Introduction
Suicide is a fatal public health problem worldwide, and 1.3% overall mortality are the result of suicide in 2019 (Who, Reference Who2021). Particularly, 77% of these deaths happened in low- and middle-income countries. For adolescents, this proportion is up to 88%. In China, over 60,000 adolescents died by suicide in 2018 (Sun et al., Reference Sun, Ren, Li, Jiang, Liu and You2020). Therefore, it has great significance to explore risk factors of suicide for developing targeted prevention programs, especially for adolescents (Liu et al., Reference Liu, Walsh, Sheehan, Cheek and Sanzari2022; Turecki and Brent, Reference Turecki and Brent2016). To date, more scholars support that suicide is a developing process, which includes three stages of suicide behaviours before suicide death: suicide ideation (thoughts concerning ending one’s own life), suicide plans (plans about how to kill oneself) and suicide attempts (actual actions attempted to complete suicide) (Chen et al., Reference Chen, Jiang, Liu, Ran, Yang, Xu, Lu and Xiao2021b; Sveticic and De Leo, Reference Sveticic and De Leo2012). From this perspective, individuals involved in suicide behaviours or not could be classified into four groups: none group (without suicide ideation, plans and attempts), suicide ideator (having suicide ideation only, neither plans nor attempts), suicide planner (having suicide ideation and plans, but no attempts) and suicide attempter (having suicide ideation, plans and attempts concurrently) (Peng et al., Reference Peng, Guo, Cheng, Wang, Tan, Rong, Kang, Ding, Wang and Yu2022).
Nowadays, a lot of prior work has supported that childhood maltreatment is a robust predictor of suicide behaviours (Chen et al., Reference Chen, Li, Cao, Liu, Liu, Chen, Luo, Liang, Guo, Zhong, Wang and Zhou2021a; Grillault et al., Reference Grillault, Godin, Dansou, Belzeaux, Aouizerate, Burté, Courtet, Dubertret, Haffen, Llorca, Olie, Roux, Polosan, Schwan, Leboyer, Bellivier, Marie-Claire and Etain2022; Stagaki et al., Reference Stagaki, Nolte, Feigenbaum, King-Casas, Lohrenz, Fonagy and Montague2022; Van Bentum et al., Reference Van Bentum, Sijbrandij, Saueressig and Huibers2022). Childhood maltreatment is typically defined as any action of commission or omission by parents or other caregivers that brings direct or potential harm, which generally encompasses five core subtypes: physical abuse, emotional abuse, sexual abuse, physical neglect and emotional neglect (Bernstein et al., Reference Bernstein, Stein, Newcomb, Walker, Pogge, Ahluvalia, Stokes, Handelsman, Medrano, Desmond and Zule2003). According to a recent meta-analysis, the pooled prevalence of five childhood maltreatment subtypes for Chinese primary and high school students from high to low is 47% (physical neglect) to 12% (sexual abuse) (Wang et al., Reference Wang, Cheng, Qu, Zhang, Cui and Zou2020).
