Does the physical activity–adolescent mental health model generalize across different social environments?

Insufficient physical activity, high prevalence of psychological problems, and inadequate sleep duration among adolescents are major public health concerns. To provide theoretical guidance for addressing these issues, Dong Rubao et al. proposed a holistic health-based model in which physical activity promotes adolescent mental health through the multiple mediating effects of general self-efficacy and sleep duration. If valid, this model could offer both theoretical and practical guidance for integrated solutions to insufficient physical activity, inadequate sleep, and poor mental health among adolescents. At the same time, it may maximize health promotion effects while saving time for adolescents with heavy academic workloads, and encourage the adoption of a holistic perspective in health promotion efforts. However, the original study was based only on cross-sectional survey data, leaving uncertainty about the model’s validity across different social contexts. Since adolescent physical activity, sleep duration, and mental health are influenced by diverse factors, particularly the physical activity environment, further validation is needed. This study aimed to examine the validity of the adolescent health promotion model across different social environments and to provide evidence for its practical application. From 2021 to 2023, we randomly extracted 1,073 sets of survey data per year from an adolescent health database containing measures of physical activity, sleep duration, general self-efficacy, and mental health. The data were collected in three distinct contexts: the normalized prevention and control stage of COVID-19 in 2021, the outbreak stage in 2022, and the post-control stage in 2023. These datasets were used to test the theoretical model constructed by Dong Rubao et al. Cross-sectional surveys conducted over three consecutive years provided data on adolescent level of physical activity (PAL), sleep duration (SD), general self-efficacy (GSE), and mental health (MH). Normality and correlation analyses were performed using SPSS 26.0. Bootstrap inference for model coefficients was conducted using PROCESS v4.1 by Andrew F. Hayes (Model 4). Effect size tests for the structural equation model were conducted in AMOS 26.0. The average sleep duration(SD) for adolescents was 440.15, 430.14, and 441.80 mins/day respectively (F = 9.62, p < 0.001), PAL was 2.54, 2.35, and 2.46 respectively (F = 18.87, p < 0.001), GSE was 2.65, 2.36, and 2.71 respectively (F = 74.53, p < 0.001), and MH was 1.54, 1.86, and 1.57 respectively (F = 87.31, < 0.001); ANOVA results showed significant differences. There were significantly correlated with each other among average of PAL, SD, GSE and MH. The total effects of the model were − 0.196, − 0.338, and − 0.221, respectively, all with P < 0.001; the direct effects were − 0.089, − 0.230, and − 0.117, respectively, all with P < 0.001; the total indirect effects were − 0.107, − 0.107, and − 0.104, respectively; with PAL→SD→MH was − 0.077, − 0.051, and − 0.063, and PAL→GSE→MH was − 0.031, − 0.057, and − 0.041, all with a significant meaning in the Bias-Corrected 95%CI test. However, during the COVID-19 outbreak, few fit indices of the model did not reach ideal values. SD and GSE play multiple mediating roles in Physical and Mental Health Promotion Model for Adolescent. The adolescent health promotion model is generally valid across different social environments. However, in contexts lacking sufficient PA opportunities, the model shows relatively poorer fit. In practice, health promotion among adolescents should address insufficient PA, high prevalence of mental health problems, and inadequate sleep duration as interrelated issues, in order to achieve more comprehensive effects.