Dialectical behavioral therapy-based prevention of non-suicidal self-harm among secondary students in China: a randomized controlled trial

Abstract

Background

Non-suicidal self-injury (NSSI) in adolescents is a significant public health issue worldwide. Dialectical Behavior Therapy (DBT) is an evidence-based intervention for treating NSSI. However, there are currently few independent studies on DBT interventions specifically targeting NSSI, particularly in Shantou city, within the Chaoshan area of southeast China. This study aimed to evaluate the effectiveness of DBT in reducing NSSI among secondary school students.

Methods

A total of 220 7th-grade students from four classes in two schools in Shantou City were selected using randomized cluster sampling. Two classes from one school were assigned to the DBT-based intervention group, while two classes from the other school were assigned to the control group, which received general education intervention. Comparative analyses of quantitative and categorical data were conducted using Student’s t-test and chi-square tests.

Results

The occurrence of NSSI in DBT group and control group was 23.60% and 24.50%, respectively. Most students who engaged in NSSI sustained mild injuries (71.43%). The occurrence of NSSI in the DBT group decreased from 23.60% to 10.91% after the DBT-based interventions (χ² = 6.23, P < 0.05). The frequency of NSSI in the DBT group was also reduced (from 1.57 ± 4.411 to 0.54 ± 1.895) after the intervention(t = 3.628, P<0.001). In the DBT group, depression scores significantly decreased from 9.59 ± 11.15 to 6.51 ± 3.08 (t = 2.217, P < 0.05). The total Sub-Optimal Health Status (SHS) score was also significantly reduced, from 16.43 ± 12.74 to 12.04 ± 7.80 (t = 3.083, P < 0.05). Among its subdimensions, the psychological state score declined from 5.37 ± 4.78 to 3.70 ± 4.08 (t = 2.792, P < 0.05), while the fatigue score dropped from 7.06 ± 5.28 to 5.06 ± 3.92 (t = 3.186, P < 0.05). However, the control group did not show significant differences in the above aspects after the conventional interventions.

Conclusions

The six-month DBT-based intervention markedly reduced NSSI and improved mental-health indicators, suggesting that DBT offers an effective, scalable framework for school-based prevention in culturally diverse Chinese settings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-24978-1.

Introduction

Non-suicidal self-injury (NSSI) refers to self-harming behavior that does not result in direct fatal consequences. Psychiatrist Karl Menninger first conceptualized the behavior, describing it as “incomplete suicide” [1]. NSSI significantly differs from suicide in three key aspects: intent, repetition, and lethality [2, 3].

First, in terms of intent, there is no desire to die when engaging in NSSI [4]. Some individuals use NSSI as a way to signal distress, draw attention from others, and indirectly seek assistance [5]. Additionally, NSSI exists on a continuum of self-injury. Some individuals who contemplate suicide may gradually become desensitized through engaging in NSSI, potentially leading to suicide. The risk of suicide among individuals with NSSI is 60 times higher than in the general population [6]. Thirdly, NSSI often exhibits repetition, with individuals who have a history of NSSI being more likely to engage in it again [7, 8]. NSSI typically involves low-lethality behaviors, such as hitting limbs, cutting or rubbing the skin, biting oneself, or punching hard objects like walls or glass [9]. The fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) classifies NSSI as a clinical disorder [10]. DSM-5 categorizes the triggers of NSSI into four types: (1) psychological triggers, such as negative emotions or thoughts (depression, anxiety, tension, anger, etc.); (2) impulsivity; (3) thoughts of self-harm; (4) attempts to alleviate interpersonal problems or negative emotions.

NSSI is a critical issue faced by adolescents [11, 12]. Adolescents tends to begin engaging in NSSI between the ages of 11 and 13 with its prevalence peaking between ages 14 and 15 [13, 14]. Globally, the prevalence of NSSI ranges from 3.1% to 15.5% among adolescents aged 12 to 21 years [15–19]. In developing countries, the prevalence of adolescent NSSI is higher, with rates ranging from 15.5% to 31.3% [20]. In China, the prevalence of NSSI has been rising since 2008 [21].

