Danilo Mandic and Imperial College colleagues have developed a novel in-ear PPG device for continuous blood glucose level measurement, using the infrared wavelength of a pulse oximeter.
In a recent proof of concept study, non-diabetic, pre-diabetic, type I diabetic, and type II diabetic states were considered. Recordings spanned 9 days, in both fasting and post carbohydrate consumption states. Blood Glucose Levels from PPG were estimated using regression-based machine learning models. An average of 82% of the BGLs estimated from PPG were in region A of the Clarke error grid plot, and 100% of the estimated BGLs in clinically acceptable regions A and B. The researchers believe that this demonstrates the potential of the ear canal for non-invasive blood glucose monitoring.