Respiratory ailments such as COVID-19 hamper the body's ability to draw oxygen from the lungs, which is why patients' blood oxygen levels often need to be checked. New research now suggests that people could track those levels at home, using their smartphone.
Ordinarily, blood oxygen saturation is monitored utilizing a device known as a pulse oximeter, which is clipped onto either a finger or the ear lobe. These tools are usually applied and monitored by trained technicians in hospitals or clinics. Smartphone-connected oximeters do exist, but they're one more item that home users would have to buy.
Seeking a simpler and less expensive alternative, scientists from the University of Washington and the University of California - San Diego developed an experimental app. They began with six volunteers aged 20 to 34 – three male, three female, one of whom was African American and the rest of whom were caucasian.
To train the deep-learning-based algorithm utilized by the app, four of those people wore a standard oximeter on one finger, and placed another finger on the same hand over a smartphone's camera lens and flash. Over a 15-minute period, they breathed a mixture of oxygen and nitrogen, to slowly reduce their blood oxygen levels.
As they did so, the phone recorded flash-lit video of the blood pulsing in and out of their finger. The app continuously analyzed how much of the light was absorbed by the blood, in the green, red and blue color channels. The amount that was absorbed varied with the amount of oxygen in their blood.
When blood oxygen readings obtained by the oximeter were compared to the light intensity measurements from the phone, it was possible to ascertain which specific readings corresponded to which specific light measurements. The app was therefore able to learn how much oxygen was present in the blood at any one time, based on how much light was being absorbed by the blood.
When the app was subsequently tested on all six volunteers, it was found to be capable of measuring blood oxygen levels down to 70%, which is the lowest value that commercial oximeters are required to measure. By contrast, a healthy person should have blood oxygen levels of at least 95%, and typically require medical attention if those levels drop below 90%.
In its current form, the app is able to accurately spot low levels 80% of the time. It is believed that this number should improve significantly as the technology is developed further, which will involve training the app on a much more extensive dataset.
"This way you could have multiple measurements with your own device at either no cost or low cost," said the University of Washington's Dr. Matthew Thompson, co-author of a paper on the research. "In an ideal world, this information could be seamlessly transmitted to a doctor's office. This would be really beneficial for telemedicine appointments or for triage nurses to be able to quickly determine whether patients need to go to the emergency department or if they can continue to rest at home."
The paper was recently published in the journal npj Digital Medicine.
Source: University of Washington
