This study was conducted from September to December 2023. Six cities—Changzhi, Taizhou, Jishou, Nanchang, Suzhou, and Xianyang—were selected based on China's six major administrative divisions. A stratified cluster random sampling method was used to conduct field tests and surveys in these six cities. To ensure a balanced gender and age distribution of the sample, a total of 18 secondary schools were selected across various cities. Using a random lottery method, 216 classes were selected from each grade, and all students were included in their respective classes. Prior to the formal investigation, each school will organize an offline briefing session, where testers will explain in detail to participants the purpose, process, potential risks, and data confidentiality mechanisms of the study. Both students and their guardians will be present and will voluntarily sign a written informed consent form. All questionnaires were in paper form, distributed on site by test administrators who explained how to fill them out, supervised the completion process, and collected them centrally. The inclusion criteria for this study were: students aged 13–18 who were currently enrolled in school, and both the participant and their guardian consented to participate in the study. Exclusion criteria included: participants with severe physical or mental health conditions that could potentially influence study outcomes, as well as those who completed the questionnaire incompletely or had missing key data. Comparisons of demographic characteristics between included and excluded participants showed no significant differences, supporting the assumption of random complete non-response. Statistical power analysis was conducted using R Core Team tools, determining the target sample size to be 954. A total of 6,500 questionnaires were distributed across six cities, with 5,713 valid responses, yielding a response rate of 87.89%. Among these, 2,908 were male and 2,805 were female, with an average age of 15.11 ± 1.70 years. This study was approved by the Human Experiment Ethics Committee of East China Normal University (HR761-2022). All participants and their guardians were informed of the study's purpose and signed informed consent forms. All procedures adhered to current ethical guidelines for research on adolescent mental health and screen time, and all data were anonymized and de-identified (see Figure 2).

The study controlled for gender and age, taking into account the impact of demographic variables such as gender and age on the analysis results.

This study used Statistical Product and Service Solutions 26.0 statistical software to process and analyze the data. First, the software was used to check whether the research data contained common method bias. A critical value of less than 40% indicates that the study did not contain common method bias. Second, descriptive statistics and correlation analysis were performed on the relevant variables, using the PROCESS SPSS macro model compiled by Hayes (42) to examine the mediating effect of BMI and waist circumference on the relationship between screen time and adolescent mental health, and Model 7 to examine the moderating effect of sleep quality on the relationship between screen time and adolescent BMI and waist circumference. Controlling for gender and age, screen time as the independent variable, mental health as the dependent variable, BMI and waist circumference as mediating variables, and sleep quality as the moderating variable. To test the research hypothesis, the Bootstrap resampling method was used, with the number of resampling set to 5,000. If the 95% confidence interval (CI) did not include zero, the moderating mediating effect was determined to be statistically significant.

Research data collected through questionnaires are susceptible to common method bias. Therefore, Harman's single-factor test was used to examine common method bias. The analysis results showed that eight factors met the criterion of eigenvalue greater than 1. Without principal component factor rotation, the first factor accounted for 26.39% of the variance, which was below the critical value of 40% (43). Therefore, the analysis results indicate that there is no common method bias in this study.

Table 1 shows the analysis of 5,713 adolescents, including 2,908 males (50.90%) and 2,805 females (49.09%). Significant differences were observed between adolescents of different genders in terms of screen time, BMI, waist circumference, mental health scores, and sleep quality scores (p < 0.001). Significant differences were also observed among adolescents aged 13–18 years in terms of screen time, BMI, waist circumference, mental health scores, and sleep quality scores (p < 0.05).

Table 2 Pearson correlation coefficients between relevant variables. Screen time was significantly negatively correlated with adolescent mental health (r = −0.10, p < 0.01) and significantly positively correlated with BMI (r = 0.03, p < 0.05). The PSQI total score was significantly positively correlated with BMI (r = 0.18, p < 0.001). Adolescent BMI was significantly negatively correlated with mental health (r = −0.14, p < 0.001). Adolescent waist circumference was significantly negatively correlated with mental health (r = 0.04, p < 0.01), and waist circumference was not correlated with PSQI total score or screen time (r = 0.01, r = −0.005, p values both > 0.05).

