Authors: Jie Meng, Xiaofan Ma, Genglin Yang, Jing Liu, Xiaoting Du, Xiaoyan Gao, Xue Li
Categories: Research, Adolescents, Sleep procrastination, Eating behavior, Depressive symptoms
Source: BMC Psychology
Authors: Jie Meng, Xiaofan Ma, Genglin Yang, Jing Liu, Xiaoting Du, Xiaoyan Gao, Xue Li
Depressive symptoms are a common mental health concern among adolescents, with prevalence steadily increasing and showing a trend toward younger age groups. This cross-sectional study aimed to examine the relationship between sleep procrastination and depressive symptoms among adolescents, with a particular focus on the mediating role of eating behavior. Stratified cluster sampling was used to select participants, who completed a battery of questionnaires, including the General Circumstances Investigation, Sleep Procrastination Scale (SPS), Patient Health Questionnaire-9 (PHQ-9), and Sakata Eating Behavior Scale Short Form. Logistic regression and restricted cubic spline models were employed to examine the relationship between sleep procrastination and depressive symptoms, and a mediation model was applied to analyze the indirect effect of eating behavior.The results indicated that sleep procrastination was positively associated with depressive symptoms, exhibiting a significant nonlinear association (p < 0.05). Mediation analysis revealed a significant indirect effect of sleep procrastination on depressive symptoms via eating behavior (β = 0.097, 95% CI [0.058, 0.143]), and this indirect association was not significantly moderated by sex. In conclusion, the indirect effect of sleep procrastination on depressive symptoms via eating behavior was statistically significant, consistent with a partial mediation model. These findings suggest that sleep procrastination is associated with depressive symptoms both directly and indirectly via eating behavior.
Adolescent mental health is a critical focus of research in global public health. In recent years, mental health problems among adolescents worldwide have shown an upward trend [1], particularly depressive symptoms and other negative emotions [2]. Depressive symptoms in adolescents refer to a mental disorder occurring during adolescence, typically characterized by persistent low mood and loss of interest or pleasure [3]. Over the past decade, the global prevalence of self-reported depressive symptoms among adolescents has risen to 34% [4]. Depressive symptoms not only induce negative emotions and anhedonia in adolescents but also impair their academic development, social adaptation, and interpersonal communication abilities [5]. The exacerbation of depressive symptoms in adolescents can lead to emotional disorders and, in severe cases, may even trigger extreme behaviors such as suicide or self-harm, imposing profound negative impacts on individuals, families, and society as a whole [6–8].
Research indicates that numerous factors may be associated with the rising incidence of depressive symptoms among adolescents, including gender, family economic status, academic pressure, sleep patterns, dietary habits, and interpersonal relationships [4, 9–11].As a critical physiological foundation for maintaining emotional stability in adolescents, sleep has been extensively linked to depressive symptoms through behavioral abnormalities [12, 13]. Sleep procrastination is defined as the act of going to bed later than intended without external constraints [14].Previous studies have demonstrated that sleep procrastination shortens total sleep duration, disrupts sleep rhythms, and impairs prefrontal cortex regulation, thereby increasing the risk of depressive symptoms [15]. Reduced sleep duration and circadian rhythm disruption associated with sleep procrastination may further exacerbate depressive symptoms in adolescents [16]. Dietary behavior is another key factor implicated in adolescent mental health. Lower levels of positive emotional eating have been associated with more severe depressive symptoms [17].For instance, a study on overweight and obese adolescents found that emotional eating and binge-eating behaviors exacerbate emotional disturbances, including increased depressive symptoms, anxiety, and anger [18]. Emotional eating patterns were most strongly associated with eating disorders, binge eating, and depressive symptoms [17].Therefore, eating behavior plays a significant role in the prevention and treatment of depressive symptoms [19], and the adoption of healthy dietary habits may reduce the risk of developing such symptoms [20].Furthermore, eating behavior, as a vital component of adolescent health behaviors, interacts dynamically with sleep and emotions. Eating disorders serve as an important mediator influencing adolescent impulsivity [21]. Research indicates that sleep procrastination may disrupt the hypothalamic-pituitary-adrenal (HPA) axis, increasing cravings for high-sugar, high-fat foods and leading to uncontrolled eating behaviors such as nighttime eating and binge eating [22]. In turn, unhealthy dietary patterns further exacerbate imbalances in neurotransmitter synthesis, reinforcing the risk of depressive moods [23].This suggests that eating behavior may play a key mediating role in the relationship between sleep procrastination and depressive symptoms.