Since both subtypes of childhood maltreatment and stages of suicide behaviours are not exclusive, disentanglement of the characteristics and pattern of the associations is warranted (Grillault et al., Reference Grillault, Godin, Dansou, Belzeaux, Aouizerate, Burté, Courtet, Dubertret, Haffen, Llorca, Olie, Roux, Polosan, Schwan, Leboyer, Bellivier, Marie-Claire and Etain2022). Although many related previous studies have been conducted in community and clinical samples (Berardelli et al., Reference Berardelli, Sarubbi, Rogante, Erbuto, Giuliani, Lamis, Innamorati and Pompili2022; Grillault et al., Reference Grillault, Godin, Dansou, Belzeaux, Aouizerate, Burté, Courtet, Dubertret, Haffen, Llorca, Olie, Roux, Polosan, Schwan, Leboyer, Bellivier, Marie-Claire and Etain2022; Hassan et al., Reference Hassan, Stuart and De Luca2016; Walsh et al., Reference Walsh, Sheehan and Liu2021; Xie et al., Reference Xie, Wu, Zheng, Guo, Yang, He, Ding and Peng2018), several key research questions remain controversial and inconclusive. First, are the effects of different subtypes of childhood maltreatment on suicide behaviours equivalent or heterogeneous? Some prior literature supports that all forms of childhood maltreatment have positive impact on suicidality (Gong et al., Reference Gong, Zhang, Li, Wang, Wu, Guo and Lu2020; Lee et al., Reference Lee, Kim, Chang, Hong, Lee, Hahm, Cho, Park, Jeon, Seong, Park and Kim2021; Liu et al., Reference Liu, Fang, Gong, Cui, Meng, Xiao, He, Shen and Luo2017; Miller et al., Reference Miller, Esposito-Smythers, Weismoore and Renshaw2013), while others indicate that not all childhood maltreatment subtypes could predict suicide behaviours (Behr Gomes Jardim et al., Reference Behr Gomes Jardim, Novelo, Spanemberg, Von Gunten, Engroff, Nogueira and Cataldo Neto2018; Berardelli et al., Reference Berardelli, Sarubbi, Rogante, Erbuto, Giuliani, Lamis, Innamorati and Pompili2022; Hadland et al., Reference Hadland, Wood, Dong, Marshall, Kerr, Montaner and Debeck2015; Hassan et al., Reference Hassan, Stuart and De Luca2016; Zatti et al., Reference Zatti, Rosa, Barros, Valdivia, Calegaro, Freitas, Cereser, Da Rocha, Bastos and Schuch2017). Second, for a certain childhood maltreatment subtype, are the effects on different involvement of suicide behaviours similar or specific (Liu et al., Reference Liu, Fang, Gong, Cui, Meng, Xiao, He, Shen and Luo2017)? For example, a prior work reveals that frequent sexual abuse victims are more likely to attempt suicide rather than have suicide ideation or plans (Lee et al., Reference Lee, Kim, Chang, Hong, Lee, Hahm, Cho, Park, Jeon, Seong, Park and Kim2021). Third, which subtype of childhood maltreatment is the strongest risk factor of suicide behaviours? Some recent meta-analyses conclude that sexual abuse has stronger effect than other subtypes (Angelakis et al., Reference Angelakis, Austin and Gooding2020, Reference Angelakis, Gillespie and Panagioti2019). However, another meta-analysis demonstrates that emotional abuse could be the most robust predictor (Liu et al., Reference Liu, Fang, Gong, Cui, Meng, Xiao, He, Shen and Luo2017). Fourth, whether the associations differ across sex and residence (Liu et al., Reference Liu, Fang, Gong, Cui, Meng, Xiao, He, Shen and Luo2017; Mcmahon et al., Reference Mcmahon, Hoertel, Olfson, Wall, Wang and Blanco2018; Salokangas et al., Reference Salokangas, Luutonen, Heinimaa, From and Hietala2019)? For example, some literature supports that adult females have higher prevalence of suicide ideation and suicide attempts than males, and rural China has three times higher rates of suicide than do urban China (Phillips et al., Reference Phillips, Yang, Zhang, Wang, Ji and Zhou2002; Turecki and Brent, Reference Turecki and Brent2016). However, it is little known among Chinese adolescents.
To date, few previous studies have explored the specific impact of five childhood maltreatment subtypes on suicide ideation, plans and attempts between urban and rural adolescents in China. To fill the gaps, we conducted a nationwide study based on a large-size sample of Chinese adolescents across five representative provinces. The primary aim is to quantify the associations between five subtypes of childhood maltreatment (physical abuse, emotional abuse, sexual abuse, physical neglect and emotional neglect) and different suicide behaviour involvement (none group, suicide ideators, suicide planners and suicide attempters). Furthermore, the second aim is to examine the potential sex difference and residence difference in the associations between childhood maltreatment and suicide behaviours.
Methods
Study design and data collection
A multistage cluster sampling method was adopted from April to December 2021. In Stage 1, China was divided into five main geographic regions (eastern, southern, western, northern and central regions). Five provinces (Jiangsu, Guangdong, Yunnan, Gansu and Hubei) were randomly chosen in each region (Figure S1 in Supplementary material). In Stage 2, two cities were selected randomly in every selected province. In Stage 3, we selected one district in the urban areas and one county in the rural areas from each selected city. In Stage 4, one junior high school and one senior high school were selected randomly in each sample district or county. In Stage 5, we randomly chose four or six classes from each grade (7th to 12th) in each selected school based on enrolment size (Figure S2 in Supplementary material). Finally, all students in the selected class were invited to participate in the research voluntarily. Informed written consents were signed by every participant and their guardians before the filed survey.
