Internet Addiction and Suicidal Ideation in Chinese Adolescents: Sex Differences and the Interactive Roles of Trauma and Depression

Abstract

This study investigated the association between Internet addiction (IA) and suicidal ideation (SI) among Chinese adolescents, focusing on sex differences and the interactive roles of childhood trauma and depressive symptoms. A large-scale multicenter survey involving 24,470 students (mean age = 14.37 ± 1.52 years; 48.40 percent female) from three Chinese cities was conducted using standardized instruments to assess IA, SI, childhood trauma, and depression. Logistic regression and interaction analyses showed that IA was significantly associated with a higher likelihood of SI, with stronger associations observed among female adolescents. Childhood trauma and depressive symptoms significantly moderated the IA–SI relationship (interaction odds ratio [OR] = 6.81/6.88). However, stratified analysis showed that the association between IA and SI was stronger among adolescents without childhood trauma (OR = 3.38) than among those with childhood trauma (OR = 2.63), and the association was also stronger among adolescents without depressive symptoms (OR = 2.88) compared to those with depressive symptoms (OR = 1.80). These differences were primarily attributed to the higher baseline risk of SI among adolescents with childhood trauma and depressive symptoms, which attenuated the relative effect of IA. This effect was more pronounced during early adolescence. These findings suggest that IA may represent a potentially modifiable behavioral risk factor for SI in adolescents, particularly among females. The results highlight the importance of integrating digital behavior and mental health factors in adolescent suicide risk prevention, with a focus on comprehensive intervention strategies for adolescents with psychological vulnerabilities.

Keywords

adolescents suicidal ideation internet addiction childhood trauma depression sex differences

Introduction

With the rapid development of digital technologies, the Internet has become an indispensable part of daily life. Among adolescents in China, the Internet usage has reached 94.9 percent.¹ Internet addiction (IA) is generally defined as excessive and poorly controlled engagement in Internet-based activities that leads to functional impairment or psychological distress.² It is estimated to affect 8.1–26.5 percent of Chinese adolescents (aged 10–24 years).^3–5 Internet gaming disorder, a specific form of problematic online behavior, was formally recognized in the 11th edition of the International Classification of Diseases in 2018.⁶ This recognition highlights growing global concern regarding problematic online behaviors and their public health implications.

IA has been linked to various psychological problems and adverse outcomes,⁷ among which suicide has attracted particular attention due to its severe consequences. Suicide is now the second leading cause of death among adolescents worldwide.⁸ Suicidal ideation (SI) is the strongest predictor of suicidal behavior.⁹ Studies indicate that 15–24 percent of Chinese adolescents (most under 18 years of age; mean age ≈ 15 years), have experienced SI, underscoring its substantial prevalence during adolescence.^10,11

However, the association between IA and SI remains inconsistent. Some studies have reported a positive association,^12,13 whereas others have found no significant relationship.¹⁴ One potential explanation is that previous studies often relied on binary classifications of IA without considering its severity spectrum. Emerging evidence suggests that only severe IA increases the risk of SI.¹⁵ These inconsistent findings highlight the need for large-scale studies examine the association between different levels of IA severity and SI.

The biological–psychosocial model of suicide provides a comprehensive framework, emphasizing the interplay of multiple risk factors.¹⁶ Within this model, childhood trauma and depressive symptoms are recognized as key distal and proximal factors, respectively.^17,18 In addition, childhood trauma and depressive symptoms may also influence adolescents’ Internet use. Some studies have indicated that individuals who have experienced childhood trauma or have depressive symptoms tend to exhibit more severe IA.^19,20 However, it remains unclear whether IA interacts with these factors in influencing SI.

Sex differences have been observed, with IA being more prevalent among males and SI being more common among females.^21,22 Prior research suggests that the pathways linking IA to SI may differ by sex. Female adolescents with IA may be more susceptible to internalizing pathways, such as depression, social comparison, and emotional dysregulation, all of which are strongly linked to SI.²³ In contrast, males with IA tend to exhibit higher impulsivity, externalizing behaviors, and gaming-related risk tendencies. These factors may contribute to alternative pathways leading to SI.²⁴ These differences suggest that the IA–SI association may vary by sex; however, this issue remains insufficiently explored.

Guided by the biopsychosocial model of suicide,¹⁶ this study examined SI as the outcome of interacting behavioral, psychological, and contextual factors. IA was conceptualized as a potentially modifiable behavioral risk factor, whereas childhood trauma and depressive symptoms represent psychological vulnerability. These two variables were examined as moderators to assess whether the IA–SI association varies across different levels of vulnerability. Sex and age were included given established developmental and sex differences in IA and SI, and family and demographic characteristics were treated as covariates to account for potential confounding. This framework allows for the assessment of both the independent association between IA and SI and its variation across key subgroups.

Materials and Methods

Participants and data collection

Data were collected between June and August 2024 using a multicenter cluster sampling design. Secondary school students were recruited from 28 schools in three regions of China: Wuyuan, Yantai, and Rizhao. Paper-based questionnaires were administered during class meetings and completed under researcher supervision. Participation was voluntary, and written informed consent was obtained from both students and their parents/guardians.

The questionnaire took ∼15 minutes to complete. Of the 25,382 questionnaires distributed, 912 were excluded because of substantial missing data, inconsistent responses, or nonresponse. A total of 24,470 valid questionnaires were retained, yielding a valid response rate of 96.4 percent. The final sample had a mean age of 14.37 years (standard deviation [SD] = 1.52), and 48.40 percent were female.