After controlling for covariates, BMI was included as a mediator variable, and sleep quality was included as a moderator variable. Screen time had a significant negative effect on adolescents' mental health [β = −0.20, SE = 0.07, p < 0.01, 95% CI = (−0.34, −0.06)], with an effect size ranging from small to moderate. Additionally, BMI partially mediated the relationship between screen time and mental health [β = −0.89, SE = 0.12, p < 0.001, 95% CI = (−1.11, −0.66)], with a large effect size. Sleep quality moderated the relationship between adolescents' screen time and BMI [β = 0.04, SE = 0.01, p < 0.001, 95% CI = (0.02, 0.07)], with a small effect size. Statistical results are presented in Table 5, structural paths in Figure 4, and moderation slope plots in Figure 5 for detailed information.

After controlling for covariates, waist circumference acted as a mediating variable, as shown in Figure 3. Screen time had a significant negative effect on adolescents' mental health [β = −0.93, SE = 0.12, p < 0.001, 95% CI = (−1.16, −0.70)], with a large effect size; Waist circumference has a significant negative effect on adolescents' mental health [β = −0.03, SE = 0.08, p < 0.001, 95% CI = (−0.05, −0.02)], with a small effect size. Screen time does not have a significant effect on waist circumference [β = 0.29, SE = 0.36, p > 0.05, 95% CI = (−0.41, 1.00)], and waist circumference does not have a significant moderating effect on the relationship between adolescents' screen time and mental health [β = −0.10, SE = 0.06, p > 0.05, 95% CI = (−0.23, 0.02)]. Statistical results are shown in Table 5, and structural paths are shown in Figure 4 for detailed information.

Additionally, to further explore the moderating effect of sleep quality on the relationship between screen time and BMI, this study used simple slope analysis to examine the moderating effects of different sleep qualities (low and high). For adolescents with low sleep quality, screen time had a significant positive effect on BMI [β_simple slope = 0.12, SE = 0.05, p < 0.05, 95% CI = (0.03, 0.21)], with a small effect size. For adolescents with high sleep quality, screen time had a significant negative effect on BMI [β_simple slope = −0.11, SE = 0.05, p < 0.05, 95% CI = (−0.2166, −0.0104)], with a small effect size. Statistical results are presented in Table 6, and the slope plots are shown in Figure 5 for further details.

Research has found that screen time is negatively correlated with mental health, a finding consistent with recent longitudinal evidence from other studies. For example, Santos et al. found that social media use is negatively correlated with mental health and pointed out that girls are at higher risk of depression (10). Although our study is cross-sectional and cannot prove causality, evidence from intervention studies supports our findings. For example, a two-week intervention study showed that reducing screen use effectively improved the mental health of children and adolescents (44), providing strong causal evidence for the association between screen time and mental health. However, different studies have varying definitions and measurement methods for screen time and mental health, which may also be a reason for the inconsistent results. For example, some studies define screen time as the total time spent using all electronic devices (45), while others focus only on specific activities, such as social media or video games (46). This study used self-report questionnaires, which may be subject to recall bias. In addition, a meta-analysis found a weak but significant correlation between screen time and internalizing and externalizing behavior problems in children (45). In summary, despite methodological differences, the findings of this study are consistent with the mainstream literature, namely that excessive screen time has a potentially negative impact on adolescents' mental health.

The negative correlation between screen time and mental health found in this study can be further understood from the perspective of Bronfenbrenner's ecological systems theory (47). This theory emphasizes that individual development is influenced by multiple environmental systems at different levels. At the microsystem level, prolonged screen use may reduce face-to-face interaction between adolescents and their families and peers, thereby weakening important social support (48). At the mesosystem level, excessive engagement in online gaming or social media may impair sleep quality, which in turn affects academic performance and may lead to more profound mental health issues (49). At the broader external and macro system levels, parental work stress (50) and family economic status (51) and the widespread acceptance of technology and social media in mainstream society (52) may indirectly influence adolescents' screen usage patterns. The impact of screen time on adolescent mental health is not a single-path phenomenon, but rather the result of complex interactions and interplay across multiple levels, including individual, family, peer, school, and sociocultural factors. Our research findings align with this theoretical framework, indicating that to fully understand and effectively address this issue, it is necessary to adopt multi-level, systematic intervention measures rather than focusing solely on individual behavior.

