AI-Integrated Mobile Application for Diabetes Self-Management in Local Languages (Preprint)

UNSTRUCTURED Project Title AI-Integrated Mobile Application for Diabetes Self-Management in Local Languages Project Description I am developing an AI-integrated mobile medical application designed to help people with diabetes safely manage and adjust their insulin therapy. The application will be available in Amharic and English, making it accessible to patients in Ethiopia and other African settings where local-language digital health tools are limited. The app, called “Patient Diabetic Insulin Guide,” will support patients in daily insulin self-management through guided calculations, voice input, and structured health records. Key Features Insulin Dose Guidance The application will help patients calculate appropriate insulin doses for meals based on estimated carbohydrate intake and guide correction insulin when blood glucose levels are high. It will also analyze daily glucose records to support safe adjustment of basal insulin for the following day. Patient Monitoring and Records Patients will be able to record their blood glucose values, insulin doses, and meals within the application. By reviewing these records, the system can support safer day-to-day insulin adjustment and improve consistency in diabetes self-management. Locally Adapted Food Carbohydrate Guidance The application will include a culturally relevant reference for common Ethiopian foods with estimated carbohydrate content, allowing patients to more accurately estimate carbohydrate intake when calculating meal insulin doses. This localized food guidance will make insulin adjustment more practical and realistic for patients using their typical diets. Planned Innovation: Voice-Enabled AI Assistance The project aims to integrate audio-based machine learning functionality that works directly on mobile devices. Patients will be able to interact with the application using voice input in local languages. For example, patients can speak their blood glucose values, meals, or insulin information into the application. The system will convert speech into structured data and then generate insulin adjustment guidance based on those inputs. This voice-enabled approach can significantly improve accessibility for individuals with limited literacy, visual challenges, or difficulty using typed interfaces. Project Aim The primary aim of this project is to improve diabetes self-management and reduce diabetes-related complications by providing accessible, culturally adapted digital guidance. By supporting accurate insulin dose adjustments, encouraging consistent glucose monitoring, and providing locally relevant dietary guidance, the application seeks to reduce episodes of severe hyperglycemia and hypoglycemia, improve long-term glucose control, and ultimately decrease the risk of chronic diabetes complications such as kidney disease, neuropathy, retinopathy, and cardiovascular disease. Expected Impact This project aims to expand access to AI-supported diabetes management tools in local languages, empower patients to participate more actively in their care, and improve the safety of insulin use in resource-limited settings. By combining clinical guidance, culturally adapted food information, and voice-enabled interaction, the application can support better diabetes outcomes for patients in Ethiopia and potentially across other African regions.