Assessment
A custom-designed questionnaire was designed to collect demographic characteristics, high-risk health-related behaviours, suicide behaviours, childhood maltreatment and other psychological problems. Demographic variables included sex (male or female), residence (urban or rural), age (years old), grade (7th to 12th), family composition (living in a family with two biological parents, one biological parent or others), father and mother’s education level (primary school or less, junior high school, senior high school and college or more), family income (average family income per month in RMB, less than 1999, 2000–3999, 4000–5999, 6000–7999 and more than 8000). High-risk health-related behaviours included smoking (yes or no) and drinking alcohol (yes or no) (Turecki and Brent, Reference Turecki and Brent2016).
Suicide behaviours
Suicide behaviours were measured based on three items regarding suicidality derived from the Global School–Based Student Health Survey (Ma et al., Reference Ma, Guo, Guo, Jiao, Zhao, Ammerman, Gazimbi, Yu, Chen, Wang and Tang2021). Three stages of suicide behaviours were measured by the three following questions in the past year. Suicide ideation: “Have you ever seriously had thoughts of committing suicide?” Suicide plans: “Have you ever made a specific plan about how you would kill yourself?” Suicide attempts: “Have you ever tried to commit suicide?” The answer of these questions was defined as No (0 times) vs. Yes (1 to 4 times) (Lee et al., Reference Lee, Kim, Chang, Hong, Lee, Hahm, Cho, Park, Jeon, Seong, Park and Kim2021; Wan et al., Reference Wan, Chen, Ma, Mcfeeters, Sun, Hao and Tao2019). Due to some participants could simultaneously have suicide ideation, plans and (or) attempts, we classified all sample into four categories: None group (without any suicide behaviours), Suicide ideator (having suicide ideation only), Suicide planner (having suicide ideation and plans but no attempts) and Suicide attempter (having all suicide ideation, plans and attempts) (Sveticic and De Leo, Reference Sveticic and De Leo2012; Wang et al., Reference Wang, He, Yu, Qiu, Yang, Qiao, Sui, Zhu and Yang2014).
Childhood maltreatment subtypes
Childhood maltreatment subtypes were measured through the Childhood Trauma Questionnaire-Short Form (CTQ-SF) (Bernstein et al., Reference Bernstein, Stein, Newcomb, Walker, Pogge, Ahluvalia, Stokes, Handelsman, Medrano, Desmond and Zule2003), which contained five subscales: physical abuse, emotional abuse, sexual abuse, physical neglect and emotional neglect. Each subscale had five items, and each item was rated on a 5-point Likert scale (1 = never true, 2 = rare true, 3 = sometimes true, 4 = often true and 5 = always true). Therefore, the score of each subscale ranged from 5 to 25, and greater scores indicated more severe history of abuse or neglect. The Cronbach’s alpha coefficient of the CTQ-SF in the current study was 0.85.
Other psychological problems
Other psychological problems included depression and anxiety. The nine-item Patient Health Questionnaire (PHQ-9) was used to measure major depression disorders (Kroenke et al., Reference Kroenke, Spitzer and Williams2001). Each item had a respond scored from 0 to 3. Total score ranged from 0 to 27, and a higher score of PHQ-9 indicated that the participant has more major depressive symptoms (Adewuya et al., Reference Adewuya, Ola and Afolabi2006). In the current study, the Cronbach’s alpha of PHQ-9 was 0.90. The seven-item Generalized Anxiety Disorder Scale (GAD-7) was used to screen generalized anxiety disorders (Spitzer et al., Reference Spitzer, Kroenke, Williams and Lowe2006). Each item had a respond scored from 0 to 3. Therefore, total score of GAD-7 ranged from 0 to 21, and a higher total score showed that the participant has more generalized anxiety disorders (Lowe et al., Reference Lowe, Decker, Muller, Brahler, Schellberg, Herzog and Herzberg2008). In this study, the Cronbach’s alpha for GAD-7 was 0.94.