In the mid-1980s, Berkman introduced the concept of a third health state - Sub-optimal Health Status (SHS) [22], also called “sub-health status” – that is distinct from both health and illness. SHS is characterised by subjective discomfort in the absence of detectable clinical pathology and signals a latent vulnerability to illness, including early structural decline, reduced physiological reserves, and psychological disequilibrium [23]. Rapid socioeconomic and lifestyle changes in China have markedly increased the burden of SHS [24]. Among adolescents, cultural shifts, academic pressure, and evolving behaviour patterns have further exacerbated the problem [25]. Although general health status is closely linked to suicidal ideation [26], the relationship between SHS and NSSI remains under-explored. A survey of middle-school students revealed a positive association between SHS and self-injurious behaviour [27]. Adolescents with SHS were 4.23 times more likely to engage in NSSI than their healthy peers (OR = 4.225, 95% CI: 2.983–5.984) [28]. A recent Chinese cross-sectional study further demonstrated that the psychological sub-health component of SHS significantly increased NSSI risk [29]. Moreover, an intervention trial in adolescents with depression and NSSI found that positive-psychology treatment not only improved depression and SHS but also significantly reduced the incidence of NSSI [30]. Collectively, these findings suggest that early identification and amelioration of SHS may serve as an upstream strategy for NSSI prevention, and routine SHS monitoring should be integrated into adolescent-health programmes.

NSSI is not only a social concern but also a pressing public-health problem. Dialectical Behavior Therapy (DBT) is a cognitive-behavioral treatment that blends behavioral science, dialectical philosophy, and Zen practices [31]. Although originally developed for borderline personality disorder, DBT is now widely employed to address suicidal behaviour [32]. The standard DBT program comprises four structured components: skills training group, individual psychotherapy, telephone consultation, and a therapist consultation team [33]. Among these, the skills training group is considered the program’s cornerstone and covers four modules: mindfulness, interpersonal effectiveness, emotion regulation, and distress tolerance [34]. DBT facilitates change through dialectical learning, emphasizing emotion regulation, distress tolerance, and the cultivation of adaptive cognitions [35]. The approach is applicable across a wide age range [36]. In a United States trial involving high-risk suicidal adolescents, six months of DBT markedly reduced suicidal ideation [37]. A subsequent study confirmed similar reductions in suicidal ideation and concomitant improvements in depression, impulsivity, and trauma-related symptoms within the same population [38] Given its demonstrated effects on self-harm and related risk factors, DBT represents a promising intervention for NSSI.

Empirical studies that evaluate DBT specifically for adolescent NSSI remain scarce in China [39]. Shantou City, situated in the Chaoshan sub-region of south-eastern China, provides a unique linguistic and cultural context for such work [40]. The local Chaoshan dialect creates pronounced clinician–patient communication barriers [41]. Although cross-cultural comparisons show broadly similar NSSI prevalence [42], culture-specific risk factors can vary [43, 44], potentially moderating treatment response. In the Chaoshan region, fluency in the local dialect is considered a social obligation, and some patients may refuse diagnosis or treatment if the physician does not speak the dialect. Similarly, healthcare providers who speak the dialect have been observed to respond impatiently when patients rely on other languages or dialects [41]. Such reciprocal linguistic mismatches shape help-seeking behaviour and may attenuate the efficacy of DBT or other psychosocial interventions. Evaluating DBT in Chaoshan could therefore guide local policy on adolescent NSSI prevention while offering lessons for other multilingual regions of China.

Traditional DBT was originally designed as a modular intervention for individuals with complex psychological conditions, enabling adaptability across diverse populations and settings [33, 45]. Empirical evidence confirms the efficacy and role of its core component—skills training groups—in addressing borderline personality disorder (BPD), depression, attention-deficit/hyperactivity disorder (ADHD), and NSSI [46]. In one clinical trial, DBT incorporating a skills-training group significantly outperformed protocols lacking this component in reducing self-harm and alleviating psychiatric symptoms [47]. Furthermore, another study also suggest that DBT skills training groups conducted in school settings can reduce adolescents’ intentions to engage in risky behaviors [48]. Guided by this evidence, the present study adopts DBT skills training to curb NSSI among secondary-school students. Effectiveness will be assessed via NSSI incidence, SHS, and indices of psychological wellbeing, aiming to provide theoretical support for the prevention and management of NSSI among Chinese adolescents.

Methods

Participants and Procedure

A total of 220 7th-grade students from four classes in two schools in Shantou City were selected using randomized cluster sampling. Participants were enrolled in the study only if written informed consent was obtained from both the students and their parents or legal guardians. To prevent information leakage, each group was drawn from two classes in different schools. Using a cluster randomization design, two intact classes from one school were allocated to receive the DBT-based intervention, whereas two parallel classes from other school were assigned to the control group, which received general education intervention.

To safeguard data confidentiality, protective measures were implemented during recruitment and data collection. During recruitment, each participant received a unique identification code to replace personal identifiers, and all personally identifiable information (e.g., names and contact details) was stored separately from research data in secure, access-restricted files. Participants were informed in writing that their data would remain strictly confidential and would be used solely for research purposes. During data collection, electronic files were accessible only to authorized personnel, and all data were transmitted and stored using encrypted systems.

