The association between obesity and problematic smartphone use among school-age children and adolescents: a cross-sectional study in Shanghai

Background The study aimed to investigate the association between content-based problematic smartphone use and obesity in school-age children and adolescents, including variations in the association by educational stage and sex. Methods Two-stage non-probability sampling was used to recruit 8419 participants from nineteen primary schools, five middle schools, and thirteen high schools in Shanghai in December 2017. Obesity was identified by body mass index (BMI), which was obtained from the school physical examination record, while problematic smartphone use was measured by the Revised Problematic Smartphone Use Classification Scale as the independent variable. Results The rates of obesity varied with educational stages, while problematic smartphone use increased with educational stages. Male students reported higher obesity rates (37.1%vs19.4%, P < 0.001) and greater problematic smartphone use scores (25.65 ± 10.37 vs 22.88 ± 8.94, P < 0.001) than female students. Problematic smartphone use for entertainment (smartphone users addicted to entertainment games, music, videos, novels and other applications) was positively associated to obesity status for primary school [odds ratio (OR), 1.030; 95% confidence interval (95% CI), 1.005–1.057] and high school students (OR, 1.031; 95% CI, 1.004–1.059). For female students, problematic smartphone use for entertainment was positively associated with obesity status (OR, 1.046; 95% CI, 1.018–1.075). Conclusions Problematic smartphone use may be associated with obesity in children and adolescents. The association differed based on the educational stage and sex, and the difference possessed dimensional specificity. Supplementary Information The online version contains supplementary material available at 10.1186/s12889-021-12124-6.


Problematic Smartphone Use Classification Scale (PSUCS)
showed the items of PSUCS, which were adopted to measure problematic smartphone use (PSU) among Chinese middle school students. The RPSUCS for Chinese school-age children and adolescents was developed based on PSUCS. games, watching videos, reading novels, etc. 6. Using entertainment apps is my favorite way to relax and reduce stress. 7. I always keep up with updates from online novels and/or games. 8. I spend a lot of time on mobile entertainment. 9. I have a strong sense of achievement and satisfaction after completing a game, video series or novel. PSU on compulsive behavior 10. I cannot resist the impulse to check for message updates in social apps (QQ, WeChat, Weibo, etc.). 11. I always reply to the news on social apps, even sneaking back during class, otherwise I will feel uneasy 12. Even thinking about using entertainment apps (gaming, video, novel, etc.) distracts me from other activities. PSU on information collection 13. I will read the news pushed by the mobile phone apps in the first time. 14. I spend a lot of time on a daily basis searching for and reading various news and information. 15. If I don't read news, I will feel restless.

Exploratory Factor Analysis (EFA)
EFA was conducted in the results of 8 419 students by Principal Component Analysis and Varimax Rotation. Items that met any of the following conditions would be deleted: (1) the items of which factor loads on all dimensions being less than 0.6; (2) the items of which factor loads being high on two or more dimensions while the differences of the factor loads being less than 0.1; (3) the items of which the number in one dimension being less than 3.
In the first EFA, the Kaiser-Meyer-Olkin (0.920) and Bartlett's tests (p<0.001) showed that the data was suitable for factor analysis. A three-factor structure was extracted from the first EFA. The factor loads of item 8 on the three extracted dimensions were 0.593, 0.553, and 0.055, while those of item 12 were 0.598, 0.217, and 0.309. Therefore, item 8 was deleted because it met the first and second conditions mentioned above, while item 12 was deleted according to the first condition.

Confirmatory Factor Analysis (CFA)
The proposed three-factor model of the RPSUCS was tested by computing CFA with AMOS 22.0. According to the modification indices, residuals e5 and e6 were linked to amend the model (Figure 1).

Internal consistency reliability
Internal consistency reliability was measured using the Cronbach α coefficient. The Cronbach's α coefficient of the total scale, PSU on social network, PSU on entertainment, and PSU on information collection respectively, were 0.876, 0.863, 0.852, and 0.750.

Test-retest reliability
A sample of 540 participants was retested to assess the test-retest reliability of RPSUCS two months later. Test-retest reliability was indicated by the interclass correlation coefficient (ICC) of total score and score of three dimensions at the two tests. The ICCs of total scale, PSU on social network, PSU on entertainment, and PSU on information collection were 0.763, 0.850, 0.645, and 0.609, respectively.