Data analyses
First, demographic characteristics of sample and the proportion of suicide behaviour involvement were described by descriptive statistics. Continuous variables were depicted by mean (SD). Second, the chi-square test was used to compare the proportion of suicide behaviour involvement across different categories. ANOVA was used to compare the mean scores between different groups. Spearman’s correlation was used between two variables.
Third, to assess the associations between each subtype of childhood maltreatment and suicide behaviour involvement, a set of Multinomial logistic regression analysis was conducted for every subtype separately to estimate odds ratios (ORs) and 95% confidence intervals (95% CIs). Besides, to examine the influence of sex × residence in the associations, we conducted a subgroup analysis among urban males, urban females, rural males and rural females, separately. The significance level was set at p < 0.05, and all tests were two-sided. These analyses mentioned above were ran by SPSS 26.0.
Fourth, given potential multicollinearity between five subtypes of childhood maltreatment (Vachon et al., Reference Vachon, Krueger, Rogosch and Cicchetti2015), Structural equation modelling was conducted to assess direct effects of five childhood maltreatment subtypes on suicide behaviours when we simultaneously entered all five subtypes into the model and adjusted the interaction between every two subtypes. Standardized β was used to estimate effect size. Structural equation modelling was ran by AMOS 21.0.
Results
Characteristics of participants
Of 21,207 respondents who sent back the questionnaire and provided the consent form, 831 were excluded because their age was over 18 or less than 12 years, 952 were excluded because they did not respond to any items regarding suicide behaviours (three items in total), 444 were excluded because the missing data of the whole questionnaire was over 15%. Finally, 18,980 participants’ questionnaires were qualified, and the actual valid response rate was 89.50% (18,980/21,207).
A total of 18,980 participants were almost equally distributed across sex (9472 males [49.9%] vs. 9508 females [50.1%]) and residence (9539 from urban areas [50.3%] vs. 9441 from rural areas [49.7%]). Their age ranged from 12 to 18, and mean (SD) of age was 14.98 (1.64). Other socio-demographic characteristics of the sample are presented in Table 1.
Table 1. General characteristics of participants and proportion of suicide behaviours
Note: PHQ-9: the nine-item Patient Health Questionnaire and GAD-7: the seven-item Generalized Anxiety Disorder Scale.
Proportion of suicide behaviours involvement
Among 18,980 participants, 4630 (24.4%) reported suicide behaviours in the past year. Specifically, 2021 (10.6%) were suicide ideators, 1595 (8.4%) were suicide planners and 1014 (5.3%) were suicide attempters (Table 1). The proportion of suicide behaviour involvement was significantly different by all variables, except for father’s education (p = 0.173) and mother’s education (p = 0.070). Participants across sex × residence had significantly different proportion of suicide behaviours involvement (Table 2). Particularly, rural females had the highest proportion of suicide ideator (13.8%) and planner (11.5%) according to pairwise comparisons. Besides, scores of five childhood maltreatment subtypes were significantly different between males and females across urban and rural areas.
Table 2. The proportion of suicide behaviours and the score difference of childhood maltreatment among participants by sex × residence
Note: a, urban male; b, urban female; c, rural male; d, rural female in pairwise comparison.
Multinomial logistic regression analysis of suicide behaviours involvement
Spearman’s correlations between suicide behaviours and other variables are displayed in Table 3. Results of multinomial logistic regression analysis are shown in Table 4. For total participants, all subtypes of childhood maltreatment were significantly associated with suicide behaviour involvement, except for the ORs between sexual abuse and suicide ideator as well as suicide planner (p > 0.05). In subgroup analysis, sexual abuse was significantly associated with suicide attempter for urban males (OR = 1.184, 95% CI 1.121–1.251, p < 0.001), urban females (OR = 1.167, 95% CI 1.046–1.302, p = 0.006) and rural females (OR = 1.112, 95% CI 1.107–1.197, p = 0.010), except for rural males (p = 0.084). Physical neglect was significantly associated with suicide ideator only for urban males (OR = 1.041, 95% CI 1.003–1.081, p = 0.036). Heat map of associations between childhood maltreatment subtypes and suicide behaviour involvement among participants by sex × residence is depicted in Figure S3 (in Supplementary material).
Table 3. Correlations between suicide behaviours and other variables
*** Note:p < 0.001.
PHQ-9: the nine-item Patient Health Questionnaire and GAD-7: The seven-item Generalized Anxiety Disorder Scale.
