Instant Heart Rate

2020: Release of the application

In the middle of August, 2020 researchers from the University of California provided a new method of detection of diabetes to San Francisco. They used the public Instant Heart Rate application from Azumio company for registration of measurements of a photoplethysmography to which added an algorithm of deep learning.

When the user places a finger-tip over a small lamp and the phone camera, the application carries out a photoplethysmography, fixing change of color of a finger-tip at each blow of heart. Researchers found out that they the frequency of serdtsebiyeniye, along with variability of a warm rhythm and some morphological features of a wave form at a photoplethysmography, correlate with many physiological indicators which diabetes, such as aging an endoteliya and atherosclerosis of microvessels influences. Therefore they solved conducted a research and to learn whether it is possible to use a photoplethysmography as an exact biomarker for detection of diabetes.

The application for the smartphone revealing diabetes is released

Researchers developed and checked an algorithm of deep learning, using nearly 3 million records of a photoplethysmography from 53,870 patients. The algorithm managed to define correctly presence of diabetes at 82% of sick patients and also to confirm its absence at 97% of patients without diabetes. The efficiency of an algorithm grew even more when using optional data of the patient, such as age, sex, BMI and also information on associated diseases. The additional analysis even showed correlation between assessment of an algorithm and A1c of examinees.^[1]

Additional researches should show how it is possible to use in the best way this new digital biomarker, but the received results already look promising.^[2]

Notes