New technology such as artificial intelligence recently has been deployed in healthcare to generate faster and more nuanced diagnostic tools for help with conditions such as Alzheimer’s.
An ongoing issue for senior living and care providers, however, is that these tools often fail to reduce biases against racial and ethnic minorities. That often is because the knowledge that AI and other software “trains on” pulls from data sets that often lack sufficient data on those groups.
Approximately 25% of nursing home residents are non-white, according to a 2019 study. And only 12% of assisted living residents are not non-Hispanic white individuals, according to the federal government.
To help address this issue, a new tool that attaches to a smartphone is designed to screen for Alzheimer’s in a way that avoids typical racial biases.
The attachment accounts for racial disparities not by skin tone, but by eye color. That’s because pupil-based diagnosis, or pupillometry, is easy to generate via light and imaging, and such eye coloration often correlates with skin tone.
The device, which was developed by engineers at the University of San Diego, essentially is a filter that makes it easier for smartphone cameras to capture measurements of the pupil’s movements and dilations.
Beyond even removing explicit racial bias, the goal of the smartphone attachment is to make Alzheimer’s diagnostic tools accessible for everyone, the researchers stated, adding praise for smartphone technology.
The attachment is being designed to work on any smartphone, and the next phase of development will include making the device easier to use for older adults, the researchers said.
“The high plasticity of smartphone applications allows for continued ingenuity in providing new and different stimuli,” the study authors stated. “The smartphone can be programmed to provide a wide variety of visual or auditory stimuli for a pupil response task.”
Although the smartphone attachment is not commercially available yet, the researchers indicated that their prototype could be “improved and simplified for low-cost production.”
