The study assessed the effect of skin tone on the accuracy of five fingertip pulse oximeters used in the NHS England COVID Oximetry @home scheme in adult intensive care patients. Peripheral oxygen saturation (SpO2) measurements were compared with arterial oxygen saturation (SaO2) as the gold standard, with skin tone measured objectively with a spectrophotometer using an individual typology angle. All oximeters overestimated SpO2 at lower SaO2 values and underestimated at higher values. On average, SpO2 values were 0.6–1.5 percentage points higher in patients with darker skin tone (−44° angle) versus lighter skin tone (46°). The incidence of false negative results at SpO2 thresholds ≤ 92% and ≤ 94% increased with darker skin tone; the proportion of SpO2 measurements > 94% despite SaO2 ≤ 92% was 5.3 to 35.3 percentage points higher (7.6–62.2% vs. 1.2–26.9%, frequency ratio 2.3–7.1). Conversely, false positives decreased with darker skin tone. The study identified higher SpO2 measurements in patients with darker skin tones, which may lead to differences in false negative and positive results in detecting hypoxemia.