Table 4. Multinomial logistic regression of suicide behaviours among participants by sex × residence (OR [95% CI])
Note: Model 1 was adjusted for sex, residence, age, grade, family composition, family income, smoking, drinking alcohol, depression and anxiety. Model 2 to Model 5 were adjusted for age, grade, family composition, family income, smoking, drinking alcohol, PHQ-9 score and GAD-7 score. Each subtype of childhood maltreatment was the only independent variable entered into the models.
Structural equation modelling for suicide behaviours
Figure S4 (in Supplementary material) showed the results of structural equation modelling. After controlling for covariates and the interaction between five subtypes of childhood maltreatment, the direct effects of childhood maltreatment subtypes on suicide behaviours from high to low were emotional abuse (β = 0.363, p < 0.001), physical abuse (β = 0.100, p < 0.001) and sexual abuse (β = 0.033, p = 0.003). However, the effects of physical neglect (β = 0.036, p = 0.507) and emotional neglect (β = 0.029, p = 0.587) on suicide behaviours were not significant.
Discussion
This is the first study to quantify the associations between five core subtypes of childhood maltreatment and three stages of suicide behaviours among Chinese adolescents. The current study has several key and new findings. First, different subtypes of childhood maltreatment have specific and unique associations with suicide behaviours. For example, sexual abuse is significantly related to suicide attempter but is not significantly related for suicide ideator and planner. This result may reveal an acute effect of sexual abuse on suicidality. Second, the associations between childhood maltreatment subtypes and suicide behaviours differ between males and females across urban and rural China. Third, the direct effects of childhood maltreatment subtypes on suicide behaviours are non-equivalence. Specifically, emotional abuse has the greatest effect, while the effects of physical and emotional neglect are not significant. These findings could benefit for policy-makers in preventing suicidality for Chinese adolescents with or without childhood maltreatment experiences.
Our results demonstrate that the associations between different childhood maltreatment subtypes and suicide behaviour involvement vary considerably, which is congruent, with some existing literature concluded that various forms of childhood maltreatment maintain a unique and independent influence on suicidal risk (Angelakis et al., Reference Angelakis, Austin and Gooding2020; Zatti et al., Reference Zatti, Rosa, Barros, Valdivia, Calegaro, Freitas, Cereser, Da Rocha, Bastos and Schuch2017). It might be possible that different subtypes of childhood maltreatment evoke the feeling of constraint and desperation in a different way (Mcmahon et al., Reference Mcmahon, Hoertel, Olfson, Wall, Wang and Blanco2018; O’Connor and Portzky, Reference O’connor and Portzky2018). On contrary, some prior literature shows that all subtypes of childhood maltreatment have a similar or equivalent effect on suicide behaviours (Guo et al., Reference Guo, Wang, Gao, Huang, Li and Lu2018; Hoertel et al., Reference Hoertel, Franco, Wall, Oquendo, Wang, Limosin and Blanco2015). The inconsistency may stem from different study designs (prospective or cross-sectional), sampling methods (random or convenient) and mixed samples (adolescents or adults, community or clinical samples) (Miché et al., Reference Miché, Hofer, Voss, Meyer, Gloster, Beesdo-Baum, Wittchen and Lieb2020). On the other hand, the definition and measurement of childhood maltreatment are varied across previous studies (Grillault et al., Reference Grillault, Godin, Dansou, Belzeaux, Aouizerate, Burté, Courtet, Dubertret, Haffen, Llorca, Olie, Roux, Polosan, Schwan, Leboyer, Bellivier, Marie-Claire and Etain2022; Miller et al., Reference Miller, Esposito-Smythers, Weismoore and Renshaw2013; Park et al., Reference Park, Shekhtman and Kelsoe2020). The most widely used assessment of five childhood maltreatment subtypes is the CTQ-SF (Bernstein et al., Reference Bernstein, Stein, Newcomb, Walker, Pogge, Ahluvalia, Stokes, Handelsman, Medrano, Desmond and Zule2003). However, some scholars treat subtypes of childhood maltreatment as five dichotomous variables (Ahouanse et al., Reference Ahouanse, Chang, Ran, Fang, Che, Deng, Wang, Peng, Chen and Xiao2022; Chen et al., Reference Chen, Jiang, Liu, Ran, Yang, Xu, Lu and Xiao2021b; Xie et al., Reference Xie, Wu, Zheng, Guo, Yang, He, Ding and Peng2018), while others treat them as five continuous variables (Gong et al., Reference Gong, Zhang, Li, Wang, Wu, Guo and Lu2020; Guo et al., Reference Guo, Wang, Gao, Huang, Li and Lu2018), which could result in heterogeneity across the existing studies.