Sample size calculation for the intervention study was determined using the formula:

Where:

λ represents the critical value derived from statistical tables based on significance level (α = 0.05), power (1-β = 0.8), and degrees of freedom (ν = k-1 = 1), yielding λ = 7.85. Angular transformations were calculated in radians. p_max and p_min denote the maximum and minimum proportions respectively.

When only the proportion difference (p_d) is known:

Although DBT’s effectiveness for NSSI has been demonstrated, evidence among Chinese adolescents remains limited, particularly in the linguistically distinctive Chaoshan region; intervention effects may therefore diverge from previous estimates [38, 39, 46]. A prior Chinese study reported an initial 28.89% reduction in the incidence of NSSI after a DBT intervention [49]. To ensure the reliability of our findings, we conservatively estimated the intervention effect size and thus hypothesized a 20% reduction in NSSI incidence. (p_d =0.2):

This yielded a calculated sample size of N ≈ 58 per group. Accounting for a 20% anticipated attrition rate and eligibility threshold. Therefore, minimum of 65 participants per group was required to ensure adequate statistical power, rounded to the nearest whole number considering practical constraints.

The sample size for both the intervention and control groups was 110. Baseline data for this study were collected between April and June 2022. At the end of the intervention, participants were surveyed again using the same questionnaire as in the baseline survey to assess the intervention effect.

The DBT group received a six-month DBT-based intervention, comprising six skills training group sessions conducted monthly from July to December 2022. Each session lasted approximately 30 to 40 min and was delivered in person by the interventionist, who had received formal training in DBT. To ensure consistency, all students in the DBT group received the intervention simultaneously in a single classroom. The DBT-based intervention incorporated multiple formats, including lectures, poster distribution, and video presentations. The content focused on core DBT modules: mindfulness, distress tolerance, emotion regulation, and interpersonal effectiveness. From July to December 2022, the control group received monthly general health education lectures over a six-month period. Each session lasted approximately 30–40 min and was conducted in a classroom group format, primarily through lectures supplemented by poster and video. The sessions were delivered in person by the same trained interventionist who led the DBT sessions. These lectures were matched to the DBT group in frequency, duration, total length, delivery format, and personnel. Their content included topics such as the definition, causes, and risk factors of NSSI, mental wellness strategies, peer pressure, and help-seeking (Table 1). Thus, the total amount of contact time was equivalent to that of the DBT group, with the only difference being the specific content. After the intervention, both groups completed the same questionnaires used at baseline to assess post-intervention outcomes.

Table 1.

Description of skills training group sessions

Session	DBT Group
1	Introduction to DBT: Introduce the theoretical underpinnings of DBT, including its dialectical philosophy, treatment hierarchy, primary strategies, intervention procedures, and therapeutic goals.
2	Mindfulness: Review the definition and benefits of mindfulness. Guide students through practical mindfulness training (e.g., mindful breathing, present-moment awareness).
3	Emotion Regulation: Help students identify, label, and appropriately express emotions. Introduce techniques for managing negative affect and increasing engagement in activities that elicit positive emotional experiences.
4	Distress Tolerance: Promote understanding of distress and its origins. Teach students specific distress tolerance skills such as distraction and self-soothing to enhance emotional resilience.
5	Interpersonal Effectiveness: Emphasize the importance of healthy peer relationships. Teach assertiveness, boundary-setting, and communication skills to improve interpersonal functioning and reduce relational stress.
6	Consolidation and Skills Integration: Review DBT skills taught in previous sessions and support students in integrating these into individualized coping strategies for long-term application.

Measures

Demographic assessment

The demographic data included the participants’ name, gender, weight, height, household economy, family residence, only-child status, and personality. The body mass index (BMI) was calculated using the formula: .

NSSI behavioral assessment

The Chinese version of the Adolescent Self-Injurious Behavior Scale [50] was used to screen for NSSI, including assessments of the modality, frequency, and level of physical harm. For this study, the scale specified the behavior occurred within the past twelve months. This scale includes 17 types of NSSI behaviors and an additional option for self-reporting others, such as hitting oneself to the point of bruising, punching hard surfaces like walls or glass, headbanging, forcefully pulling one’s hair, and cutting the skin with sharp objects like glass or knives. The frequency of NSSI is assessed on a four-level scale: none, once, 2–4 times, and 5 or more times. Recognizing the cognitive level of secondary school students and the potential for reporting bias, this study categorized the degree of physical harm resulting from NSSI into three levels as outlined in the original scale. The investigator thoroughly explained the criteria for each level to the students during the survey. Specifically, “mildly” denoted minimal or no bodily damage; “moderately” signified physical harm not requiring medical intervention; and “severely” indicated bodily harm warranting medical attention or hospitalization. Individuals who had engaged in any of the aforementioned NSSI within the past year were classified as engaging in NSSI during the baseline survey, and those who engaged in NSSI during the past six months of the intervention period were similarly classified. In this study, Cronbach’s alpha coefficient for internal consistency was 0.826, and the KMO value was 0.823. The internal consistency (Cronbach’s alpha) and sampling adequacy (KMO values) for all scales were calculated based on the baseline data to assess the reliability of the measures prior to the intervention.