Current research on sleep and depressive symptoms in adolescents has primarily focused on the association between sleep quality or duration and depressive symptoms [24–26], with most studies examining only the bivariate association between these two variables in isolation. Mediating relationships explored in existing literature are primarily concentrated in domains such as physical activity [10, 24]. Notably, few studies have investigated the relationship between sleep procrastination and depressive symptoms, nor have they examined the potential mediating role of eating behavior in this relationship. Therefore, this study aimed to explore the relationship between sleep procrastination and depressive symptoms among adolescents, with a particular focus on the mediating role of eating behavior. By clarifying these relationships, this study intended to propose targeted preventive measures for association related to the onset of adolescent depressive symptoms, thereby providing a theoretical basis and practical guidance for effectively reducing the occurrence and progression of depressive symptoms in this population.
This study adopted a cross-sectional survey design, conducted from March to July 2025, using stratified cluster sampling. First, stratified sampling was performed by randomly selecting two districts and one county (Shaibak District, Midong District, and Ürümqi County) from the seven districts and one county of Ürümqi. Then, cluster sampling was conducted by randomly selecting one secondary school from each district or county. All non-boarding students in Grades 7 to 9 at each selected school were included as participants, with no fewer than 80 students per grade. After obtaining informed consent from both parents/guardians and the participants themselves, an adolescent health status questionnaire was distributed and completed on-site. A total of 1,257 questionnaires were distributed; after excluding invalid ones with missing information or patterned responses, 1,121 valid questionnaires were ultimately obtained, resulting in a valid response rate of 89.2%.Data verification showed that only a small number of non-core variables had scattered missing values in the valid questionnaires. The univariate missing rate for each variable was < 2%, and the overall missing rate was < 1%, which was consistent with the characteristics of missing completely at random (MCAR). All participants voluntarily participated in the study and signed informed consent forms. This study was approved by the Ethics Committee of Xinjiang Medical University (Approval No.: XJYKDXR20240521007).
Basic demographic information on adolescents in Urumqi, Xinjiang, was collected using the Adolescent Health Status Questionnaire, including sex, age, height, weight, parental educational level, household economic status, and academic workload. Each participant’s BMI was calculated based on the height and weight provided in the questionnaire, using the formula BMI = weight (kg) / [height (m)]².
The Bedtime Procrastination Scale (BPS) was developed by Kroese et al. [27]. It consists of 9 items scored on a 5-point Likert scale, ranging from 1 (almost never) to 5 (almost always). Items 2, 3, 7, and 9 are reverse-scored (5–1), while the remaining items are scored normally (1–5). This scale has a single-factor structure, with the mean of 9 items serving as the scale score, ranging from 1 to 5 [28]. Higher scores indicate more severe sleep procrastination behavior.The Cronbach’s α coefficient for this scale in the present study was 0.837.
The Patient Health Questionnaire-9 (PHQ-9) is a depressive self-assessment scale developed by Kroenke et al. [29] based on the diagnostic criteria for depressive disorder outlined in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). This scale consists of 9 items designed to assess participants’ depressive symptoms over the past two weeks. It employs a 4-point Likert scale ranging from 0 (Never) to 3 (Almost every day), yielding a total score between 0 and 27. Based on established criteria from previous studies, depressive symptoms in this study were defined as scores exceeding 5 [24].This threshold demonstrates high sensitivity within the adolescent population, maximising the identification of potentially at-risk individuals.The Cronbach’s α coefficient for this research scale was 0.865.
The Sakata Eating Behavior Scale Short Form (EBS-SF) was developed by Tayama et al. [30] and adapted into Chinese by Ge Pu et al. [31]. This scale comprises seven body image perception, substitute eating, eating patterns, eating motivation, abnormal eating rhythms, satiety, and meal content, consisting of seven items in total. The Likert 4-point scale was used, with scores ranging from 1 (Strongly Disagree) to 4 (Strongly Agree), yielding a total score range of 7 to 28. Higher scores indicate more abnormal eating behavior among respondents. The Cronbach’s α coefficient for this research scale was 0.872.
In this study, a series of variables associated with sleep procrastination and depressive symptoms were included, namely sex, age, BMI, father’s educational level, mother’s educational level, household economic status, and academic workload. BMI ≤ 23.9 was classified as normal, 24.0 ≤ BMI < 27.9 as overweight, and BMI ≥ 28.0 as obese; parental educational level was divided into “high school and below” and “college or above”; household economic status was categorized based on adolescents’ subjective perceptions into three better, generally, and poor. Academic workload was classified according to adolescents’ self-assessment of study pressure into three Lighter, generally, and Heavier.