A predefined suicide risk management protocol was implemented. Students screening positive for SI received further assessment, and those at elevated risk were referred to school psychological services or appropriate professional institutions for follow-up support. All participants were provided with information on available mental health resources. The study was approved by the affiliated institution’s Research Ethics Committee (2024-R-157).

Measures

Demographic information

A self-designed questionnaire was used to collect socioeconomic, family, and demographic information, including sex, age, grade, academic performance, personality, residence, family economic status, parental marital status, parental education level, and only-child status.

Internet addiction

IA severity was assessed using Young’s Internet Addiction Test, a 20-item scale rated on a 5-point Likert scale ranging from 1 (rarely) to 5 (always).²⁵ The Chinese version has been validated among Chinese adolescents (Cronbach’s α = 0.93).²⁶ Total scores range from 0 to 100: below 40 means no addiction, 40–59 mild, 60–79 moderate, and 80 or above severe. A cutoff of 50 was used to identify IA (Cronbach’s α = 0.913).

Childhood trauma

Childhood trauma was assessed using the Chinese version of the Childhood Trauma Questionnaire-Short Form, which has been validated among Chinese adolescents.²⁷ The questionnaire contains 25 items rated on a 5-point Likert scale ranging from 1 (never) to 5 (always). The questionnaire includes 25 items rated from 1 (never) to 5 (always). Total scores range from 25 to 125, with higher scores indicating more severe trauma. A score above 45 indicates high trauma (Cronbach’s α = 0.847).

Depressive symptoms

Depressive symptoms were assessed using the Children’s Depression Inventory (CDI), which was designed for children and adolescents.²⁸ The Chinese version has been validated among Chinese children and adolescents (Cronbach’s α = 0.88).²⁹ The CDI includes 27 items rated on a 0–2 scale, with higher scores indicating more severe depressive symptoms. A cutoff score of 19 was used to identify depressive symptoms. To avoid overlap with the SI outcome, Item 9, which assesses SI, was excluded from the CDI total score in this study (Cronbach’s α = 0.889).

Suicidal ideation

SI was assessed using Item 9 of the CDI: 0 = I do not have thoughts of suicide; 1 = I have thought about suicide, but I would never act on it; and 2 = I might commit suicide. Adolescents who scored “1” or “2” were classified as having SI. Item 9 is widely used in adolescent research and has been validated as a measure of SI.^30,31

Statistical analysis

Descriptive analysis

Descriptive analyses were performed using SPSS version 27.0. Categorical variables were summarized as frequencies and percentages, whereas continuous variables were summarized as means and standard deviations. Differences between participants with and without SI, and between males and females, were examined using χ² tests or two-tailed independent samples t-tests after assessing normality.

Binary logistic regression analysis

Binary logistic regression was performed to examine the association between IA and SI, with results reported as odds ratios (ORs) and 95 percent confidence intervals (CIs). Analyses were conducted in the total sample and further stratified by age group (10–14 vs. 15–19 years) to assess developmental differences in the IA–SI association. A hierarchical modeling strategy was applied with three progressively adjusted models.

Covariates were selected based on prior evidence and the biopsychosocial framework of suicide, with the aim of accounting for demographic, family/contextual, and psychological factors that may confound the association between IA and SI. Model 1 adjusted for demographic characteristics (sex, age, grade, academic performance, and personality). Model 2 further adjusted for family socioeconomic and environmental variables (family economic status, residence, parental marital status, only-child status, and parental education level). Model 3 additionally included childhood trauma and depressive symptoms as key psychological factors to evaluate whether the IA–SI association remained after accounting for these variables.

Subgroup analysis

Subgroup analyses were conducted to explore whether the IA–SI association differed by sex. Sex differences were assessed using ORs. To provide developmental context, additional analyses were performed within age groups (10–14 vs. 15–19 years) to examine sex differences separately across early and late adolescence.

Interaction analysis

Using the cutoff value, the study examined the interactive effects of IA with childhood trauma and depressive symptoms on SI separately and plotted the corresponding trend maps. Interaction analysis was also analyzed by age group.

Results

A total of 24,470 adolescents participated in the study (mean age = 14.37 ± 1.52 years), including 12,626 males (51.60 percent) and 11,844 females (48.40 percent). Among the sample, the prevalence of IA was 19.4 percent. In addition, 42.1 percent reported experiencing SI, 19.1 percent had a history of childhood trauma, and 27.5 percent exhibited depressive symptoms. The sample was divided into four IA severity groups: no addiction (n = 14,281, 58.4 percent), mild addiction (n = 8,660, 35.4 percent), moderate addiction (n = 1,367, 5.6 percent), and severe addiction (n = 162, 0.7 percent) (Table 1).

Table 1.

Characteristics of Participants (n, %)