SHS assessment

The SHS of secondary school students was assessed using the Sub-Health Status Questionnaire-25 (SHSQ-25) developed by Wei Wang [51]. The questionnaire measures SHS across five dimensions: fatigue, psychological state, cardiovascular system, digestive system, and immune system. SHSQ-25 comprises 25 items, each rated on a 5-point Likert scale (0 = never to 4 = always). The total SHSQ-25 score ranges from 0 to 100, with higher scores indicating more severe SHS. In this study, Cronbach’s alpha coefficient for internal consistency was 0.937, and the KMO value was 0.951. Complete scoring details remain in Supplementary material.

Self-esteem assessment

The Chinese version of the Rosenberg Self-Esteem Scale (RSES-10) [52] was used in this study. The scale consists of 10 items, with items 1, 2, 4, 6, and 7 being positively scored and items 3, 5, 8, 9, and 10 being negatively scored. Each item is scored on a 4-point Likert scale, ranging from 1 to 4. The RSES-10 was used to measure student’s self-esteem, with higher scores indicating higher levels of self-esteem. In this study, Cronbach’s alpha coefficient for internal consistency was 0.918, and the KMO value was 0.917. Complete scoring details remain in Supplementary material.

Anxiety assessment

The Chinese version of the 7-item Generalized Anxiety Disorder scale [53] (GAD-7) was used to assess the anxiety levels of secondary school students(seven items scored 0–3, total 0–21), with higher scores indicating more severe anxiety. In this study, Cronbach’s alpha coefficient for internal consistency was 0.912, and the KMO value was 0.924. Complete scoring details remain in Supplementary material.

Depression assessment

The Chinese version of the Beck Depression Inventory-Ⅱ [54] (BDI-Ⅱ; 21 items scored 0–3, total 0–63) was used to assess depression levels among students, with higher scores indicating more severe depression. In this study, Cronbach’s alpha coefficient for internal consistency was 0.945, and the KMO value was 0.975. Complete scoring details remain in Supplementary material.

Statistical Analysis

To ensure data accuracy, all questionnaire responses were independently entered by two researchers using Epidata 3.1. The two datasets were compared, and any discrepancies were resolved by cross-checking with the original paper forms. After verification and de-identification, the finalized dataset was exported and analyzed using SPSS 23.0. Descriptive statistical analyses were performed on both quantitative and categorical data, using means and standard deviations for continuous variables and frequencies and proportions for categorical variables. For within-group comparisons in both the DBT and control groups, paired sample t-tests were used to assess changes in continuous variables, and McNemar’s test was applied for dichotomous categorical variables. For between-group comparisons, independent sample t-tests were used for continuous variables, while Pearson’s chi-square test was applied to categorical variables. When the assumptions for the chi-square test were violated—specifically, when more than 20% of expected cell frequencies were less than 5, or any individual expected frequency was less than 1—the Fisher–Freeman–Halton exact test was employed as a more appropriate alternative. A two-tailed P value of <0.05 was considered statistically significant.

Results

Characteristics of Participants

The baseline characteristics of the control and intervention groups are shown in Table 2. The occurrence of NSSI in the intervention and control groups was 23.60% and 24.50%, respectively. Most of the students who engaged in NSSI were mildly injured (71.43%). However, differences in gender, only-child status, BMI, household economy, family residence, personality, emotional regulation ability, left-behind status, marital status of parents, family atmosphere, NSSI occurrence, and NSSI severity were not statistically significant (P_all >0.05). Additionally, there were no statistically significant differences in age, NSSI frequency, SHS scores, depression scores, self-esteem scores, and anxiety scores between the two groups (P_all >0.05).

Table 2.