Some previous studies believe that sexual abuse has a stronger effect on suicidality than other subtypes (Angelakis et al., Reference Angelakis, Austin and Gooding2020; Mcmahon et al., Reference Mcmahon, Hoertel, Olfson, Wall, Wang and Blanco2018). However, after considering the potential multicollinearity among five subtypes of childhood maltreatment, our results indicate that emotional abuse might have the most harmful effect on suicide behaviours. The difference may be related to the cultural and social background between China and other countries (Stoltenborgh et al., Reference Stoltenborgh, Van Ijzendoorn, Euser and Bakermans-Kranenburg2011). Traditional Chinese culture, such as Confucian values, strongly inhibits the disclosure of sexual abuse experiences, especially for adolescents (Ji et al., Reference Ji, Finkelhor and Dunne2013). On the other hand, compared with physical and sexual abuse, emotional abuse might be hard to detect because of less obvious signs. However, emotional abuse in adolescents can lead to shame, hopelessness, avoidance and a dysfunctional attachment pattern, which will increase the risk of depression and anxiety. These two psychological disorders are well-recognized risk factors of suicidality (Chen et al., Reference Chen, Jiang, Liu, Ran, Yang, Xu, Lu and Xiao2021b; Lee et al., Reference Lee, Bae, Rim, Lee, Chang, Kim and Won2018). In addition, some specific behaviours of emotional abuse, such as parental rejection and low parental warmth or responsiveness, may lead to low self-esteem, a negative coping style and a desperate worldview (Lee et al., Reference Lee, Kim, Chang, Hong, Lee, Hahm, Cho, Park, Jeon, Seong, Park and Kim2021). Thus, emotional abuse could be a strong predictor of internal psychopathology development and might disturb the psychosocial well-being during childhood and adolescence (Ahouanse et al., Reference Ahouanse, Chang, Ran, Fang, Che, Deng, Wang, Peng, Chen and Xiao2022).
Notably, we found that sexual abuse is only significantly related to suicide attempter but not for suicide ideator and planner. However, other four subtypes of childhood maltreatment are concurrently associated with suicide ideator, planner and attempter. This finding may mean that the effect of sexual abuse on suicide behaviours is acute and is different between other childhood maltreatment forms. For victims of sexual abuse, they may have suicide ideation, plans and attempts simultaneously in a short period of time and become a suicide attempter. Therefore, they are less likely to be a suicide ideator or planer. The innovative finding should be confirmed with more follow-up studies (Lee et al., Reference Lee, Kim, Chang, Hong, Lee, Hahm, Cho, Park, Jeon, Seong, Park and Kim2021). Regarding sex difference, some previous studies believe that the associations between childhood maltreatment subtypes and suicidality differ between males and females (Liu et al., Reference Liu, Fang, Gong, Cui, Meng, Xiao, He, Shen and Luo2017; Mcmahon et al., Reference Mcmahon, Hoertel, Olfson, Wall, Wang and Blanco2018; Miché et al., Reference Miché, Hofer, Voss, Meyer, Gloster, Beesdo-Baum, Wittchen and Lieb2020; Salokangas et al., Reference Salokangas, Luutonen, Heinimaa, From and Hietala2019). However, to our knowledge, little is known that whether the associations differ between urban and rural adolescents. In the current study, it is interesting to note that the proportion of suicide behaviour involvement is significantly different between males and females across urban and rural China. Notably, rural females may be the most vulnerable population of suicidality. Therefore, suicide prevention programs could give priority to female adolescents in rural China.