Harman’s single-factor test was conducted as a preliminary assessment of common method bias [32]. Exploratory factor analysis (EFA) was conducted on all measurement items, which extracted 16 factors with eigenvalues greater than 1. The first factor explained 18.71% of the total variance, a proportion that is below the widely accepted threshold of 50%. Although this result indicates that common method bias is unlikely to be a major confounding factor in the present study, Harman’s single-factor test has well-documented limitations and thus cannot definitively eliminate the possibility of bias inherent in self-reported data. Therefore, the findings of this study should be interpreted with caution. Future research is recommended to adopt multi-source data collection or longitudinal study designs to further mitigate this limitation and enhance the robustness of the research results.
The measurement model was evaluated using confirmatory factor analysis (CFA). The initial three-factor model, including sleep procrastination, eating behavior, and depressive symptoms, yielded the following fit χ²/df = 5.72, CFI = 0.913, TLI = 0.902, RMSEA = 0.064, SRMR = 0.052. All fit indices met the acceptable criteria for psychological research (CFI/TLI > 0.90, RMSEA < 0.08, SRMR < 0.08), indicating that the three-factor structure has good construct validity.
Data analysis was performed using IBM SPSS Statistics 27.0 and R software. Normally distributed continuous variables were presented as mean ± standard deviation, while non-normally distributed continuous variables were reported as median (interquartile range). Categorical variables were described as frequencies and proportions. Between-group differences in continuous variables were assessed using the Mann-Whitney U test (for two groups) and the Kruskal-Wallis H test (for three or more groups). Spearman’s correlation analysis was applied to examine bivariate correlations. Logistic regression was used to analyze the relationship between sleep procrastination and depressive symptoms with three sequential Model No covariate adjustment; Model Adjusted for sex within individual factors; Model Adjusted for other relevant factors, including mother’s educational level, household economic status, and academic workload. Restricted cubic spline (RCS) models were also used to examine the association.Mediation analyses were conducted using the Generalized Additive Model (GAM) to account for the nonlinear relationship among variables; depressive symptoms were treated as a binary variable with a logit link function; 5,000 bootstrap resamples were performed to estimate the 95% confidence intervals of the effects, where the regression coefficients (β) reported in the tables represent log odds ratios, and the significance of effects was determined by whether the confidence intervals included zero.The significance level for this study was α = 0.05.
This study included a total of 1,121 participants, comprising 506 males and 615 females. The results showed that the depression symptoms score was higher in females [6 (3, 10)] than in males [4 (1, 9)]; adolescents whose mothers had an associate degree or higher education had higher depression symptoms scores [6 (3, 9)] than those whose mothers had a lower educational level [5 (2, 10)]; adolescents in the poor household economic status group had higher scores [9 (9, 14)] than those in the average [6 (3, 9)] and better [5 (2, 9)] household economic status groups; and adolescents with a heavier academic workload had higher scores [9 (5, 13)] than those with average [5 (5, 9)] and lighter [3 (3, 6)] academic workload. Sex, maternal educational level, household economic status, and academic workload were statistically associated with depressive symptoms in adolescents (p < 0.05) See Table 1.
Table 1Basic demographic characteristicsVariable
n
Depressive symptoms[M(P25,P75)]p valuescoreH/Z/rSex-6.431< 0.001 Male5064(1,9) Female6156(3,10)Age14(14,15)6(3,10)-0.0230.451BMI0.4040.817 ≤23.99826(3,10) 24-27.9965(3,9) ≥ 28437(2,9)Paternal educational level-8920.372 High school and below8135(2,10) Associate degree or higher3086(3,9)Maternal educational level-2.2290.026 High school and below7755(2,9) Associate degree or higher3466(3,10)Household economic status15.233< 0.001 Better3525(2,9) Average7236(3,9) Poor469(9,14)Academic workload85.194< 0.001 Lighter723(3,6) Average7945(5,9) Heavier2559(5,13)
Results indicated that the odds ratio (OR) for sleep procrastination in Model 1 was 3.380 (95% CI [2.67, 4.27]); after adjusting for covariates in Models 2 and 3, the OR gradually decreased, with an OR of 2.981 in Model 3. However, all OR values remained greater than 1, and all associations were statistically significant (p < 0.05). These results suggested a positive association between sleep procrastination and depressive symptoms. See Table 2. A restricted cubic spline model with 5 knots was used in this study. The knots were placed at the 0th, 25th, 50th, 75th, and 100th percentiles of sleep procrastination scores, with 4 degrees of freedom. The median sleep procrastination score (2.44) was set as the reference point to analyze the nonlinear association between sleep procrastination and depressive symptoms. When sleep procrastination scores were below this reference point, the OR for depressive symptoms was less than 1, indicating a weaker association compared with the reference level. As sleep procrastination scores increased, the OR exhibited a nonlinear trend of first increasing and then decreasing (p for nonlinearity < 0.05). These findings indicated that sleep procrastination was significantly associated with depressive symptoms in a nonlinear manner See Fig. 1.