Baseline characteristics of participants

Characteristics	DBT group(n = 110)	Control group(n = 110)	t/χ2	P
Gender			0.018	0.893
Boy	59(53.64)	58(52.73)
Girl	51(46.36)	52(47.27)
Age (x ± s)	13.73 ± 1.21	13.75 ± 0.99	−0.182	0.855
BMI			2.355	0.502
Normal	46(41.80)	53(48.20)
Slim	48(43.60)	47(42.70)
Overweight	7(6.40)	3(2.70)
Obesity	9(8.20)	7(6.40)
Whether an only child			3.124	0.077
Yes	12(10.90)	5(8.50)
No	98(89.10)	105(95.50)
Personality			0.029	0.985
Introvert	23(20.90)	22(20.00)
Neutral	71(64.50)	72(65.5)
Extrovert	16(14.50)	16(14.50)
Emotional regulation ability			3.893	0.143
Good	67(60.90)	55(50.00)
Ordinary	32(29.10)	46(41.80)
Bad	11(10.00)	9(8.20)
Family residence			0.021	0.885
Urban	36(32.70)	35(31.80)
Rural	74(67.30)	75(68.20)
Household economy			2.563	0.464
< 3000	13(11.80)	7(6.40)
3000–5000	21(19.10)	18(16.40)
5000–8000	41(37.30)	47(42.70)
> 8000	35(31.80)	38(34.50)
Whether a left-behind child
Yes	5(4.50)	4(3.60)	0.116	0.734
No	105(95.50)	106(96.40)
Family atmosphere			0.726	0.867
Very good	43(39.10)	42(38.20)
Good	46(41.80)	42(38.20)
Ordinary	17(15.50)	21(19.10)
Bad	4(3.60)	5(4.50)
The marital status of parents			2.405	0.300
First Marriage	102(92.70)	107(97.3)
Divorced	5(4.50)	2(1.80)
Others	3(2.70)	1(0.90)
NSSI occurrence	26(23.60)	27(24.50)	0.025	0.875
NSSI severity			3.948	0.267
No	81(73.64)	83(75.50)
Mildly	18(16.36)	22(20.00)
Moderately	4(3.64)	1(0.90)
Severely	7(6.36)	4(3.60)
NSSI frequency	1.57 ± 4.411	1.01 ± 3.521	1.047	0.296
SHS
Fatigue scores	7.06 ± 5.28	8.06 ± 5.95	−1.318	0.189
Cardiovascular system scores	0.88 ± 1.36	1.15 ± 1.91	−1.177	0.240
Digestive system scores	1.45 ± 1.87	1.61 ± 1.94	−0.602	0.548
Immune system scores	1.65 ± 1.73	1.85 ± 1.69	−0.827	0.409
Psychological state scores	5.37 ± 4.78	6.25 ± 5.48	−1.272	0.205
Total SHS scores	16.43 ± 12.74	18.92 ± 14.78	−1.339	0.182
Depression scores(x ± s)	9.59 ± 11.15	12.29 ± 11.64	−1.176	0.080
Anxiety scores(x ± s)	3.42 ± 4.49	3.94 ± 5.03	−0.792	0.429
Self-esteem scores(x ± s)	29.98 ± 5.70	29.00 ± 6.13	1.219	0.224

Within-Group Comparisons Before and After Intervention

The occurrence of NSSI in the DBT group decreased from 23.6% to 10.91% after the DBT-based interventions, a statistically significant reduction (χ² = 6.23, P < 0.05). In the control group, the occurrence of NSSI decreased from 24.55% to 20.91% after the general interventions, but this change was not statistically significant (χ² = 0.414, P > 0.05). In the DBT group, NSSI frequency decreased significantly from 1.57 ± 4.411 before the intervention to 0.54 ± 1.895 after the intervention (t = 3.268, P < 0.05), whereas anxiety scores and self-esteem scores did not change significantly (P_all>0.05). (Table 3)

Table 3.

Within-Group comparisons of NSSI and related outcomes before and after intervention in DBT and control groups

Variables	DBT group		t/χ2	P	Control group		t/χ2	P
	Before intervention	After intervention			Before intervention	After intervention
Whether NSSI behavior has occurred in the past half-year, n(%)			6.23	0.013*			0.414	0.520
Yes	26(23.60)	12(10.91)			27(24.55)	23(20.91)
No	84(76.40)	98(89.09)			83(75.45)	87(79.09)
NSSI frequency(x ± s)	1.57 ± 4.411	0.54 ± 1.895	3.268	0.001**	1.01 ± 3.521	0.94 ± 3.189	0.929	0.335
SHS
Fatigue scores	7.06 ± 5.28	5.06 ± 3.92	3.186	0.002**	8.06 ± 5.95	7.01 ± 3.54	1.597	0.112
Cardiovascular system scores	0.88 ± 1.36	0.56 ± 1.13	1.889	0.060	1.15 ± 1.91	1.17 ± 1.42	1.420	0.905
Digestive system scores	1.45 ± 1.87	1.07 ± 1.41	1.714	0.088	1.61 ± 1.94	1.65 ± 1.98	−0.176	0.861
Immune system scores	1.65 ± 1.73	1.64 ± 1.49	0.083	0.934	1.85 ± 1.69	1.48 ± 1.43	1.720	0.087
Psychological state scores	5.37 ± 4.78	3.70 ± 4.08	2.792	0.006**	6.25 ± 5.48	6.23 ± 4.72	0.040	0.968
Total SHS scores	16.43 ± 12.74	12.04 ± 7.80	3.083	0.002**	18.92 ± 14.78	17.55 ± 8.21	0.852	0.396
Depression scores	9.59 ± 11.15	6.51 ± 3.08	2.217	0.029*	12.29 ± 11.64	10.06 ± 9.19	1.575	0.117
Anxiety scores	3.42 ± 4.49	3.10 ± 4.08	0.531	0.596	3.94 ± 5.03	3.73 ± 4.97	0.310	0.757
Self-esteem scores	29.98 ± 5.70	30.24 ± 5.82	−0.345	0.735	29.00 ± 6.13	30.28 ± 5.24	−1.838	0.061

