A Personalized Adviser-Interpreter System for Monitoring Diabetic Patients' Weight Loss

Diabetes and obesity are chronic diseases that need continuous follow-up. Analysis of the large volumes of patient data collected over time is time consuming and effortful for the physician. Automatic data abstraction and providing a summary of important events in the patient's follow-up history saves the physician's time and helps in making more informed decisions. The purpose of this study was to design an interpreter system that generates a tailored textual report about patients' adherence to a therapeutic regimen. The system has been designed in three phases: determining report structure, interpreting data, and generating textual-graphical reports for the patient and physician. The output report was evaluated by 12 diabetes nutrition specialists and eight diabetic patients. We used a questionnaire and a semi-structured interview approach to assess the quality of the report in terms of content validity, understandability, and practicality.