A smartphone app may be a non-invasive method to identify people with diabetes by detecting vascular changes, a new study suggests.
The method works by using the phone’s flashlight and camera to apply an optical technique called photoplethysmography (PPG). When a person’s finger goes over the light and camera, the app captures color changes in the fingertip that corresponds with each heartbeat and changes in blood volume. PPG is already used clinically to measure heart rate and oxygen saturation levels.
Researchers obtained nearly 3 million PPG recordings from more than 50,000 patients who reported having been diagnosed with diabetes by a health care provider. Researchers used the data to both develop and validate an algorithm to detect the presence of diabetes using the smartphone-measured PPG signals.
The algorithm correctly identified the presence of diabetes in up to 81 percent of patients in two separate datasets. Among the patients that the algorithm predicted did not have diabetes, the app confirmed that roughly 95 percent of patients did not have the disease across the validation datasets. The predictive performance improved further after accounting for other patient data such as age, gender, body mass index, and race/ethnicity.
The study, published in Nature Medicine, could one day provide a low-cost, in-home alternative to blood draws and clinic-based screening tools for diabetes, which affects more than 32 million Americans. The study was funded by NHLBI.