* for P < 0.05, ** for P < 0.01, *** for P < 0.001

The DBT group showed significant decreases in fatigue score, psychological state score, and total SHS score after the interventions compared to the period before, with all differences being statistically significant (P_all<0.05). Depression scores in the DBT group decreased from 9.59 ± 11.15 before the intervention to 6.51 ± 3.08 after the intervention, a statistically significant reduction (t = 2.217, P < 0.05). The control group showed no significant differences in NSSI frequency, NSSI occurrence, total SHS score and any of its dimensions, anxiety scores, depression scores, or self-esteem scores before and after the intervention (P_all>0.05). (Table 3)

Between-Group Comparisons After Intervention

The results indicated that while the DBT group (0.54±1.895) had a slightly lower frequency of NSSI compared to the control group (0.94±3.189), this difference was not statistically significant (t=1.131, P>0.05). However, the DBT group exhibited significantly lower fatigue scores, psychological state scores, digestive system scores, total SHS scores, and depression scores compared to the control group (P_all<0.01), indicating significant improvements in these areas. There were no significant differences in anxiety scores, self-esteem scores, or immune system scores between the groups (P_all>0.05). (Table 4)

Table 4.

Between-Group comparisons of NSSI and related outcomes after intervention

Variables	DBT group	Control group	t	P
NSSI frequency	0.54 ± 1.895	0.94 ± 3.189	−1.131	0.259
SHS
Fatigue scores	5.06 ± 3.92	7.01 ± 3.54	−3.970	0.000***
Cardiovascular system scores	0.56 ± 1.13	1.17 ± 1.42	−3.622	0.000***
Digestive system scores	1.07 ± 1.41	1.65 ± 1.98	−2.600	0.01**
Immune system scores	1.64 ± 1.49	1.48 ± 1.43	0.713	0.477
Psychological state scores	3.70 ± 4.08	6.23 ± 4.72	−4.248	0.000***
Total SHS scores	12.04 ± 7.80	17.55 ± 8.21	−5.193	0.000***
Depression scores	6.51 ± 8.223	10.06 ± 9.188	−3.023	0.003**
Anxiety scores	3.10 ± 4.082	3.73 ± 4.082	−1.023	0.307
Self-esteem scores	30.24 ± 4.08	30.28 ± 5.24	−0.049	0.961

* for p < 0.05, ** for p < 0.01, *** for p < 0.001

Comparison of Pre–Post Change Scores Between Groups

When comparing the differences between the DBT and control groups before and after the intervention, the DBT group exhibited a significantly greater reduction in NSSI frequency compared to the control group (t = 2.950, P < 0.01), indicating the effectiveness of the DBT intervention in reducing NSSI occurrences. Although the DBT group showed improvements in total SHS scores, fatigue scores, psychological state scores, and depression scores, these differences were not statistically significant between the two groups. (P_all>0.05). (Table 5)

Table 5.

Between-Group comparisons of Pre–Post change scores in NSSI and related outcomes

Variables	DBT group difference	Control group difference	difference	t	p
NSSI frequency	1.03 ± 3.32	0.07 ± 0.07	0.96	2.950	0.004**
SHS
Fatigue scores	2.03 ± 6.374	1.08 ± 6.453	0.95	1.093	0.275
Cardiovascular system scores	0.33 ± 1.863	−0.03 ± 2.309	0.36	1.254	0.211
Digestive system scores	0.38 ± 2.505	−0.04 ± 2.889	0.42	1.147	0.253
Immune system scores	0.03 ± 2.336	0.37 ± 2.102	−0.34	−1.153	0.250
Psychological state scores	1.67 ± 6.472	0.07 ± 6.783	1.6	1.790	0.075
Total SHS scores	4.44 ± 15.104	1.46 ± 15.694	2.98	1.431	0.154
Depression scores	3.08 ± 14.58	2.23 ± 15.87	0.85	0.420	0.678
Anxiety scores	0.33 ± 6.46	0.21 ± 7.61	0.12	0.120	0.901
Self-esteem scores	0.27 ± 8.28	1.28 ± 7.3	−1.01	−0.958	0.339

