Your phone camera might have another useful feature beyond taking selfies.
Engineers at UC San Diego and the University of Washington have published a proof-of-concept study indicating that a smartphone is capable of detecting blood oxygen levels almost on par with the industry-standard medical device.
If you’ve ever gone into surgery, you’ve likely had a pocket-size pulse oximeter clamped onto your finger to track the percentage of oxygen in your red blood cells. This kind of oxygen monitoring was especially useful for triaging patients with acute respiratory conditions during the COVID-19 pandemic.
Edward Wang, a professor of electrical and computer engineering at UC San Diego and one of the lead authors of the study, said a big part of it is to see what is possible with the tools already available, like a smartphone.
“The biggest difference between a medical device, how it’s made and how the smartphone-based health monitor works is that we don’t have a perfect system and we have to use machine learning and a lot of data to sort of help us get to somewhere close to what would have been the perfect physical system, which is the medical device,” said Wang, who also is a faculty member of the university’s Design Lab.
For people recovering from COVID-19 or other conditions like asthma, which can make it harder for oxygen to get to the lungs, pulse oximeters can help them with regular observation at home. When a patient’s oxygen level drops to 90 percent or below, medical attention is needed. A healthy person has an oxygen level of about 95 percent.
In the study, the researchers used a smartphone’s camera — with the flash on — to take a video of a person’s blood flow on one finger and track it with a pulse oximeter on another finger. The smartphone technique uses light to help capture the amount of oxygen in the blood.
“The camera records how much that blood absorbs the light from the flash in each of the three color channels it measures: red, green and blue,” Wang said.
As oxygen binds to hemoglobin, it changes the saturation of oxygen in the blood, and that translates to a change in the color of the blood.
From there the researchers used a machine learning algorithm that tracks the pattern of changes at different oxygen levels. During a 15-minute period, each participant breathed in a controlled mixture of oxygen and nitrogen to slowly reduce oxygen levels to simulate the changes.
The smartphone used in the study was able to detect a blood oxygen level as low as 70 percent, the lowest value that pulse oximeters should be able to measure, according to the U.S. Food and Drug Administration.
The study tested the smartphone technique on six people ages 20-34. The participants were evenly divided among those who identify as male or female. One participant identified as African American, while the rest identified as Caucasian.
Variables such as skin tone or muscle density can affect the results of a pulse oximeter reading, so there is more work to be done on the technology and machine learning to account for them, Wang said.
The researchers noted it will be awhile before the smartphone technique can be widely used in households or clinics via an app. However, part of what they are doing is showing how these kinds of tools can be made more …….