- Many oximeters have a bias against dark skin because their manufacturers calibrated them in trials most of whose participants were white people.
- Habib Naqvi of the UK NHS said there is a possibility that a sizeable fraction of India’s COVID-19 deaths during the second wave could be traced to false oximeter readings.
- He urged officials in India to review oximeter readings the way the UK currently is, and advised health workers to take multiple readings when working with dark-skinned people.
The director of the Race and Health Observatory of the UK National Health Service has said the bias of oximeters against people of dark skin is a matter of critical concern around the world, including in India.
The observatory is currently reviewing potential inaccuracies in the way some medical devices, including oximeters, and drugs respond to people with dark skin, which in turn could affect medical diagnosis and access to treatment.
With reference to oximeters – portable devices that can quickly measure the blood’s oxygen saturation, a value affected by COVID-19 – Habib Naqvi said all the research thus far shows “the darker the pigmentation of the person, the more inaccurate the reading”. That is, he added, “They overstate the levels of oxygen in the blood.”
This means if a dark-skinned person’s oxygen saturation is already below the common threshold of 94%, they could be denied access to important COVID care based on oximeter readings alone.
In a 30-minute interview to The Wire, Naqvi drew on evidence available in the UK, where he is based, to say: “Black people are twice as likely as white people to get COVID-19 while Asians are 1.5-times more likely”. Accurate oximeter readings can mark the difference between life and death for susceptible members of these communities.
On November 20, science editor of The Sunday Times Ben Spencer had written: “It is hard not to come to the conclusion that Black and Asian people were denied lifesaving treatment because the device clipped to the end of their finger was designed for someone of a different race”.
Oximeters have a bias against dark skin – a ‘feature’ Spencer called “racist” – because their manufacturers tested and calibrated them in trials most of whose participants were white people.
These devices work by emitting radiation of a fixed frequency and then observing its interaction with the finger.
Dark pigmentation absorbs more light than white, so oximeter readings with people with dark skin can be and is often inaccurate.
Research by scientists at the University of Michigan reached a similar conclusion – that oximeters were thrice as likely to misread African-American patients’ oxygen saturation levels as those of white patients. Spencer had written “12% of Black patients” in this study “who were considered to have safe oxygen levels were in fact dangerously hypoxic”.
Naqvi said in his interview there is “a huge possibility” that a sizeable fraction of COVID-19 deaths in India during the second wave could be traced to false oximeter readings. So he also expressed hope that India and other countries review oximeter readings the way UK currently is.
In the meantime, because these reviews could take months to complete, Naqvi said nurses and doctors using oximeters with people with brown and black skin must take multiple readings, and inform their diagnoses with other information relevant to the patient, including their medical history.
The bias against dark skin is not limited to oximeters. According to Naqvi, drugs called ACE-inhibitors have been known to be effective with Black people while morphine-based painkillers may not work so well with people of Chinese heritage.
The reasons, again, are similar: they were tested mostly with white people. Naqvi also said the manuals used to train doctors are often designed assuming the patient is white.
Editor’s note: The issue of devices with implicit racial bias was prominently in the news in 2015. At the time, a sensor-based soap-dispenser at a hotel in Atlanta didn’t respond when a Black man tried to use it, but did when a white man tried.
Irrespective of whether India chooses to review oximeter readings, it has its own biases to contend with – particularly against women and people of other genders.
For example, during India’s ongoing COVID-19 epidemic itself, women’s pathogen exposure risk), unequal access to healthcare and lack of data about the effect of certain medical conditions on women specifically undermined efforts to understand how the disease affected women’s bodies or if they could benefit from different kinds of treatment.