Subgroup	Total	Suicidal ideation		Sex
Subgroup	Total	No (n = 14,159)	Yes (n = 10,311)	Male (n = 12,626)	Female (n = 11,844)
Age
10–14	13,703 (56.0)	7,951 (56.2)	5,752 (55.8)	6,986 (55.3)	6,717 (56.7)
15–19	10,767 (44.0)	6,208 (43.8)	4,559 (44.2)	5,640 (44.7)	5,127 (43.3)
Grade
Middle	18,657 (76.2)	10,929 (77.2)	7,728 (74.9)	9,676 (76.6)	8,981 (75.8)
High	5,813 (23.8)	3,230 (22.8)	2,583 (25.1)	2,950 (23.4)	2,863 (24.2)
Performance
Upper tier	4,708 (19.2)	3,019 (21.3)	1,689 (16.4)	2,516 (19.9)	2,192 (18.5)
Middle tier	13,177 (53.8)	7,773 (54.9)	5,404 (52.4)	6,417 (50.8)	6,760 (57.1)
Lower tier	6,585 (26.9)	3,367 (23.8)	3,218 (31.2)	3,693 (29.2)	2,892 (24.4)
Personality
Introverted	3,710 (15.2)	1,884 (13.3)	1,826 (17.7)	1,809 (14.3)	1,901 (16.1)
Extroverted	6,044 (24.7)	3,837 (27.1)	2,207 (21.4)	3,488 (27.6)	2,556 (21.6)
Neutral	14,716 (60.1)	8,438 (59.6)	6,278 (60.9)	7,329 (58.0)	7,387 (62.4)
Residence
Urban	5,250 (21.5)	3,242 (22.9)	2,008 (19.5)	2,722 (21.6)	2,528 (21.3)
Town	10,661 (43.6)	6,225 (44.0)	4,436 (43.0)	5,569 (44.1)	5,092 (43.0)
Rural	8,559 (35.0)	4,692 (33.1)	3,867 (37.5)	4,335 (34.3)	4,224 (35.7)
Parents’ marital status
First marriage	20,809 (85.0)	1,2341 (87.2)	8,468 (82.1)	10,793 (85.5)	10,016 (84.6)
Divorced	1,935 (7.9)	939 (6.6)	996 (9.7)	964 (7.6)	971 (8.2)
One deceased	343 (1.4)	176 (1.2)	167 (1.6)	169 (1.3)	174 (1.5)
Remarried	1,383 (5.7)	703 (5.0)	680 (6.6)	700 (5.5)	683 (5.8)
Only child
Yes	3,886 (15.9)	2,367 (16.7)	1,519 (14.7)	2,359 (18.7)	1,527 (12.9)
No	20,584 (84.1)	11,792 (83.3)	8,792 (85.3)	10,267 (81.3)	10,317 (87.1)
Father’s education
Middle school or below	14,759 (60.3)	8,184 (57.8)	6,575 (63.8)	7,668 (60.7)	7,091 (59.9)
High school/Vocational	6,349 (25.9)	3,798 (26.8)	2,551 (24.7)	3,198 (25.3)	3,151 (26.6)
Associate degree or above	3,362 (13.7)	2,177 (15.4)	1,185 (11.5)	1,760 (13.9)	1,602 (13.5)
Mother’s education
Middle school or below	16,336 (66.8)	9,092 (64.2)	7,244 (70.3)	8,413 (66.6)	7,923 (66.9)
High school/Vocational	5,204 (21.3)	3,168 (22.4)	2036 (19.7)	2,691 (21.3)	2,513 (21.2)
Associate degree or above	2,930 (12.0)	1,899 (13.4)	1,031 (10.0)	1,522 (12.1)	1,408 (11.9)
Family economic status
Superior	748 (3.1)	501 (3.5)	247 (2.4)	408 (3.2)	340 (2.9)
Good	10,056 (41.1)	6,289 (44.4)	3,767 (36.5)	5,126 (40.6)	4,390 (41.6)
Average	12,206 (49.9)	6,723 (47.5)	5,483 (53.2)	6,279 (49.7)	5,927 (50.0)
Poor	1,460 (6.0)	646 (4.6)	814 (7.9)	813 (6.4)	647 (5.5)
Internet addiction
None addiction	14,281 (58.4)	9,955 (70.3)	4,326 (42.0)	7,113 (56.3)	7,168 (60.5)
Mild addiction	8,660 (35.4)	3,845 (27.2)	4,815 (46.7)	4,677 (37.0)	3,983 (33.6)
Moderate addiction	1,367 (5.6)	330 (2.3)	1,037 (10.1)	748 (5.9)	619 (5.2)
Severe addiction	162 (0.7)	29 (0.2)	133 (1.3)	88 (0.7)	74 (0.6)
Childhood trauma
No	19,787 (80.9)	12,851 (90.8)	6,936 (67.3)	10,217 (80.9)	9,570 (80.8)
Yes	4,683 (19.1)	1,308 (9.2)	3,375 (32.7)	2,409 (19.1)	2,274 (19.2)
Depressive symptoms
No	17,737 (72.5)	12,575 (88.8)	5,162 (50.1)	9,459 (74.9)	8,278 (69.9)
Yes	6,733 (27.5)	1,584 (11.2)	5,149 (49.9)	3,167 (25.1)	3,566 (30.1)

Logistic regression results showed that IA severity was significantly associated with SI (Table 2). The adjusted odds ratios (aORs) changed only slightly from Model 1 to Model 2 (e.g., mild: 3.06–2.99; moderate: 7.68–7.44; severe: 10.62–10.26). However, after further adjustment for childhood trauma and depressive symptoms in Model 3, the associations were attenuated, with aORs decreasing to 2.14, 3.28, and 3.29 for mild, moderate, and severe IA, respectively. This association remained consistent across age groups.

Table 2.

