Piotr Ladyzynski, Poland
Nalecz Institute of Biocybernetics and Biomedical Engineering, Polish Academy of Sciences Department of Modeling and Supporting of Internal Organs FunctionsModerator of 1 Session
Presenter of 1 Presentation
Technologies and systems for meal counting and estimating prandial insulin dose
Abstract
Background and Aims
Effective insulin therapy in people with diabetes requires them to be able to estimate post prandial insulin doses compensating for consumed meals. The aim of this work is to review existing technologies and systems for automating meal counting and estimating prandial insulin doses.
Methods
Automatic bolus calculators estimate prandial insulin doses based on carbohydrate content in meals and a carbohydrate-to-insulin ratio provided by the user. In recent years, several systems designed to automate the counting of meals, i.e. estimation of energy and carbohydrate content (in some cases also protein and fat) in a meal have been developed. In these systems several different technologies have been used such as food image processing and analysis, voice description of meals followed by voice-to-text conversion and the meal content analysis, or monitoring food intake activities (e.g. chewing or swallowing) using signal processing from various sensors.
Results
It was demonstrated in a few pilot clinical trials that systems for meal counting based on the computer vision and food image processing techniques or the voice description of meals are feasible and applicable. Using such systems it is possible to calculate content of meals with high enough accuracy to estimate insulin doses that effectively control postprandial glycemia. Whereas systems monitoring food intake activities are at the earlier stage of development and distinguishing a large number of foods using such systems is unlikely.
Conclusions
New insights on the effect of dietary macronutrients on postprandial glycemia demonstrates that prandial insulin doses should account for carbohydrates, protein and fat content in meals. Some of the technologies and systems outlined above may facilitate an accurate meal counting and thus, contribute to improvement of the insulin therapy results.