Abstract
BACKGROUND: The promotion of healthy dietary behaviors in adolescence is critical, which have long-term implications for lifelong health. Integration is an important method for improving limited theories of dietary behavior change. The present study proposes an integrated model aimed at identifying the diverse determinants of healthy dietary behaviors in adolescents and assesses its stage-specific nature as the potential for effective interventions. METHODS: A two-year prospective study was conducted in southwest China, involving a sample of 1990 adolescents (mean age: 15.06 years) from two randomly selected secondary schools. A total of 765 adolescents were exposed to the three-phase interventions while 1225 adolescents participated as the control group. The interventions aimed to promote healthy dietary behaviors in adolescents, which included health education sessions and health handbooks targeting specific constructs of the integrated model. The follow-up surveys after corresponding interventions were performed every six months. Self-reported frequency of healthy dietary behaviors, psychological constructs integrated from Health Action Process Approach, Health Belief Model, and Theory of Planned Behavior, as well as anxiety symptoms were measured. RESULTS: Within structural equation modeling controlling for past behavior, the integrated model accounted for 61.7% variance of behavioral intentions and 19.1% variance of healthy dietary behaviors, and showed discontinuity patterns across behavior change phases. The pivotal constructs included outcome expectancies, perceived severity, subjective norms, action self-efficacy, behavioral intentions, action planning, and maintenance self-efficacy. Anxiety was an emotional barrier in the dietary behavior modification (β = -0.113, P < .001). Interventions within the staged integrated model, led to increased adoption of healthy dietary behaviors in intervention group compared to control group (33.40% vs. 25.70%, P < .05), indicating certain effectiveness, particularly in targeting action self-efficacy (stronger direct effects on behavioral intentions and maintenance self-efficacy, β = 0.489 to 0.704, P < .001). CONCLUSIONS: The staged integrated model provides a detailed understanding of the determinants of healthy dietary behaviors in adolescence, highlighting anxiety as an emotional barrier that impedes positive cognition and healthy dietary behaviors. It provides valuable guidance for future interventions targeting specific constructs across behavior change phases, with particular emphasis on enhancing action self-efficacy.