Relationships Between Internet Addiction and Suicidal Ideation

Group	n (%)	OR (95 % CI)
Group	n (%)	Unadjusted	Model 1^a	Model 2^b	Model 3^c
Total
None addiction	14,281 (58.36)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)
Mild addiction	8,660 (35.39)	2.88 (2.73–3.05)***	3.06 (2.89–3.24)***	2.99 (2.82–3.18)***	2.14 (2.01–2.28)***
Moderate addiction	1,367 (5.58)	7.23 (6.36–8.23)***	7.68 (6.73–8.76)***	7.44 (6.51–8.49)***	3.28 (2.83–3.79)***
Severe addiction	162 (0.67)	10.55 (7.05–15.80)***	10.62 (7.06–15.96)***	10.26 (6.81–15.45)***	3.29 (2.12–5.09)***
10–14
None addiction	8,819 (64.36)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)
Mild addiction	4,117 (30.04)	3.17 (2.94–3.42)***	3.27 (3.02–3.54)***	3.19 (2.94–3.56)***	2.18 (1.99–2.38)***
Moderate addiction	669 (4.88)	8.13 (6.73–9.83)***	8.34 (6.87–10.12)***	8.06 (6.63–9.80)***	3.21 (2.59–3.99)***
Severe addiction	98 (0.72)	9.91 (5.93–16.56)***	10.15 (6.04–17.07)***	9.71 (5.76–16.36)***	2.88 (1.64–5.08)***
15–19
None addiction	5,462 (50.73)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)
Mild addiction	4,543 (42.19)	2.71 (2.49–2.94)***	2.66 (2.44–2.89)***	2.61 (2.40–2.84)***	1.79 (1.62–1.97)***
Moderate addiction	698 (6.48)	6.67 (5.58–7.96)***	6.40 (5.35–7.65)***	6.20 (5.18–7.42)***	2.31 (1.83–2.92)***
Severe addiction	64 (0.60)	11.67 (6.08–22.40)***	10.48 (5.45–20.15)***	10.44 (5.41–20.11)***	2.91 (1.45–5.82)***

Cultural background of parents, only child, childhood trauma, and depressive symptoms.

Model 1: adjusted for sex, age (in total model), grade, performance, and personality.

Model 2: adjusted for sex, age, grade, performance, personality, residence, family economic status, and parental marriage.

Model 3: adjusted for sex, age, grade, performance, personality, residence, family economic status, and parental marriage.

***p < 0.001.

CI, confidence interval; OR, odds ratio.

Subgroup analyses indicated that IA was associated with SI in both males and females after adjustment for covariates (Table 3). In the overall sample, females exhibited higher aORs than males across all IA severity levels (mild: 2.27 vs. 1.99; moderate: 4.10 vs. 2.88; severe: 6.42 vs. 2.61), with sex differences observed in the mild and moderate addiction groups. Age-stratified analyses similarly showed generally higher aORs among females than males (10–14 years: mild 2.53 vs. 1.93, moderate 3.79 vs. 2.94, severe 3.79 vs. 2.46; 15–19 years: mild 2.07 vs. 2.07, moderate 4.37 vs. 2.89, severe 17.00 vs. 2.85). However, sex differences in early adolescence were primarily evident in the mild addiction group (2.53 vs. 1.93, p < 0.001), whereas no sex differences were observed in late adolescence (p > 0.05).

Table 3.

Relationships Between Internet Addiction and Suicidal Ideation by Sex

Group	Male			Female			ROR^c	p
Group	n (%)	OR (95 % CI)^a	aOR (95 % CI)^b	n (%)	OR (95 % CI)^a	aOR (95 % CI)^b	ROR^c	p
Total
None addiction	7,113 (56.34)	1.00 (Reference)	1.00 (Reference)	7168 (60.52)	1.00 (Reference)	1.00 (Reference)
Mild addiction	4,677 (37.04)	2.83 (2.61–3.06)***	1.99 (1.82–2.18)***	3,983 (33.63)	3.30 (3.04–3.58)***	2.27 (2.07–2.49)***	0.88	0.04
Moderate addiction	748 (5.92)	6.58 (5.60–7.74)***	2.88 (2.40–3.46)***	619 (5.23)	10.62 (8.43–13.39)***	4.10 (3.18–5.29)***	0.70	0.03
Severe addiction	88 (0.70)	8.45 (5.27–13.54)***	2.61 (1.57–4.35)***	74 (0.62)	23.97 (9.66–59.49)***	6.42 (2.48–16.63)***	0.41	0.10
10–14
None addiction	4,403 (63.03)	1.00 (Reference)	1.00 (Reference)	4,416 (65.74)	1.00 (Reference)	1.00 (Reference)
Mild addiction	2,182 (31.23)	2.90 (2.60–3.23)***	1.93 (1.70–2.18)***	1935 (28.81)	3.96 (3.53–4.44)***	2.53 (2.22–2.89)***	0.76	0.01
Moderate addiction	349 (5.00)	7.25 (5.71–9.21)***	2.94 (2.24–3.86)***	320 (4.76)	11.74 (8.38–16.44)***	3.79 (2.61–5.50)***	0.78	0.28
Severe addiction	52 (0.74)	8.35 (4.51–15.47)***	2.46 (1.24–4.85)**	46 (0.69)	17.60 (6.30–49.18)***	3.79 (1.29–11.14)*	0.65	0.51
15–19
None addiction	2,710 (48.05)	1.00 (Reference)	1.00 (Reference)	2,752 (53.68)	1.00 (Reference)	1.00 (Reference)
Mild addiction	2,495 (44.24)	2.83 (2.51–3.19)***	2.07 (1.82–2.36)***	2,048 (39.95)	2.88 (2.56–3.24)***	2.07 (1.82–2.36)***	1.00	1.00
Moderate addiction	399 (7.07)	6.21 (4.97–7.78)***	2.89 (2.25–3.72)***	299 (5.83)	9.87 (7.16–13.60)***	4.37 (3.09–6.19)***	0.66	0.06
Severe addiction	36 (0.64)	8.64 (4.14–18.01)***	2.85 (1.31–6.18)**	28 (0.54)	49.68 (6.74–366.17)***	17.00 (2.21–131.06)**	0.17	0.11

Unadjusted model.

