The effects of an empowering self-management model on sense of coherence and self-efficacy among adolescent girls with internet addiction in Bushehr, Iran: a randomized controlled trial

Abstract

Background

Despite the positive applications of the internet, excessive use and its associated harmful consequences, particularly internet addiction among adolescents, have attracted significant attention. The present study aimed to assess the effect of an empowering self-management model on the sense of coherence and self-efficacy among adolescent girls with internet addiction in Bushehr (Iran),2022.

Methods

This clinical trial was conducted on 80 adolescent girls with internet addiction in the first grade of secondary school in Bushehr. In a two-stage cluster sampling design, schools were randomly selected and assigned to either the intervention or control group. Participants voluntarily completed Young’s(1998) internet addiction test and were included in the study if they scored 46 or higher. Based on an empowering self-management model, the intervention group participated in a five-stage plan: (1) self-awareness of changes and understanding their personal level of performance and expectations; (2) optimal goal setting; (3) planning; (4) adjusting physical, psychological, and social structures; and (5) evaluation. Data collection instruments comprised the Antonovsky’s sense of coherence Questionnaire (SOC-13) and the Sherer’s general self-efficacy scale (GSES), administered at baseline and post-intervention to assess changes in sense of coherence and self-efficacy. Data were analyzed using SPSS 26.0 at the significance level of 0.05.

Results

The two groups showed no significant differences in most demographic variables, except for age and having a personal room (p < 0.05). After adjusting for age and access to a personal room, a repeated-measures analysis of covariance(ANCOVA) revealed that the intervention group experienced a significantly greater improvement in sense of coherence than the control group (adjusted mean change:25.57 ± 1.1 vs. -2.67 ± 1.1; p < 0.001; partial eta squared [η²] = 0.799). Similarly, self-efficacy increased significantly more in the intervention group than in the control group (adjusted mean change:17.05 ± 0.76 vs. -2.77 ± 0.76; p < 0.001; η² = 0.807). After the intervention, severe internet addiction in the intervention group dropped to zero, with 10% of participants classified as mild, while no change was observed in the control group (p = 0.03). Accordingly, the mean change in the internet addiction score after the intervention was − 12.85 ± 0.79 in the intervention group and + 1.70 ± 0.28 in the control group (p < 0.001).

Conclusion

The self-management empowerment model significantly improved the sense of coherence and self-efficacy among adolescent girls with internet addiction. However, given the several potential biases, short follow-up period and the homogeneity of participants, these findings should be interpreted with caution. Further studies involving male adolescents and more diverse samples are warranted to confirm the generalizability of the results.

Trial registration

IRCT20130422013092N11 registered date: Sep 12, 2022.

Background

Adolescence is a complex transitional phase from childhood dependency to adulthood, marked by developmental potential, personal agency, and social responsibility [1]. Puberty introduces physiological changes that contribute to stress and psychological challenges [2]. According to Erikson’s theory, identity development, a key psychosocial task, is shaped through social interactions [3]. Adolescents’ strong need for peer acceptance drives them to seek social approval, often through virtual identities fostered by early exposure to computers and the internet [4, 5]. The COVID-19 lockdown notably increased adolescents’ internet use, contributing to the emergence of internet addiction as a contemporary behavioral concern [6, 7].

The term “internet addiction,” first coined by Yang (1998) [8], refers to excessive and pathological internet use [9]. The prevalence of internet addiction has been rising, particularly among adolescents, who represent a significant portion of affected individuals [10, 11]. The American Academy of Pediatrics warns against excessive digital use, as it may diminish imagination, creativity, and physical development opportunities [12]. Internet addiction has thus become a critical social and health issue globally, attracting multidisciplinary research attention [13]. In the U.S., about 25% of adolescents experience this addiction [14], while Iranian data suggest a growing trend among youth, especially girls [15]. Boys typically prefer online gaming, whereas girls spend more time on social media [16]. Psychological and behavioral consequences include mood and anxiety disorders, depression, stress, sleep problems, ADHD, and social withdrawal [16–18], with girls more susceptible to depression and anxiety due to unhealthy internet habits [19, 20]. Given that adolescence, especially among girls, is marked by heightened stress and psychological vulnerability, it is essential to explore protective factors such as the sense of coherence, which facilitates effective stress adaptation [21].

