Construcción de un modelo de identificación temprana de depresión posparto basado en tecnología de salud móvil y evaluación de sus efectos de aplicación en visitas comunitarias posparto

This study developed a mobile health-based early identification model for postpartum depression (PPD) and evaluated its effectiveness in community postpartum visits. A randomized controlled trial included 102 postpartum women (intervention: n=54; control: n=48). The intervention group used the "Maternal Love Guardian" app, integrating clinical risk factors and digital phenotyping (behavioral, emotional, cognitive, and voice data) to stratify PPD risk. Primary outcomes included detection rates, model performance, and intervention adherence. The model achieved 90.0% sensitivity and 84.1% specificity (AUC=0.871) at 3 weeks postpartum. The intervention group had significantly shorter PPD identification time (11.8 vs. 26.9 days, P<0.001) and higher early intervention rates (25.9% vs. 8.3%, P=0.019). Digital phenotyping revealed key differences in sleep, step count, social activity, and speech rate in PPD cases. At 12 weeks, symptom improvement was faster (4.3 vs. 6.9 weeks, P=0.009), and mother-infant interaction scores improved (77.8 vs. 72.1, P=0.033). Community providers reported 23% reduced consultation time (P<0.001). Cost-effectiveness was favorable (8,924 RMB per QALY). The mobile health model enhances PPD detection, accelerates intervention, and improves outcomes efficiently.