Adjusted for sex (in total model), age, grade, performance, personality, residence, family economic status, parental marriage, cultural background of parents, only child, childhood trauma, and depressive symptoms.

Calculated by adjusted OR.

*p < 0.05; **p < 0.01; ***p < 0.001.

aOR, adjusted odds ratio; CI, confidence interval; OR, odds ratio; ROR, relative odds ratio.

Moderation analyses showed that childhood trauma and depressive symptoms influenced the association between IA and SI (Table 4). The IA–SI association was larger among adolescents without childhood trauma (OR = 3.38 vs. 2.63) and without depressive symptoms (OR = 2.88 vs. 1.80). As shown in Figure 1, adolescents with childhood trauma or depressive symptoms had a higher prevalence of SI across all IA severity levels, indicating a higher baseline risk in these psychologically vulnerable groups. However, the increase in SI prevalence across IA severity levels appeared steeper among adolescents without childhood trauma or depressive symptoms, suggesting a stronger relative IA–SI association in these lower baseline-risk groups. These findings support a subadditive interaction pattern, in which the relative contribution of IA appears attenuated among adolescents with elevated baseline SI risk. Interaction estimates indicated that the joint presence of IA and these psychological factors was associated with a higher risk than IA alone (interaction OR = 6.81/6.88), with stronger interaction patterns observed in early adolescence.

FIG. 1.

Association between severity levels of Internet addiction and suicidal ideation in adolescents.

Table 4.

Interaction and Stratified Analysis Between Internet Addiction and Suicidal Ideation

Group	Suicidal ideation		Nonsuicidal ideation		OR (95 % CI)^a	OR (95 % CI)^a for interaction
Group	IA ≥50	IA <50	IA ≥50	IA <50	OR (95 % CI)^a	OR (95 % CI)^a for interaction
Total
Childhood trauma						6.81 (5.97–7.78)***
Yes	1,380 (40.89)	1,995 (59.11)	287 (21.94)	1,021 (78.06)	2.63 (2.25–3.07)***
No	1,620 (23.36)	5,316 (76.64)	1,063 (8.27)	11,788 (91.73)	3.38 (3.10–3.68)***
Depressive symptoms						6.88 (6.18–7.65)***
Yes	2,103 (40.84)	3,046 (59.16)	467 (29.48)	1,117 (70.52)	1.80 (1.59–2.05)***
No	897 (17.38)	4,265 (82.62)	883 (7.02)	11,692 (92.98)	2.88 (2.60–3.19)***
10–14
Childhood trauma						8.03 (6.63–9.73)***
Yes	756 (37.31)	1,270 (62.69)	132 (17.25)	633 (82.67)	2.93 (2.36–3.65)***
No	748 (20.08)	2,978 (79.92)	467 (6.50)	6,719 (93.50)	3.55 (3.12–4.03)***
Depressive symptoms						8.16 (6.97–9.55)***
Yes	1,119 (37.66)	1,852 (62.34)	204 (24.94)	614 (75.06)	1.93 (1.61–2.31)***
No	385 (13.84)	2,396 (86.16)	395 (5.54)	6,738 (94.46)	2.87 (2.46–3.35)***
15–19
Childhood trauma						5.70 (4.74–6.85)***
Yes	624 (46.26)	725 (53.74)	155 (28.55)	388 (71.45)	2.31 (1.85–2.88)***
No	872 (27.17)	2,338 (72.83)	596 (10.52)	5,069 (89.48)	3.21 (2.85–3.61)***
Depressive symptoms						5.83 (5.04–6.75)***
Yes	984 (45.18)	1,194 (54.82)	263 (34.33)	503 (65.67)	1.69 (1.42–2.02)***
No	512 (21.50)	1,869 (78.50)	488 (8.97)	4,954 (91.03)	2.85 (2.48–3.27)***

Adjusted for sex, age, grade, performance, personality, residence, family economic status, parental marriage, cultural background of parents, only child, childhood trauma, and depressive symptoms.

***p < 0.001.

CI, confidence interval; IA, internet addiction; OR, odd ratio.

It should be noted that the severe addiction group had a small sample size (n = 162), resulting in lower statistical power (75.5 percent). This reduced power may affect the stability of the estimates for this group, and caution is recommended when interpreting the results for severe addiction.

Discussion

This large-scale multicenter study examined the relationship between IA and SI in Chinese adolescents, focusing on sex differences and the moderating roles of childhood trauma and depressive symptoms. IA was significantly associated with SI, particularly among females. Childhood trauma and depressive symptoms were found to moderate the IA–SI relationship, with a stronger association observed during early adolescence. These findings enhance our understanding of how these moderating effects vary across developmental stages and sexes in the context of adolescent suicide risk in China.

