Methodological evaluation of the psychosomatic index for the classification of bang-estimated diabetes risk in the peruvian population

Introduction: Type 2 diabetes mellitus is an increasing public health problem in Peru, and most risk estimation models emphasize metabolic factors while overlooking emotional components such as depression. This study aimed to evaluate the discriminative performance of the Psychosomatic Index (PSI) for classifying adult Peruvians according to the risk estimated by the Bang index, used as a reference model. Materials and methods: An analytical study using secondary data from the 2022–2024 Demographic and Family Health Surveys (ENDES) (n= 95,685). Records with complete information on blood pressure, anthropometric measures, and PHQ-9 scores were included. A composite index was constructed using factor analysis, and its discriminative performance was assessed through ROC curves (AUC), Youden index, sensitivity, specificity, and other performance metrics, in comparison with the Bang index. Results: Factor analysis identified three components: hemodynamic (systolic blood pressure), anthropometric (body mass index), and affective (PHQ-9), which constituted the PSI= (SBP × 0.915 + BMI × 0.971 + PHQ-9 × 0.914). The model showed acceptable statistical adequacy (KMO= 0.559; Bartlett’s test: p< 0.001). In women, the PSI achieved an AUC of 0.849, outperforming BMI, waist circumference, and PHQ-9. In men, the waist-to-height ratio showed the highest performance (AUC= 0.851). By age group, the PSI showed excellent performance in individuals ≤ 39 years (AUC= 0.916–0.993) and moderate performance in those ≥ 40 years (AUC= 0.737–0.836). Conclusions: The PSI showed discriminative performance for classifying adults according to Bang-index-estimated risk, among women and young adults, supporting its exploratory value as a cardiometabolic–psychoemotional tool requiring validation against biochemical standards future.