GlucoseGo Kids: Machine-learning derived tools for supporting exercise in children with type 1 diabetes (GlucoseGo Kids)

Research summary

Exercising regularly and is an essential part of managing Type 1 Diabetes, as it improves blood sugar control, reduces insulin requirements and risk of complications of diabetes and also improves cardiovascular and mental health. Despite this, very few people with type 1 diabetes exercise regularly and thus do not get these benefits. Managing glucose levels within safe ranges when exercising remains challenging, particularly for children. Fear of having hypoglycaemic episode (where glucose levels drop dangerously low) is a key barrier to exercise for children with type 1 diabetes. We have developed a simple visual tool to help people with type 1 diabetes understand the likelihood of having a hypoglycaemic episode based on their glucose levels and how long they plan to exercise for. This is called 'GlucoseGo'. This is based on a prediction model which was developed using information over 16000 exercise bouts from people with type 1 diabetes, and which can predict likelihood of hypoglycaemia very accurately. However the data that we used to develop this prediction model was largely from adults, and some children over 12 years old doing very structured exercise. In order to make sure GlucoseGo can be used by all people with type 1 diabetes we need to check that it works as well for children as it does for adults. In this study we will collect data on exercise using a simple self-report diary in children (under 18) with type 1 diabetes, and also ask them to share with us data from the continuous glucose monitor that they wear as part of their usual care. We can use these two sources of information to test whether GlucoseGo can assess likelihood of hypoglycaemia around exercise for children as well as it can in adults.

Principal Investigator

Ms Anne Marie Frohock

Contact us

Email: childrensresearch@ouh.nhs.uk

IRAS number

351880