Understanding the acute influence of lifestyle on daily glucose fluctuations
Using Continuous Glucose Monitoring and contextual data to increase insight in unhealthy glucose patterns for individuals with type 2 diabetes (Gluco-insight)
Together, TNO, LUMC, Roche, Ekomenu and Reinier Haga MDC will investigate individual causes of (un)healthy glucose patterns in type 2 diabetes (T2D) patients. Patients will be equipped with a continuous glucose monitor that measures their glucose levels while they continue daily life (observational phase) or do lifestyle challenges such as diet and exercise. Different analysis and modelling techniques will be used to identify individual relations between glucose profiles and lifestyle.
The outcome of this project will pave the way for more personalised lifestyle management support for T2D patients. Such a personalised approach is required as people differ considerably in their biological responses. Additionally, lifestyle modifications have been shown to be effective in a research setting in terms of improving diabetes-related health markers and medication reduction. Further, improving self-management and health of T2D patients not only increases their quality of live, it also supports the affordability of health care.
This project provides a first step towards self-management support in improving glucose patterns for people with T2D. The core of the project is an experimental study to acquire continuous glucose data and contextual data during observational and intervention phases. This data is used for analysing and modelling the influence of contextual variables on glucose profiles. As no golden standard exists, several techniques will be applied with the aim to explain individual (N-of-1) glucose patterns, and to identify individual predictors of (un)healthy glucose parameters. These methods include physiological modelling, Bayesian networks (N-of-1), and machine learning techniques. The models will be subjected to in-silicotests to look at their strengths and weaknesses and potential to be used in self-management support tools.
In the end this project will offer a rich dataset of factors affecting glucose patterns, as well as models explaining (parts of) the complexity of individual glucose patterns. These models could allow for more accurate, tailored lifestyle advice. Additionally, the in-silicotests will provide insight in the minimal required dataset for such tailored advice, which can contribute to improved personalized T2D management. Potentially, this project strengthens the business case for glucose pattern-based e-health applications.