Mobile Platform for Personalized Nutritional Recommendations Using a Machine-Learning Cloud Service for Patients with Chronic Diseases

Diaz V.; Carrasco F.; Solis R. · Proceedings - 2025 13th International Conference in Software Engineering Research and Innovation, CONISOFT 2025 (2025) · DOI: 10.1109/conisoft66928.2025.00043

Resumen

Managing personalized nutrition for adults with chronic conditions remains a critical challenge in primary care. Currently, nutrition apps lack the ability to tailor recommendations to specific medical needs. This study presents MottiNut, a mobile application that integrates artificial intelligence and cloud-based machine learning to deliver customized dietary guidance for patients with chronic diseases such as type 2 diabetes, hypertension, and cardiovascular conditions. The system was deployed and clinically validated at a public health center in Lima, Peru. Over a 12 -week period, patients using MottiNut demonstrated significant clinical improvements: fasting glucose levels decreased by 23.3 mg/dL, glycated hemoglobin (HbA1c) was reduced by 1.2 %, LDL cholesterol decreased by 14.1 mg dL, and triglycerides dropped by 28.7 mg/dL. Additionally, patients experienced a 4.8 kg average weight loss and a 13.9 mmHg reduction in systolic blood pressure. The application also achieved a high System Usability Scale (SUS) score of 78.4 and a 91.7 % retention rate. Adherence to nutritional plans improved by 35 %, based on digital logs and MARS-5 assessments. This AI-powered solution enhances patient engagement through intuitive interfaces and dynamic, personalized recommendations while supporting nutritionists in delivering more efficient, individualized care.

Instituciones: Universidad Peruana de Ciencias Aplicadas