Table 2Relationship between sleep procrastination and depressive symptomsSleep ProcrastinationOR(95%CI)p valueModel 13.380(2.670, 4.270)< 0.001Model 23.238(2.552, 4.107)< 0.001Model 32.981(2.336, 3.805)< 0.001Model No covariate adjustment; Model Adjusted for sex within individual factors; Model Adjusted for other relevant factors, including mother’s educational level, household economic status, and academic workload
Fig. 1RCS model of the relationship between sleep procrastination and depressive symptoms
This study employed a Generalized Additive Model (GAM) with Bootstrap sampling (5000 iterations) to calculate 95% confidence intervals and examine the indirect effect of eating behavior in the relationship between sleep procrastination and depressive symptoms. The mediation analysis showed that sleep procrastination was significantly associated with eating behavior (a = 0.177, p < 0.001), and eating behavior was significantly associated with depressive symptoms (b = 0.550, p < 0.001). The indirect effect between sleep procrastination and depressive symptoms via eating behavior was 0.097 (95% CI [0.058, 0.143]), a finding consistent with a partial mediation model, accounting for 12.4% of the total effect. See Table 3; Fig. 2.
Table 3Mediating effects of eating behavior on the relationship between sleep procrastination and depressive symptomsEffectPathβSE95% CIProportionDirect effect (c′)Sleep procrastination →Depressive symptoms0.6850.081[0.532, 0.847]87.6%Indirect effect (a×b)Sleep procrastination→ Eating behavior→ Depressive symptoms0.0970.023[0.058, 0.143]12.4%Total effect (c)0.7820.150[0.618, 0.956]100%
Fig. 2Mediating model of sleep procrastination, eating behavior, and depression symptoms. ***p < 0.001. Note. a, b, and c’ are all direct effects
This study employed moderated mediation analysis to examine the indirect effect of eating behavior in the relationship between sleep procrastination and depressive symptoms. Results indicated that findings were consistent with a significant partial mediation model. The indirect effect was significant in both males and in males, the indirect effect was 0.072 (95% CI [0.018, 0.125]), accounting for 9.9% of the total effect; in females, the indirect effect was 0.106 (95% CI [0.049, 0.163]), accounting for 13.6% of the total effect. However, sex did not significantly moderate either pathway in this mediation model (sleep procrastination × sex interaction term β = 0.076, p = 0.201; eating behavior × sex interaction term β = −0.038, p = 0.789). This suggested that the association between sleep procrastination and eating behavior, and the association between eating behavior and depressive symptoms, were of comparable magnitude in males and females. Although the indirect effect was numerically slightly higher in females than in males (Δβ = 0.034), this difference was not statistically significant. See Tables 4 and 5.
Table 4Analysis results of the moderated mediation modelPathPredictorβ/EstimateSEt/z valuep valueSleep procrastination → Eating behaviorSleep procrastination × Sex (Interaction)0.0760.0591.2800.201Eating behavior → Depressive symptomsEating behavior × Sex(Interaction)-0.0380.141-0.2680.789
Table 5Analysis of conditional indirect effectsPathMaleFemaleβ95% CIβ95% CISleep procrastination→ Eating behavior(a)0.129[0.028, 0.232]0.205[0.117, 0.293]Eating behavior→ Depressive symptoms(b)0.561[0.346, 0.791]0.521[0.326, 0.730]Sleep procrastination →Depressive symptoms(c’)0.663[0.441, 0.917]0.677[0.470, 0.899]Indirect effect (a×b)0.072[0.018, 0.125]0.106[0.049, 0.163]Total effect (c)0.735[0.495, 1.012]0.784[0.572, 1.014]Proportion9.9%13.6%
The prevalence of depressive symptoms among adolescents is rising globally, and evidence suggests that lifestyle factors such as sleep and dietary patterns are closely associated with depressive symptoms [, 35]. This study constructed a mediation model to explore the relationship between sleep procrastination and depressive symptoms in adolescents. We aimed to describe the characteristics of sleep procrastination, examine its association with depressive symptoms, explore whether eating behavior is consistent with a potential indirect effect, and further test the potential moderating role of sex.The results indicated a nonlinear association between sleep procrastination and depressive symptoms, showing a trend of first increasing and then decreasing. In addition, the indirect effect of sleep procrastination on depressive symptoms via eating behavior was statistically significant, consistent with a partial mediation model, and this indirect association was not significantly moderated by sex. Notably, the observed nonlinear association between sleep procrastination and depressive symptoms differed from the linear association commonly reported in previous studies [36]. This discrepancy may be attributed to differences in measurement instruments, sample characteristics, or analytical approaches across studies.
