Diabetic distress is a psychological condition commonly experienced by individuals with Type-2 Diabetes Mellitus (T2DM), significantly impacting their quality of life. Despite its high prevalence, diabetic distress is often underdiagnosed, which can lead to poor patient outcomes. This study aims to predict diabetic distress and emotional burden in T2DM patients by integrating multimodal data sources, including demographic information, laboratory results, and survey-based assessments. Various machine learning algorithms, such as CatBoost, LightGBM, Decision Tree, and Random Forest, were employed to train prediction models on this dataset. The integration of Explainable Artificial Intelligence (XAI) techniques ensures the transparency of predictions, allowing healthcare professionals to understand the reasoning behind model outputs. The findings of this study could enhance the recognition of diabetic distress and emotional burden in T2DM patients, offering valuable insights for clinical interventions aimed at improving patient well-being and disease management.

Keywords: Diabetic distress, emotional burden, Type-2 Diabetes Mellitus, machine learning, multimodal data, Explainable AI, CatBoost, LightGBM, Random Forest, Decision Tree, healthcare, patient well-being, emotional health.