The concept of “sense of coherence” (SOC), introduced by Antonovsky (1993), provides a framework for understanding how individuals cope with stress and maintain well-being [22]. SOC consists of three components: comprehensibility (cognitive understanding of life’s events), manageability (belief in available resources to handle demands), and meaningfulness (motivation to engage with challenges) [23]. Developing throughout life, SOC plays a crucial role during adolescence in stress adaptation [24]. A study in Norway found boys have higher SOC than girls, who reported more frustration and psychosomatic distress [25]. Social exclusion (ostracism) in adolescents increases internet addiction, which elevates stress and reduces SOC [26]. Moreover, higher SOC scores correlate with less internet use under stress [21].

Self-efficacy is also a vital internal resource that empowers individuals to control their behaviors, mediating the effects of stress and maladaptive coping on internet addiction [27]. It reflects an individual’s perceived ability to manage challenges and maintain resilience during stressful situations, preventing acceptance of negative self-beliefs [28]. As a personality trait, self-efficacy influences vulnerability to various problems, including internet addiction [29]. Empirical evidence shows that self-efficacy predicts internet addiction risk, and enhancing it alongside social skills can effectively mitigate addiction in students [13, 30]. Among the theoretical frameworks addressing internet addiction, Davis’ (2001) Pathological internet use model and Caplan’s (2003) theory of problematic internet use are prominent, both grounded in a cognitive-behavioral approach. Davis’ model posits that pre-existing psychosocial difficulties, such as depression and social anxiety, lead to maladaptive cognitions and behaviors, resulting in excessive internet use. Caplan emphasizes individuals’ preference for online over face-to-face social interactions and links social skill deficits to addiction risk. Therefore, studies incorporating these models alongside positive psychology concepts such as sense of coherence and self-efficacy are essential for a comprehensive understanding of internet addiction [31, 32]. A systematic review and meta-analysis found that Internet addiction in teenagers and college students is strongly associated with interpersonal difficulties, with higher addiction levels linked to poorer quality relationships [33]. Although school-based prevention programs show promise, evidence remains limited, emphasizing the need for skill-building and protective factors. School counsellors highlight that comprehensive interventions, including family support and digital literacy education, are essential, and strategies should be multifaceted rather than solely focused on restricting Internet access [34, 35]. Collectively, these findings underscore the urgent need for well-designed interventions to mitigate the social and cognitive harms of adolescent Internet addiction.

Empowering adolescents, particularly regarding internet use, is a vital strategy for preventing and managing internet addiction by strengthening cognitive beliefs and behavioral control [36]. Originating from the 1970 s patient rights movement, empowerment is both a process and outcome focused on enhancing individuals’ strengths, competence, and self-efficacy [37]. As a nursing intervention, empowerment builds on realistic self-expectations [38], while self-management enables individuals to utilize personal and environmental resources to achieve health goals [39]. The self-management empowerment model, based on grounded theory, encompasses awareness, independence, role function, adaptability, satisfaction, control, and self-management, key elements of personal power [40]. Previous studies confirm the model’s effectiveness among adolescents with chronic conditions [41, 42]. Given the heightened vulnerability of adolescents, especially girls, to internet addiction amid puberty and the COVID-19 pandemic, interventions targeting mental and physical health are crucial. Thus, the present study evaluated an empowerment-based self-management education’s impact on sense of coherence and self-efficacy in adolescent girls with internet addiction in Bushehr, 2022.

Methods

Study design

This study was a randomized controlled clinical trial (RCT) with a parallel group design, with an allocation ratio of 1:1 (equal groups), and a superiority framework, aiming to show that education based on the self-management empowerment model improves outcomes related to sense of coherence, and self-efficacy among adolescent girls with internet addiction, compared to usual education. The study protocol was registered in the Iranian Clinical Trial Registration System the IRCT registration code: (http://www.irct.ir/, IRCTID: IRCT20130422013092N11 registered Sep 12, 2022), and approved by the Ethics Committee of Bushehr University of Medical Sciences (IR.BPUMS.REC.1401.078). The CONSORT (Consolidated Standards of Reporting Trials) guidelines were used to design, conduct, and report this study.

