British experts have joined what’s growing into a global call for action on medical devices that are prone to unfair biases, including blood oxygen monitors and certain artificial intelligence (AI) enabled tools, to prevent harm to ethnic minorities and women.
A recent report detailed the findings of an Independent Review of Equity in Medical Devices, scrutinising the extent and impact of ethnic and other unfair biases in the performance of equipment commonly used in the NHS, reports The Independent.
It focused on optical devices such as pulse oximeters, AI-enabled devices and certain genomics applications, where evidence suggested there was substantial potential for harm.
The panel found that pulse oximeters – blood oxygen monitors – can overestimate the amount of oxygen in the blood of people with darker skin tones, resulting in possible treatment delays if dangerously low oxygen levels are missed.
The experts say they did not specifically look at the use of these devices during the pandemic, but because there was an overwhelming number of people with very low oxygen levels, “the likelihood is that that inaccuracy was large at that time”.
Daniel Martin, professor of perioperative and intensive care medicine, Peninsula Medical School, University of Plymouth, said: “We can only say that there’s association between the harm and the inaccuracy and not causation.
“But it’s a reasonably strong signal that there’s a potential of harm there, particularly during Covid when oxygen levels are so very low.”
The review makes a number of recommendations in relation to the devices, including patients being advised to be alert for other symptoms, like shortness of breath, chest pain and fast heart rate.
It also suggests researchers and manufacturers should work to produce devices that are not biased by skin tone.
On AI-enabled devices, the review found evidence of potential biases against women, ethnic minority and disadvantaged socio-economic groups, and highlighted potential under-diagnosis of skin cancers for people with darker skin.
There is also a long-standing problem of under-diagnosis of heart conditions in women, which AI algorithms in medical devices could worsen, the panel suggests.
The University of Liverpool’s Professor Dame Margaret Whitehead, chairwoman of the review, said: “The advance of AI in medical devices could bring great benefits, but it could also bring harm through inherent bias against certain groups, notably women, people from ethnic minorities and disadvantaged socio-economic groups.
“Our review reveals how existing biases and injustices in society can unwittingly be incorporated at every stage of the lifecycle of AI-enabled medical devices, and then magnified in algorithm development and machine learning.
“Our recommendations, therefore, call for system-wide action by many stakeholders that needs to be implemented as a …priority with full government support.”
The advice also suggests that there be renewed efforts to increase skin tone diversity in medical imaging databanks used for developing and testing optical devices for dermatology, including in clinical trials, and for improved tools for measuring skin tone being incorporated into optical devices.
The Department of Health and Social Care said significant action is already being taken to overcome potential disparities in the performance of medical devices.
This includes the Medicines and Healthcare products Regulatory Agency (MHRA) now requesting that approval applications for new medical devices describe how they will address bias.
NHS guidance has been updated to highlight potential limitations of pulse oximeter devices on patients with darker skin tone.
Equity in Medical Devices summary (Open access)
The Independent article – Experts call for action on medical devices prone to biases (Open access)
See more from MedicalBrief archives:
Race versus skin tone debate in resolving pulse oximeters’ false readings
How software bias leads to under-diagnosis in black men’s lung problems
Pulse oximeters deliver unreliable readings across ethnic groups
Pulse oximetry accuracy varies between race groups – US cohort study