This study found a positive correlation between screen time and BMI. The use of electronic devices can help adolescents alleviate stress and anxiety, but prolonged screen time may lead to internet addiction (53) and is associated with sedentary behavior, where the body is in a low-energy state and lacks physical activity (54) and increase the risk of overweight and obesity (55). This study also found a positive correlation between BMI and mental health. Research indicates that a higher BMI may lead to negative body image perceptions among adolescents, which in turn may affect their self-esteem and social interactions, thereby exacerbating mental health issues (56). Body image is closely related to self-esteem, especially in early adolescence, when this relationship is particularly evident (57). At this stage, adolescents are undergoing rapid physical changes and seeking social acceptance. Low self-esteem and psychological distress are often significantly related to dissatisfaction with body shape (58). A negative body image not only affects adolescents' self-confidence, but can also lead to social media addiction, further exacerbating mental health issues such as anxiety and depression (58).

This study found that BMI mediates the relationship between screen time and mental health in adolescents, consistent with similar findings in previous studies (59). Based on social cognitive theory and self-determination theory, we can gain a deeper understanding of the psychological mechanisms underlying this phenomenon. Social cognitive theory emphasizes the interdependent relationship between individuals, behavior, and the environment (60). Increased screen time is associated with sedentary behavior, often accompanied by reduced physical activity and insufficient energy expenditure, leading to an increase in BMI (61). A high BMI further affects adolescents' physical self-efficacy and body image perception (62), making them more prone to negative emotions and psychological distress (63). Therefore, BMI is not only a physiological indicator, but also indirectly affects mental health through cognitive and emotional mechanisms. Secondly, self-determination theory suggests that an individual's mental health depends on the fulfillment of basic psychological needs (autonomy, competence, and relatedness) (64). Excessive screen use often limits physical activity and real-life social opportunities, thereby weakening adolescents' sense of competence and social connection experienced through physical activity (65). An increase in BMI may further reduce motivation to participate in physical activity and social occasions, making it more difficult to satisfy basic psychological needs and ultimately having a negative impact on mental health (66).

In summary, screen time has a direct negative impact on adolescents' mental health and indirectly affects their mental health through BMI. BMI plays a partial mediating role in this relationship, but the indirect effect is small, indicating that its impact is relatively limited. Nevertheless, even small effects can have significant implications for public health in large sample populations. Previous studies have shown that excessive screen time among adolescents is associated with an increased risk of obesity (19), leading to higher BMI among adolescents. Adolescents with higher BMIs are more likely to experience social withdrawal (67), social isolation (68), bullying (69), and humiliation (70). Adolescents with higher BMIs may develop mental health issues. Therefore, the results of this study confirm Hypothesis 2, namely that BMI plays a partial mediating role in the relationship between screen time and mental health in adolescents.

However, this study found that waist circumference did not mediate the relationship between screen time and mental health. There was no significant correlation between adolescents' screen time and waist circumference, which is inconsistent with the findings of Padmapriya et al. (22), who found that longer screen time in adolescents was associated with higher levels of abdominal obesity. In addition, a meta-analysis showed that there was no significant association between screen time and abdominal obesity in adolescents. Although the waist circumference of the high screen time group was significantly higher than that of the low screen time group, the difference was small (71). The impact of screen time on waist circumference is not fixed and may be influenced by various factors, such as diet and sleep (72). In contrast, BMI was significantly associated with screen time and mental health in this study. This may be because BMI reflects the overall balance between energy intake and expenditure, which is more directly influenced by diet and physical activity. Waist circumference, as an indicator that concentrates on abdominal fat distribution, is often more strongly influenced by genetic, hormonal, and metabolic factors (73, 74). During the early developmental stage of adolescence, energy allocation mechanisms tend to favor overall weight gain rather than localized fat deposition. During this stage, growth hormone is active and subcutaneous fat storage capacity is strong, while the biological mechanism for visceral fat accumulation has not yet been fully activated (75, 76). This may explain why screen time is significantly positively associated with BMI, but not statistically significantly associated with waist circumference. Therefore, in short-term or cross-sectional studies, the association with screen time may not be as obvious as that with BMI. Future studies should combine longitudinal designs and consider dietary, sleep, and metabolic factors to further explore the differential roles of BMI and waist circumference in the mechanisms by which screen time affects mental health.

