Background and aim:
Modern smartphones strive in our society to the point where people depend on these devices for most tasks. MyDiabetes is a mobile application that helps type I diabetics with their daily records (glycaemia, carbohydrates, insulin, exercise, etc.), by providing advices to its users triggered and adapted to the registers entered.
Methods:
Advices are based on medical protocols and guidelines, and serve as a recall to the guidance given by the medical doctor, encompassing the medical expertise. We use general advices to suit the broad dia-betic population. Avoiding crisis can be accomplished by warning the user of misleading actions (“you exercised, perhaps you should reduce insulin intake for this meal”), and by explaining the crisis’ possi-ble causes (“is your hypoglycaemia related with extra activity?”). This allows introspective analysis, en-abling behaviour adaptations or consulting a doctor with added information.
Registers can also be mined to unveil user patterns (“you have low glycaemia values on Tuesdays”) and create specific advices that combine the medical protocols with the data mined.
Results:
We are incorporating the advice system in the existent MyDiabetes Android application. We used it to collect registers from 5 diabetic patients (from 15 volunteers) for initial data mining. Currently, it was possible to derive general patterns as “on Sundays’ mornings with heavy breakfast, hyperglycaemia is common”.
Conclusions:
One current problem is the lack of data, the addition of feedback based on these patterns and the in-clusion of the advice system aims to encourage and benefit patients driving their usage of the platform.