A Scoping Review of Depression Risk Prediction Models for Diabetes Patients

Diabetes is a metabolic disorder characterized by chronic hyperglycemia. By 2024, the global prevalence had reached 540 million, with China accounting for 140 million cases—the highest number worldwide . Under the long-term burden of disease management, individuals with diabetes are prone to a range of mental health issues, with depression being one of the most common psychological concerns. Research indicates that the risk of depression among diabetic patients is twice that of non-diabetics, with a prevalence as high as 35%. Diabetic patients with comorbid depression exhibit poorer glycemic control, reduced treatment adherence, and diminished quality of life, alongside heightened risks of complications, mortality, and suicide . A complex bidirectional relationship exists between depression and diabetes. Compounded by its insidious nature, this makes depression difficult for clinicians to identify, often leading to missed opportunities for early intervention . Predictive models, statistical frameworks that assess disease risk based on multiple variables, offer advantages in early detection and convenience. However, existing diabetes depression risk prediction models vary significantly in their construction methods, definitions of depression, and included predictors. Their quality is inconsistent, lacking a unified evaluation system, and only a handful are truly applied in clinical practice. Therefore, this study conducts a scoping review of the construction methods, predictors, presentation formats, and model performance of depression risk prediction models for diabetic patients. The aim is to provide a reference for clinical healthcare providers to select more optimal risk prediction tools.