absorption in the same way as nail polish | Use an unaffected finger |
| Cold hands or poor circulation | Reduced blood flow to the fingertip weakens the pulse signal the device relies on | Warm the hand first; rest for a minute before reading |
| Movement while measuring | Motion is one of the commonest causes of an inaccurate or fluctuating reading | Keep the hand still and relaxed for the full reading |
| Wrong probe on the wrong body part | An adult probe used on a child, or a finger probe used on an ear, can read tens of percentage points off true | Use the probe size and site the device is designed for |
| Scar tissue, wounds or skin conditions on the fingertip | Alters how light passes through the tissue | Try a different, unaffected finger |
| Bright ambient light | Strong light hitting the sensor can interfere with the reading it’s trying to take | Take readings out of direct sun or bright lamps |
| Darker skin pigmentation | Can produce small but real overestimates of true oxygen saturation | Track the trend over time rather than a single number; flag any concern to a clinician |
| Non-medical consumer devices (fitness trackers, some smartwatches) | Not built or validated to the same standard as a dedicated medical oximeter | Use a dedicated device for any reading you intend to act on |
What the regulator and the NHS have actually said
This isn’t a fringe concern. In the UK, medical devices sold for clinical use — including pulse oximeters — must meet the UK Medical Device Regulations 2002 and carry a valid CE, CE UKNI or UKCA mark; the MHRA does not itself test or approve each device, but manufacturers must demonstrate compliance. Separately, NHS England issued a Patient Safety Alert in December 2018 specifically about the risk of harm from inappropriate placement of pulse oximeter probes, warning that using a probe on the wrong body part, or an adult probe on a child, could produce readings up to 50% lower or 30% higher than the true value — a striking margin for a device many households treat as infallible.
On skin pigmentation, the position is more nuanced. The regulator has acknowledged that people with lighter skin may show small differences in reported results compared with those with darker skin, though it states it is not aware of incidents where skin colour caused harm during clinical care. NHS guidance for people with darker skin recommends watching how a reading changes over time rather than treating a single figure as definitive, and raising any concern with a healthcare professional rather than relying on the device alone. Notably, the MHRA does not recommend home use of pulse oximeters by the public outside clinical advice — a caveat worth knowing before you treat a home reading as a substitute for medical judgement.
Getting a reading you can actually trust
None of this means a home pulse oximeter is useless — used correctly, it remains a genuinely useful early-warning tool for tracking a respiratory illness at home. But a trustworthy reading depends on ruling out the interference above first:
- Warm the hand and let it rest for a minute before measuring.
- Remove nail polish or use a bare finger without tattoos, henna or scarring.
- Sit still and keep the hand steady for the full reading cycle.
- Take the reading away from bright direct light.
- Use the probe size the manufacturer intends for your age group and site.
- Take two or three readings a few minutes apart and look at the trend, not a single number.
- Treat any reading you’re unsure about — especially a sudden drop — as a reason to contact a healthcare professional, not a number to argue with.
The bottom line on pulse oximeter accuracy
A home pulse oximeter is a light-based estimate, not a lab test, and its accuracy is more conditional than the simple number on the screen suggests. Nail polish, cold hands, movement, ambient light, the wrong probe and skin pigmentation can all move a reading in ways that matter, and UK regulators have formally flagged some of these risks rather than dismissing them. The sensible approach for any household relying on one is the same watchdog principle behind every device on the market: know what can distort the number before you trust it, control for it where you can, and treat a concerning reading — not a single figure but a genuine trend — as your cue to seek professional advice rather than a verdict in itself.