IA was independently associated with SI and showed a dose–response relationship, consistent with a recent study.³² Adolescents with IA typically exhibit poorer self-control and heightened impulsivity,³³ traits that significantly increase the risk of IA.³⁴ Online communication may promote reduced self-awareness and behavioral inhibition through anonymous interaction.³⁵ Consequently, repeated exposure to suicide-related content may desensitize adolescents to its negative consequences.³⁶ In addition, IA may also increase the SI risk by reducing social interactions, leading to loneliness, which is a significant risk factor for SI.³⁷

This study found a stronger association between IA and SI among females. Previous research suggests that females and males may differ in their patterns of Internet use: Females tend to engage more in social networking and online interaction, whereas males primarily use the Internet for entertainment, particularly online gaming.³⁸ These differences suggest that female adolescents may be more likely to use online activities to cope with emotional distress or interpersonal difficulties, thereby increasing their vulnerability to IA.³⁹ Furthermore, some studies suggest that females are more drawn to social or role-playing games, which involve higher emotional engagement and interpersonal interaction.³² The increasing popularity of sex-neutral or female-oriented online games, such as Honor of Kings and Onmyoji, may therefore expose females to online gaming environments with strong social and emotional components. These environments may involve psychosocial processes such as social interaction, emotional investment, and comparison,⁴⁰ which could explain the stronger IA–SI association observed among females, although this remains hypothetical. As the present study did not collect data on specific gaming behaviors, preferences, or Internet use motives, further studies incorporating behavioral surveys are needed to clarify these mechanisms.

Childhood trauma moderated the IA–SI relationship, with a pattern consistent with a subadditive interaction. As shown in Figure 1, adolescents with childhood trauma showed a consistently higher prevalence of SI across all levels of IA severity, indicating an elevated baseline risk for SI in trauma-exposed adolescents. This finding aligns with previous research showing that childhood trauma is a significant risk factor for SI.⁴¹ However, the increase in SI prevalence across IA severity levels was less steep in adolescents with childhood trauma than in those without, indicating that while trauma-exposed adolescents have a higher absolute risk of SI, the incremental effect of IA on SI is smaller in this group. This subadditive interaction may reflect the already elevated baseline risk of SI in trauma-exposed adolescents rather than suggesting that IA is unimportant. Mechanistically, adolescents with trauma exposure may be more likely to rely on online interactions for emotional support, potentially reflecting reduced trust in offline relationships.⁴² In this context, excessive Internet use may serve as an avoidant coping strategy for trauma-related distress. Although such use may temporarily reduce negative affect, persistent avoidance may interfere with the development of adaptive coping strategies, reinforce social withdrawal and emotional dysregulation, and ultimately increase vulnerability to SI.⁴³

Similarly, depressive symptoms moderated the IA–SI relationship, showing a comparable subadditive pattern. Adolescents with depressive symptoms had a significantly higher prevalence of SI across all levels of IA severity, consistent with extensive evidence linking depression to SI.⁴⁴ As with childhood trauma, although adolescents with depressive symptoms had a higher baseline prevalence of SI, the increase in SI across IA severity levels was less pronounced in this group compared with those without depressive symptoms. This suggests that the elevated baseline risk associated with depression attenuates the apparent incremental effect of IA on SI. However, this does not imply that IA is negligible in adolescents with depressive symptoms. IA may still contribute to SI through maladaptive coping pathways, as adolescents with depressive symptoms may turn to the Internet to alleviate feelings of sadness, loneliness, or hopelessness.⁴⁵ While online engagement may provide temporary relief, excessive reliance on digital reinforcement may worsen emotional dysregulation over time, reduce offline support, and increase hopelessness, further lowering the threshold for SI.⁴⁶

The interaction results suggest that childhood trauma and depressive symptoms have a greater impact during early adolescence. This period is crucial for rapid brain development, especially in areas related to emotional regulation and social functions.⁴⁷ During this stage, adolescents’ cognitive abilities are still developing, and neural connections are highly sensitive to external influences.⁴⁸ Adolescents are also in the process of developing psychological adaptation and coping strategies.⁴⁹ As a result, they may lack the resources needed to effectively cope with trauma and negative emotions. Therefore, providing adequate emotional support and mental health resources is crucial to help adolescents build resilience and improve coping mechanisms.

These findings should also be understood within the sociocultural context of China. Chinese adolescents face significant academic pressure and high parental expectations, while family communication is often more reserved. Consequently, their emotional needs may not always be met effectively. In this context, Internet use serves not only as entertainment but also as an important means of emotional regulation, social support, and self-expression. For left-behind children or those with limited offline support, digital environments may play a compensatory role in helping them cope with emotional distress. The social and psychological functions of Internet use among Chinese adolescents may therefore differ from those observed in other cultural settings. This perspective helps explain the observed IA–SI association and underscores the importance of incorporating culturally sensitive prevention strategies.

Although this study provides valuable insights into the IA–SI relationship, it has limitations. First, the cross-sectional design prevents the establishment of causal relationships; therefore, future longitudinal studies are needed. Second, the reliance on self-report questionnaires introduces the potential for social desirability bias, which may affect the accuracy of responses. Third, the small sample size in the severe addiction group (n = 162) resulted in lower statistical power (75.5 percent), which may affect the stability of the estimates for this group. Future research should include larger samples to improve statistical power and ensure more reliable estimates. Finally, although Item 9 of the CDI was used to assess SI,^30,31 it is a single-item measure that may not capture the full complexity of SI, including its frequency, severity, and duration. Future studies may benefit from using multi-item or interview-based assessments to provide a more comprehensive understanding of SI.

Conclusions

In conclusion, IA was significantly associated with SI among Chinese adolescents, particularly among females. Childhood trauma and depressive symptoms influenced this association, highlighting the complexity of suicide risk during adolescence. These findings suggest that IA may represent a potentially important behavioral correlate of SI. Integrating digital behavior assessment with psychological screening may support the early identification of at-risk adolescents and contribute to more comprehensive suicide prevention efforts.

Authors’ Contributions

F.Z. and X.X.: Conceptualization, methodology, data curation, formal analysis, investigation, writing—original draft, and writing—review and editing. Z.H.: Methodology, data curation, and investigation. T.F.: Data curation and investigation. J.L.: Investigation and writing—review and editing. Y.Y.: Methodology, investigation, and supervision. L.K.: Conceptualization, data curation, writing—review and editing, and supervision. All authors read and approved the final article.