Limitations
Several limitations should be acknowledged. First, the cross-sectional design of the current study does not allow us to draw any causal relationships between childhood maltreatment subtypes and suicide behaviours. Future studies could benefit from a longitudinal design. Second, the self-reported questionnaire in the current study has some sensitive questions for Chinese adolescents. Participants who submitted incomplete questionnaires without reports of suicide behaviours or (and) childhood maltreatment, especially for sexual abuse, may be more likely to suffer suicide risk. Future research could take multi-informant measures to obtain these sensitive information, such as prospective caregiver reports (usually from the mother) (Newbury et al., Reference Newbury, Arseneault, Moffitt, Caspi, Danese, Baldwin and Fisher2018). Third, other important information of childhood maltreatment was not evaluated, such as the age of the first maltreatment experience, duration, intensity, frequency and perpetrator of abuse or neglect (Grillault et al., Reference Grillault, Godin, Dansou, Belzeaux, Aouizerate, Burté, Courtet, Dubertret, Haffen, Llorca, Olie, Roux, Polosan, Schwan, Leboyer, Bellivier, Marie-Claire and Etain2022). In the next step, we will conduct a survey to collect more related information to further explore the impact of timing, duration, severity and offender of childhood maltreatment (Duarte et al., Reference Duarte, Belzeaux, Etain, Greenway, Rancourt, Correa, Turecki and Richard-Devantoy2020). Finally, although the study highlights independent and specific effects of different childhood maltreatment subtypes, the co-occurrence of five subtypes is ignored. In future research, a latent class (profile) analysis should be used to examine the relationship between co-occurrence of multiple childhood maltreatment subtypes and suicidality (Luk et al., Reference Luk, Bond, Gabrielli, Lacroix, Perera, Lee-Tauler, Goldston, Soumoff and Ghahramanlou-Holloway2021; Witt et al., Reference Witt, Münzer, Ganser, Fegert, Goldbeck and Plener2016).
Implications
A better understanding of the associations between different childhood maltreatment subtypes and suicide behaviours could advance the development of efficient preventive strategies. Our findings have some practical implications for Chinese adolescents. First, given that not all childhood maltreatment subtypes is significantly associated with suicide ideator, planner and attempter. The finding emphasizes the importance of differentiating the specificity of different forms of childhood maltreatment, which can help decision-maker to develop a tailored preventive approach for suicidality. For example, sexual abuse is a risk factor of suicide attempter but not for suicide ideator and planner. We should take immediate measures for those who have history of sexual abuse to keep them away from dangerous stuffs, such as knives and sleeping pills. Then, more efforts could be taken to reduce their chances of being suicide attempter. Second, emotional abuse may be the strongest predictor of suicide behaviours in five subtypes, while the effects of physical neglect and emotional neglect are not significant. Therefore, suicide interventions could focus on those who have history of emotional abuse. Increasing parental support for children and positive interaction with offspring will be a promising approach to weaken the effect of emotional abuse and prevent suicide behaviours. Third, the associations between childhood maltreatment subtypes and suicide behaviours differ by sex and residence. This finding may encourage health professionals to implement unique strategies between males and females across urban and rural China. For instance, among rural males, we could pay more attention to those who experienced physical abuse, emotional abuse and emotional neglect rather than sexual abuse when we only have limited resources for suicide intervention.
Conclusions
Five core subtypes of childhood maltreatment have specific and non-equivalence associations with suicide ideator, planner and attempter in Chinese adolescents. The finding highlights the specificity and severity of different childhood maltreatment forms on suicide behaviours. Emotional abuse may have the strongest effect and sexual abuse have an acute effect on suicide behaviours. Suicide prevention programs could focus on those who experienced emotional, physical and sexual abuse. Furthermore, the associations between childhood maltreatment subtypes and suicide behaviours differ between males and females across urban and rural China. Intervention strategies of suicide should be tailored by sex and residence. Particularly, rural females may be the most vulnerable population of suicidality, and they deserve more attention.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S2045796023000604
Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Acknowledgements
The authors thank students who took part in the survey, parents who supported the work, teachers who assisted with the field investigation and all investigators.
Author contributions
Chang Peng and Junhan Cheng contributed equally to this study.
Financial support
This work was supported by grants from the National Natural Science Foundation of China (grant number 81973066 and 82173541). The funding body had no role in the design of the study and collection, analysis and interpretation of data or in writing the manuscript.
Competing interests
None.
Ethical standards
The study procedures were carried out under the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. The study was approved by the Medical Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (No. 2021-A216).
Open access