* for p < 0.05, ** for p < 0.01, *** for p < 0.001

Discussion

NSSI occurs throughout the lifespan but is most prevalent during adolescence. As internalising psychopathology has become more common in this age group, NSSI has emerged as a critical public-health concern. A study has shown that the incidence of NSSI among adolescents is increasing year by year [55]. The results of this study showed that the prevalence of NSSI among secondary school students was 24.09%, which is similar to the prevalence reported in a meta-analysis [21] of NSSI among secondary school students in China, with a prevalence of 22.37% (95% CI: 18.84–25.70%). Minor differences between these estimates may reflect variation in academic, social, and familial stressors, along with internalising symptoms such as depression and anxiety, all of which are robustly associated with NSSI.

Of the 56 secondary-school students who reported NSSI during the previous 12 months, 40 (71.4%) sustained only minor injuries. This pattern underscores the predominantly covert and low-severity nature of adolescent NSSI and implies that prevalence figures derived from clinical settings substantially underestimate its true community burden. Geulayov [56] proposed the “iceberg” phenomenon of NSSI, emphasizing that the majority of NSSI cases occur in the community, with only a small percentage of individuals seeking help from health facilities such as hospitals. Enhanced vigilance by parents, educators, and public-health stakeholders is therefore essential to prevent and reduce NSSI among adolescents.

In this study, a six-month DBT-based intervention was conducted for the intervention group, while the control group received a six-month standard education intervention. A comparative analysis of NSSI occurrence before and after the intervention was then conducted between two groups. Since there were no statistically significant differences in the characteristics of participants at baseline, it can be more confidently concluded that any observed differences are due to the effects of the intervention, rather than resulting from inherent characteristics of participants. The intervention resulted in a notable reduction in NSSI occurrence within the intervention group, decreasing from 23.6% to 10.91%. This reduction accords with earlier findings; for example, a study of adult women in the United States reported a significant DBT-related decline in NSSI [47] Likewise, a intervention study documented a decrease in NSSI from 66.7% to 22.7% after six months of DBT [57]. By contrast, the control group showed a non-significant fall from 24.6% to 20.9%. The post-intervention reduction in NSSI was therefore significantly greater in the DBT group than in controls, consistent with previous research [31]. Collectively, these findings indicate that a school-based DBT curriculum can markedly reduce NSSI risk in secondary-school students and outperforms standard educational approaches.

SHS has a significant negative impact on adolescents’ physical and psychological well-being, manifesting as persistent fatigue, impaired concentration, weakened immunity, and irritability, among other effects. In evaluating the impact of the intervention on SHS, both the total SHS score and its individual dimensions were examined for the DBT group and the control group. The total SHS score for the DBT group demonstrated a statistically significant decrease post-intervention compared to the pre-intervention levels. In this study, the four modules of DBT’s core skills training group—mindfulness, interpersonal effectiveness, emotion regulation, and distress tolerance—helped students enhance these competencies and effectively improved their SHS. This improvement was particularly evident in the dimensions of fatigue and psychological state. A meta-analysis has demonstrated that improving mindfulness skills in adolescents can effectively enhance their psychological well-being [58]. In addition, another meta-analysis has also shown that mindfulness-based practices can significantly alleviate fatigue in participants [59]. Although direct evidence for DBT-based SHS interventions in secondary-school populations remains scarce, these findings imply that enhanced mindfulness—a core DBT component—may mediate SHS improvement in students. By contrast, the control group exhibited no significant change in total or domain-specific SHS scores. Post-intervention comparisons revealed statistically significant between-group differences in fatigue, cardiovascular, digestive, psychological, and total SHS scores. Collectively, the findings indicate that a DBT programme—integrating behavioral science and dialectical principles and tailored to secondary-school students—exerts a significant beneficial effect on SHS.

Depression, anxiety and self-esteem were reassessed in relation to internalizing problems in secondary school students after the intervention. The DBT group exhibited a statistically significant reduction in depression levels, and the difference in depression between two groups after the intervention was also statistically significant. The effect of the intervention is consistent with a study on American college students [57], in which depression scores decreased greatly from 34.74 ± 8.67 to 18.90 ± 9.19 after six months of DBT intervention. However, anxiety levels and self-esteem did not change significantly from pre-intervention levels. A study involving individuals with suicidal behavior and borderline personality disorder reported that a comprehensive DBT programme—combining individual therapy with skills training groups—significantly alleviated both depression and anxiety [60]. The absence of individual therapy in this study may explain why anxiety scores did not significantly decrease. Regarding self-esteem, previous studies have shown that DBT interventions significantly improve self-esteem in individuals with BPD, a finding inconsistent with our results [61]. One plausible explanation is that those studies employed a more comprehensive format—including individual therapy, skills training, and consultation-team meetings—that may be more potent for enhancing self-esteem. Moreover, the observation of changes only in depression levels in this study, compared with previous research, may be attributed to differences in the duration of DBT intervention, follow-up periods, and intervention settings [60, 61]. Therefore, future DBT interventions for adolescents should carefully consider strategy formulation to more effectively address internalizing problems and self-esteem improvement.