The present study found a significant association between sleep procrastination and depressive symptoms, which is consistent with previous findings [37–39].Sleep procrastination is typically characterized as bedtime procrastination, referring to the behavioral tendency of an individual to fail to go to bed on time without external obstacles. Previous research has indicated that individuals who frequently delay their bedtime show a positive association with depressive symptoms [14], which aligns with the results of the present study.Based on subjective assessments of sleep and depressive symptoms, this study revealed an associational pattern between the two variables. Sleep procrastination may also be associated with depressive symptoms through other pathways. For example, Zhai et al. reported that the most significant effect of bedtime procrastination on sleep is reduced sleep duration [28], and sleep duration is positively correlated with depressive mood [40]. Furthermore, other studies have shown that sleep procrastination is also associated with delayed bedtime, irregular sleep patterns, and disrupted circadian rhythm stability. These findings suggest that circadian rhythm disturbances may potentially influence the secretion of key neurotransmitters such as melatonin and serotonin, and impaired serotonin secretion is closely associated with core depressive symptoms, including low mood and anhedonia [41].
Furthermore, the results of this study suggest that the indirect effect between sleep procrastination and depressive symptoms via eating behaviour is statistically significant. This finding is consistent with certain mediation models, suggesting that eating behaviour may play a mediating role in the statistical association between sleep procrastination and depressive symptoms. This study found a significant positive correlation between sleep delay scores and dietary behaviour scale scores, indicating that individuals with higher levels of sleep delay also have higher scores on the dietary behaviour scale, manifesting as abnormalities in eating patterns, eating motivations and eating rhythms. Research has suggested that sleep delay may disrupt the secretion of appetite-regulating hormones, thereby being associated with abnormal dietary behaviour [42]. In this study, sleep delay scores were significantly positively correlated with eating behaviour scale scores, consistent with this theoretical perspective. Concurrently, eating behaviour scale scores were significantly positively correlated with depression symptom scale scores. Previous research has found an association between abnormal eating behaviour and depressive symptoms [43], which is consistent with the findings of this study. Furthermore, this indirect pathway did not exhibit significant gender-moderating effects. This result is consistent with the systematic review by Khalid et al. [44], which, after analysing multiple studies, noted that associations between sleep, eating and depressive symptoms are prevalent among children and adolescents, and that these associations do not exhibit consistent gender differences. Other literature suggests that sleep delay behaviour may be related to reduced self-regulation, and that self-regulation levels may simultaneously influence both eating behaviour and emotional adaptation processes [45, 46]. Based on this, it is speculated that sleep delay, abnormal eating behaviours and depressive symptoms may not be mutually independent issues, but rather are associated with common self-regulation-related factors [47]. However, as this study did not directly measure self-regulation ability, future longitudinal or experimental studies could be conducted to further test this hypothesis.
This study examined the relationship between sleep procrastination and depressive symptoms in adolescents, with eating behavior included as a potential statistical mediator. It provides preliminary evidence relevant to research on adolescent depressive symptoms, although several limitations should be noted. First, this study employed a cross-sectional design, which precludes causal or temporal inferences among variables. Second, this study did not distinguish between specific subtypes of eating behavior (e.g., emotional eating or restrictive eating), and thus could not identify which subtype may be more strongly associated with the observed statistical indirect effect. Future studies may adopt multiple mediation models to compare potential indirect effects across different eating behavior subtypes. Third, all data were collected using self-report questionnaires, which may involve subjective bias. Notably, Harman’s single-factor test has limitations and cannot fully rule out common method bias; thus, caution is still needed in interpreting the results.
This study constructed a mediation model to explore the relationship between sleep procrastination and depressive symptoms in adolescents. The study found a significant positive correlation between sleep procrastination and depressive symptoms. Additionally, eating behavior showed a significant indirect effect in the relationship between sleep procrastination and depressive symptoms, consistent with a partial mediation model. However, sex did not significantly moderate this indirect association. The findings of this study propose hypotheses for future longitudinal and experimental studies to examine the temporal order of variables and explore whether interventions targeting sleep procrastination or eating behavior may be associated with changes in depressive symptoms among adolescents.