Footnotes

Acknowledgment

The authors gratefully acknowledge the participating schools, students, and all individuals who contributed to data collection and coordination. The authors also thank all participants for their voluntary participation and the staff for their help with data collection in this study.

Ethics Approval Statement

All procedures were conducted in accordance with relevant laws and institutional guidelines, and approval was obtained from the Ethics Committee of Shandong University School of Nursing and Rehabilitation (2024-R-157).

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request. The data are not publicly available due to privacy or ethical restrictions.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

The study was supported by the Shandong Excellent Young Scientists Fund Program (Overseas) (Grant No. 2024HWYQ-010) and the Taishan Scholar Foundation of Shandong Province (Grant No. tsqn202211034).

References

State internet information office of the people’s republic of China the 49th statistical report on the development of the internet in China. China Internet Network Information Center. 2022. https://www.cnnic.com.cn/IDR/ReportDownloads/202204/P020220424336135612575.pdf

León Méndez

, Padrón

, Fumero

, et al. Effects of internet and smartphone addiction on cognitive control in adolescents and young adults: A systematic review of fMRI studies. Neurosci Biobehav Rev, 2024; 159:105572; doi: 10.1016/j.neubiorev.2024.105572

Xin

, Xing

, Pengfei

, et al. Online activities, prevalence of internet addiction and risk factors related to family and school among adolescents in China. Addict Behav Rep, 2018; 7:14–18.

, Chen

, Han

, et al. Prevalence and factors of addictive internet use among adolescents in Wuhan, China: Interactions of parental relationship with age and hyperactivity-impulsivity. PLoS One, 2013; 8(4):e61782.

Cao

, Sun

, Wan

, et al. Problematic Internet use in Chinese adolescents and its relation to psychosomatic symptoms and life satisfaction. BMC Public Health, 2011; 11(1):802.

Potenza

. Do gaming disorder and hazardous gaming belong in the ICD-11? Considerations regarding the death of a hospitalized patient that was reported to have occurred while a care provider was gaming. J Behav Addict, 2018; 7(2):206–207.

Guo

, Luo

, Wang

W-X

, et al. Association between problematic internet use, sleep disturbance, and suicidal behavior in Chinese adolescents. J Behav Addict, 2018; 7(4):965–975.

Boldrini

, Xiao

, Singh

, et al. OMICS approaches to investigate the pathogenesis of suicide. Biol Psychiatry, 2024; 96(12):919–928.

Gelaye

, Kajeepeta

, Williams

. Suicidal ideation in pregnancy: An epidemiologic review. Arch Womens Ment Health, 2016; 19(5):741–751.

10.

Liu

, Chen

, Liu

, et al. Prevalence of suicidal behaviour and associated factors in a large sample of Chinese adolescents. Epidemiol Psychiatr Sci, 2019; 28(3):280–289.

11.

Chang

, Shi

, Yao

, et al. Prevalence of suicidal ideation, suicide plans, and suicide attempts among children and adolescents under 18 years of age in mainland china: A systematic review and meta-analysis. Trauma Violence Abuse, 2024; 25(3):2090–2102.

12.

Cheng

Y-S

, Tseng

P-T

, Lin

P-Y

, et al. Internet addiction and its relationship with suicidal behaviors: A meta-analysis of multinational observational studies. J Clin Psychiatry, 2018; 79(4):17r11761.

13.

Kuang

, Wang

, Huang

, et al. Relationship between internet addiction, susceptible personality traits, and suicidal and self-harm ideation in Chinese adolescent students. J Behav Addict, 2020; 9(3):676–685.

14.

Severo

, Soares

, Affonso

, et al. Prevalence and risk factors for internet gaming disorder. Braz J Psychiatry, 2020; 42(5):532–535.

15.

Huang

, Xu

, Mei

, et al. Problematic internet use and the risk of suicide ideation in Chinese adolescents: A cross-sectional analysis. Psychiatry Res, 2020; 290:112963.

16.

Turecki

, Brent

. Suicide and suicidal behaviour. Lancet, 2016; 387(10024):1227–1239.

17.

Zheng

, Feng

, Du

, et al. Specific effects of cumulative childhood trauma on suicidality among youths. J Affect Disord, 2024; 358:260–269.

18.

Grossberg

, Rice

. Depression and suicidal behavior in adolescents. Med Clin North Am, 2023; 107(1):169–182.

19.

Yang

, Sheng

, Ge

, et al. Childhood trauma and psychological sub-health among Chinese adolescents: The mediating effect of internet addiction. BMC Psychiatry, 2022; 22(1):762.

20.

Wang

, Xia

, Wang

, et al. Prevalence and clinical correlates of internet addiction symptoms and their association with quality of life in adolescents with major depressive disorder: A multicenter cross-sectional study. Front Psychiatry, 2022; 13:819704.

21.

, An

, Chen

. Trends and influencing factors in problematic smartphone use prevalence (2012-2022): A systematic review and meta-analysis. Cyberpsychol Behav Soc Netw, 2024; 27(9):616–634.

22.

Madigan

, Korczak

, Vaillancourt

, et al. Comparison of paediatric emergency department visits for attempted suicide, self-harm, and suicidal ideation before and during the COVID-19 pandemic: A systematic review and meta-analysis. Lancet Psychiatry, 2023; 10(5):342–351.

23.

