Network analysis of exercise self-efficacy and psychosocial factors in patients with coronary heart disease after percutaneous coronary intervention: a cross-sectional study

Exercise self-efficacy is an important psychological factor associated with exercise-related behaviors in patients with coronary heart disease after percutaneous coronary intervention. However, the structural relationships between exercise self-efficacy and related psychosocial factors remain unclear. This study aimed to examine the network structure of exercise self-efficacy and its psychosocial correlates using item-level and multidimensional network models. A cross-sectional study was conducted among 340 patients with coronary heart disease after percutaneous coronary intervention in a tertiary hospital in China. Exercise self-efficacy, kinesiophobia, heart-focused anxiety, and perceived social support were assessed using validated scales. A dual-network framework was constructed to characterize both the internal structure of exercise self-efficacy and its broader psychosocial associations. An item-level network based on nine ordinal self-efficacy items and a multidimensional network integrating 13 psychosocial dimensions were estimated using Gaussian graphical models with EBICglasso. Expected influence was used to identify central components within the exercise self-efficacy network, and targeted bridge expected influence was applied to identify psychosocial factors showing stronger associations with the exercise self-efficacy community. Network accuracy, stability of assessed centrality measures, and robustness were evaluated using bootstrap procedures and sensitivity analyses. In the item-level network, “Include exercise in your daily routine” showed a relatively larger expected influence estimate within the exercise self-efficacy network. In the multidimensional network, the nodes “Avoidance of exercise,” “Avoidance,” and “Friend support” showed relatively larger absolute targeted bridge expected influence estimates, indicating stronger observed associations with the exercise self-efficacy community. Bootstrap and sensitivity analyses indicated acceptable edge-weight accuracy, good stability of the assessed centrality measures, and robust primary findings after the exclusion of extreme observations. These findings suggest that exercise self-efficacy after percutaneous coronary intervention is characterized by an interconnected structure involving both its internal components and its associations with related psychosocial factors, rather than functioning as an isolated construct. The dual-network framework provides a comprehensive perspective for understanding the structural organization of exercise self-efficacy and its psychosocial correlates and may inform future hypothesis-driven longitudinal research.