In future interventions, collaboration between educators and parents will be essential to support young people, raise awareness, establish support systems, promote mental health education, discourage excessive criticism and blame, foster a positive social environment, and encourage seeking professional help when necessary. Additionally, secondary school students should be equipped with skills to manage emotions, cope with stress, enhance self-esteem, and avoid negative emotion.

Strengths and limitations

This study, based on DBT, targeted mindfulness skills, emotion regulation, interpersonal relationships, and coping strategies to intervene in NSSI. These findings offer new perspectives for reducing the occurrence of NSSI and developing intervention measures for secondary school students in Shantou, China. This study also provides empirical evidence supporting the effectiveness of DBT in addressing NSSI within school settings. Its core mechanism—the skills training group—demonstrated stability across different linguistic and cultural backgrounds. Consequently, the results carry important practical implications for adolescent mental-health interventions in Chinese regions where Mandarin is not the competent dialect.

However, this study has some limitations. To begin with, the data were primarily collected through self-report questionnaires, which may be subject to recall bias, as students were required to recall their NSSI behaviors over the past year. Furthermore, there is a possibility of reporting bias, as students might exaggerate or conceal information to provide socially desirable responses. Although we detected significant improvements in several outcome variables, the analytic approach employed in this study does not directly model the interaction between group and time. Future studies with larger samples may benefit from the use of mixed-design ANOVA to examine such interactions and provide more comprehensive evidence regarding the intervention’s efficacy. In addition, China’s vast regional cultural differences may affect the generalizability of our findings. Since this study was conducted only in Shantou city, the results may not be generalized to all secondary school students in China. Although we recognised the distinctive linguistic and cultural context of the Chaoshan area, the intervention itself was not culturally or linguistically adapted to local characteristics. The DBT session materials and examples were developed and delivered in Mandarin, without specific reference to Chaoshan dialect or cultural idioms. Group facilitators conducted sessions in Mandarin Chinese, which all participants could understand, although some spoke Chaoshan dialect in daily life. Facilitator training emphasized general DBT skills, but did not directly address dialect-related communication barriers. This lack of cultural and linguistic tailoring could have attenuated intervention efficacy. Future work should weave culturally and linguistically sensitive elements into the DBT framework to enhance outcomes across diverse settings. Most importantly, whereas previous research shows that DBT effects on adolescent self-injury can persist for up to one year [60], our study assessed only short-term outcomes. Therefore, the durability of the intervention’s effects in this study remains uncertain. Considering prior evidence of DBT’s long-term impact, it is plausible that some benefits may be sustained over time. Nevertheless, without longitudinal data, such assumptions remain speculative. Future studies incorporating long term follow-up assessments are essential to confirm the persistence of DBT’s effects in this population.

Conclusion

The six-month DBT-based intervention demonstrate a significant impact on reducing the occurrence of NSSI among secondary school students, surpassing the effectiveness of conventional interventions. This underscores the effectiveness of DBT-based NSSI interventions in reducing incidence and improving SHS. Preventing NSSI requires concerted efforts across society to enhance public awareness, establish support systems, increase awareness among parents and educators, foster positive lifestyles, and provide comprehensive mental health education. Building on this study’s outcomes, more comprehensive and effective interventions should be developed to cultivate partnerships among families, educational institutions, and society, with the goal of preventing NSSI and promoting healthy adolescent development.

Abbreviations

NSSI

Non-suicidal self-injury

SHS

Sub-optimal health status

BMI

Body mass index

DBT

Dialectical Behavior Therapy

Authors’ contributions

L.L., P.W. and J.L. designed the study. Z.Z. collected the data. J.L. and P.W. were involved in the analysis of data. J.L. and P.W. drafted the paper, and J.L., P.W., J.Z. and L.L. iterated and commented on drafts. P.W., J.L., J.Z. and L.L. reviewed the manuscript. All authors approved the submitted manuscript.

Funding

No funding was received to assist with the preparation of this manuscript.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This randomized controlled trial strictly adheres to the Consolidated Standards of Reporting Trials (CONSORT) 2010 guidelines. The study obtained approval from the Ethics Committee of Shantou University (SUMC-2022-076) and received authorization from the leaders and class teachers of the surveyed educational institutions. And the informed consent to participate was obtained from the parents or legal guardians of all participants under the age of 16.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.