This study found that sleep quality moderates the relationship between screen time and BMI, which is consistent with Hypothesis 3. Sleep quality not only positively predicts adolescents' BMI, but also moderates the effect of screen time on BMI. Specifically, as shown in Figure 4, poor sleep quality in adolescents exacerbates the negative impact of screen time on BMI, and excessive screen time in adolescents with poor sleep quality may increase BMI. This result is consistent with Danil's findings: screen time has a more significant impact on BMI among adolescents who sleep less (77). Although this study did not directly examine sleep duration, sleep quality and sleep duration are closely related. Research shows that shorter sleep duration negatively affects an individual's sleep quality, leading to difficulty falling asleep or frequent waking during the night (78). Conversely, poor sleep quality can also make it difficult for individuals to get adequate sleep duration (79). According to circadian rhythm theory (80), human metabolism and appetite control processes are regulated by biological clock rhythms. Adolescents with poor sleep quality may experience circadian rhythm disorders due to reduced melatonin secretion, which can lead to metabolic abnormalities, increased appetite, and hunger, ultimately resulting in weight gain (81). If teenagers stare at electronic screens for long periods of time, especially at night, the blue light emitted by the screens may suppress melatonin secretion (82). Reduced melatonin secretion is associated with insulin resistance, high uric acid levels, and elevated triglyceride levels, which may lead to obesity in adolescents (83). In addition, reduced melatonin secretion may interfere with the breakdown and synthesis of adipose tissue, thereby exacerbating obesity (84). Another explanation is that poor sleep quality may be a source of stress (56), which could exacerbate overeating caused by stress (85). Lack of sleep can cause fatigue in adolescents (86), which in turn reduces physical activity levels and leads to overweight and obesity. In addition, poor sleep quality may trigger increased appetite (87) and energy metabolism disorders (88), further leading to an increase in BMI. However, if an individual has high-quality sleep and a stable circadian rhythm, their physiological functions are more likely to remain balanced (89), thereby mitigating the negative impact of screen time on BMI. Sleep quality mediates the relationship between screen time and BMI. Although the effect size is small, the results suggest that good sleep quality can partially buffer the adverse effects of screen time on BMI. For adolescents with poor sleep quality or long screen time, interventions to improve sleep should be prioritized to reduce their risk of obesity.

This study has both theoretical and practical significance. The theoretical significance of this study lies in providing a theoretical framework for the interaction of multiple factors. The research findings contribute to understanding the causes of mental health, reveal the complexity of mental health issues, and offer a new perspective on understanding adolescent mental health. Additionally, the findings of this study provide guidance for prevention and intervention strategies and serve as a reference for policy-making. The practical significance of preventing and intervening in mental health issues is as follows: first, adolescents should use electronic devices reasonably to reduce dependence and excessive use. Second, adolescents should increase their physical activity levels; if adolescents spend excessive time on screens, we recommend that they engage in at least 70 min of moderate-to-vigorous physical activity daily to reduce body fat (90). Finally, adolescents should establish regular sleep patterns to ensure adequate sleep duration and quality, thereby lowering the risk of mental health issues.

However, this study has certain limitations. First, this is a cross-sectional study and cannot explain the causal relationships between variables. To address this limitation, future studies should adopt longitudinal or experimental designs to better understand the causal pathways and dynamic interactions of these variables over time. Such studies will enhance the robustness and applicability of research findings and provide stronger evidence for developing targeted intervention measures. Second, sleep quality, screen time, and mental health were measured through questionnaire surveys, and the reliability of self-reported data may influence the results. Future studies should incorporate more objective measurement methods (e.g., wearable devices, digital tracking data) to enhance the validity of the results. Third, the study did not distinguish between different types of screens, making it impossible to determine which type of screen use (e.g., social media, video games, educational software, etc.) affects adolescents' mental health and obesity. Future studies should further refine the types of screen use and explore the mechanisms by which different types of screens affect adolescents' mental health and obesity, thereby providing a basis for developing more precise intervention measures. Fourth, this study did not fully account for the nested structure of the data. Future research should employ methods such as multilevel models to more comprehensively analyze the relationships among these variables.