Could we do it without pricking?
Around half a million people in Switzerland have diabetes and have to measure their blood sugar regularly to avoid dangerous hypoglycemia. So far, this meant taking blood samples or relying on a device attached to the body, which naturally raises the question of whether smartwatches and similar devices that many people use day and night anyway can be programmed to warn them against hypoglycemia. Yasmine Mohamed, a graduate of the programme, used health data obtained from patients' Apple Watches, glucose meters and diary entries to train machine learning models. She then designed algorithms that could provide non-invasive warnings about hypoglycemia. When integrated into smartwatches, such algorithms could offer people with diabetes a further tool to help them best manage their health in daily life.