Liu

, Liu

, Deng

, et al. Depression and suicide attempts in Chinese adolescents with mood disorders: The mediating role of rumination. Eur Arch Psychiatry Clin Neurosci, 2023; 273(4):931–940; doi: 10.1007/s00406-022-01444-2

24.

Bellaert

, Simeone

, Zhang

, et al. Sex-specific pathways from early irritability trajectories to later suicidal ideations and behaviors: Findings from the ABCD study®. J Child Psychol Psychiatry, 2026; 67(5):673–685; doi: 10.1111/jcpp.70044

25.

Young

. Psychology of computer use: XL. Addictive use of the internet: A case that breaks the stereotype. Psychol Rep, 1996; 79(3 Pt 1):899–902.

26.

, Wu

, Zhao

, et al. Validation study of Young’s Chinese version of Internet Addiction Scale. Injury Med Electron Ed, 2019; 8:17–23.

27.

Zhao

, Zhang

, Li

, et al. Reliability and validity of the Chinese version of Childhood Trauma Questionnaire. Chin J Clin Rehabil, 2005; 20:105–107.

28.

Smucker

, Craighead

, et al. Normative and reliability data for the Children’s Depression Inventory. J Abnorm Child Psychol, 1986; 14(1):25–39.

29.

, Lu

, Tan

, et al. Reliability and validity of the Chinese version of Children’s Depression Inventory. Chin J Ment Health, 2010; 24(10):775–779.

30.

Voltas

, Hernández-Martínez

, Arija

, et al. Suicidality in a community sample of early adolescents: A three-phase follow-up study. Arch Suicide Res, 2020; 24(sup2):S217–S235.

31.

Houghton

, Cowley

, Kelleher

, et al. Suicidal ideation among Irish adolescents: The CDI suicide question (item 9). Ir J Psychol Med, 2003; 20(3):106.

32.

Guo

, Tao

, Li

, et al. Associations of internet addiction severity with psychopathology, serious mental illness, and suicidality: Large-sample cross-sectional study. J Med Internet Res, 2020; 22(8):e17560.

33.

, Mo

PKH

, Zhang

, et al. Impulsivity, self-control, interpersonal influences, and maladaptive cognitions as factors of internet gaming disorder among adolescents in China: Cross-sectional mediation study. J Med Internet Res, 2021; 23(10):e26810.

34.

Gifuni

, Spodenkiewicz

, Laurent

, et al. Symptoms characteristics of personality disorders associated with suicidal ideation and behaviors in a clinical sample of adolescents with a depressive disorder. Front Psychiatry, 2023; 14:1269744.

35.

Reicher

, Spears

, Postmes

. A social identity model of deindividuation phenomena. Eur Rev Soc Psychol, 1995; 6(1):161–198.

36.

Wong

PW-C

, Fu

K-W

, Yau

RS-P

, et al. Accessing suicide-related information on the internet: A retrospective observational study of search behavior. J Med Internet Res, 2013; 15(1):e2181.

37.

Zhong

, Chan

SSM

, Liu

, et al. Nonfatal suicidal behaviors of Chinese rural‐to‐urban migrant workers: Attitude toward suicide matters. Suicide Life Threat Behav, 2019; 49(5):1199–1208.

38.

Tateno

, Teo

, Shirasaka

, et al. Internet addiction and self-evaluated attention-deficit hyperactivity disorder traits among Japanese college students. Psychiatry Clin Neurosci, 2016; 70(12):567–572.

39.

Fox

, Vendemia

. Selective self-presentation and social comparison through photographs on social networking sites. Cyberpsychol Behav Soc Netw, 2016; 19(10):593–600.

40.

Tang

, Peng

, Ma

, et al. Dispelling stigma and enduring bias: Exploring the perception of esports participation among young women. Humanit Soc Sci Commun, 2025; 12(1):144.

41.

Berardelli

, Sarubbi

, Rogante

, et al. Association between childhood maltreatment and suicidal ideation: A path analysis study. J Clin Med, 2022; 11(8):2179.

42.

, Jian

, Dong

, et al. Childhood trauma and suicidal ideation among Chinese university students: The mediating effect of internet addiction and school bullying victimisation. Epidemiol Psychiatr Sci, 2020; 29:e152.

43.

Kircaburun

, Griffiths

, Billieux

. Psychosocial factors mediating the relationship between childhood emotional trauma and internet gaming disorder: A pilot study. Eur J Psychotraumatol, 2019; 10(1):1565031.

44.

Wang

, Harrer

, Zwald

, et al. Association of recent violence encounters with suicidal ideation among adolescents with depression. JAMA Netw Open, 2023; 6(3):e231190; doi: 10.1001/jamanetworkopen.2023.1190

45.

Choi

, Cho

, Kim

J-Y

, et al. Structural alterations in the prefrontal cortex mediate the relationship between internet gaming disorder and depressed mood. Sci Rep, 2017; 7(1):1245.

46.

Tye

, Mirzabekov

, Warden

, et al. Dopamine neurons modulate neural encoding and expression of depression-related behaviour. Nature, 2013; 493(7433):537–541.

47.

Kågesten

, Gibbs

, Blum

, et al. Understanding factors that shape gender attitudes in early adolescence globally: A mixed-methods systematic review. PLoS One, 2016; 11(6):e0157805.

48.

Hornung

, Heim

. Gene–environment interactions and intermediate phenotypes: Early trauma and depression. Front Endocrinol; 5(14):17; doi: 10.3389/fendo.2014.00014

49.

Melegari

, Barni

, Piperno

, et al. Coping skills in pre- and early adolescents: The role of temperament and character. J Genet Psychol, 2021; 182(6):406–421.