Screen time, physical activity, and overweight/obesity among school-aged children and adolescents in Heilongjiang Province, China: a cross-sectional study

Introduction Childhood overweight and obesity are major public health concerns in China. The activity displacement hypothesis provides a plausible framework linking electronic-device use, physical activity, and weight status, but its applicability may vary across settings. This study examined overweight/obesity and explored whether physical-activity frequency formed a statistical pathway between daily electronic-device use and overweight/obesity among school-aged children and adolescents in Heilongjiang Province, China. Methods This cross-sectional secondary analysis used data collected between October 2024 and April 2025 from participating schools in Daqing, Heihe, and Yichun/Wangkui. Of 26,535 survey records, 14,095 had nonmissing values for the outcome, exposure, age, and prespecified covariates and were retained in the original analysis. The questionnaire recorded average total daily electronic-device use, exercise frequency, nighttime sleep duration, and parental education. Multivariable logistic regression and sex- and grade-stratified models were fitted. To harmonize exposure specification, additional logistic models treated daily electronic-device use continuously, with and without adjustment for physical-activity frequency; school-clustered robust standard errors were examined as a sensitivity analysis. An exploratory bootstrap decomposition with 1,000 iterations estimated direct and indirect statistical associations through physical-activity frequency. Results Among 14,095 participants, the prevalence of overweight/obesity was 38.2%. Male sex showed the strongest association (OR = 2.67, 95% CI: 2.48–2.87). A higher physical-activity frequency score (OR = 0.94) and a higher approximate sleep-duration score (OR = 0.96) were inversely associated with overweight/obesity. In the added continuous-exposure analysis, each additional reported hour/day was weakly associated before adjustment for physical activity (OR = 1.019, 95% CI: 1.002–1.035, p = 0.025) but not after adjustment (OR = 1.014, 95% CI: 0.998–1.031, p = 0.085); neither estimate was statistically significant with school-clustered robust standard errors. The exploratory decomposition yielded a direct probability difference of 0.0033 and an indirect probability difference of 0.0011 through physical-activity frequency; the descriptive indirect proportion was 24.35%. Conclusion The association between reported daily electronic-device use and overweight/obesity was small and sensitive to exposure specification, physical-activity adjustment, and variance estimation. The fitted models were compatible with a small indirect statistical association through physical-activity frequency, but the cross-sectional data do not establish temporal ordering or causal mediation. Longitudinal studies using validated measures are needed.