Study population and setting

The research population included adolescent girls studying in the junior high schools (grades 7–9) in Bushehr, 2022. This study employed a two-stage cluster sampling procedure. In the first stage, four schools were randomly selected from eligible junior high schools (grades 7–9) in the Bushehr Department of Education. Then, two schools (one public and one private) were randomly assigned to the intervention group, and two other schools (one public and one private) were assigned to the control group. In the second stage, all eligible adolescent girls within these schools who met inclusion criteria were invited to participate. Cluster randomization at the school level was chosen to minimize the risk of contamination between participants in different study arms, as students within the same school could easily share information about the intervention. Because allocation was performed at the school (cluster) level, allocation concealment at the individual level was not possible. The researcher who delivered the intervention also carried out the randomization procedure.

Sample size

The initial sample size was calculated based on the study by Navabi et al. [42], which reported post-intervention changes in self-efficacy of 20.52 ± 10.44 and 8.61 ± 8.60 in the intervention and control groups, respectively. However, using these exact values yielded a small sample (about 10 participants per group). Using the software G*Power 3.1.9.2, with a confidence level of 95% and a power of 80%, we obtained a more conservative and reliable estimate while accounting for the cluster-randomized design (schools as clusters). We assumed more moderate group differences (15 ± 8 vs. 10 ± 5) and recalculated the sample size accordingly. This resulted in approximately 30 participants per group. After adjusting for a 20% anticipated attrition rate, the final target sample size was set at 40 participants per group (80 in total). The design effect was expected to be small given the limited number of clusters (four schools) and relatively low intra-class correlation anticipated for psychosocial outcomes. However, the effective sample size may still be underpowered for a cluster design; therefore, LMM was applied to consider the cluster effect.

The female adolescents were asked to complete the Young’s internet addiction test (IAT), and those who scored higher than 46 were included in the study. The participants were selected from junior high schools (grades 7–9) using convenience sampling method until 20 participants were recruited from each school. The sampling period lasted from September to December of 2022. No specific blinding procedures were implemented in the present study. Nonetheless, as the primary outcomes were assessed using standardized questionnaires administered uniformly across both groups, the potential for bias is minimized. However, the possibility of detection bias, performance bias or residual bias cannot be completely excluded.

Consent and enrollment

The participants were informed that their participation in the research would be voluntary and their information would remain confidential. Then, written consent was obtained from the parents, and verbal informed consent was obtained from the adolescents. The inclusion criteria were scoring 46 or higher on the internet addiction scale, possessing a device for accessing the internet such as a smartphone and obtaining consent from both the adolescent and their parents to participate in the study. The exclusion criteria were not fully completing the questionnaire by the adolescent, using psychiatric medications, participating in other similar studies simultaneously and withdrawing from the study.

Self-management empowerment intervention

Participants in both groups completed demographic form, Antonovsky’s sense of coherence (SOC-13) questionnaire and Sherer’s general self-efficacy scale (GSES). The self-management empowerment intervention was implemented only in the intervention group over a two-month period. The training program was rigorously reviewed and approved by a panel of subject matter experts, comprising specialists in nursing, pediatrics, community medicine and health education. Participants in the intervention group underwent a structured five-step empowerment program, grounded in a self-management framework and customized to address needs identified during the preliminary assessment sessions. If a participant failed to implement 40% or more of the specified components of the intervention, they were excluded from the study. The intervention was administered to the participants by the first author. No significant modifications were made to the trial after its initiation. The five-step process of the intervention were:1) Self-awareness and recognition of performance level and self-expectations; 2) Desired goal setting; 3) Planning; 4) Adjusting physical, psychological and social structures; and 5) Evaluation [20, 22].

Step 1. self-awareness and recognition of performance level and self-expectations

Through individual interviews between the researcher and each adolescent, factors such as lifestyle, living environment, adolescent’s performance in managing life’s challenges, their level of independence, coping mechanisms used in response to problems and barriers, effectiveness of support systems and their expectations from themselves, their family and others were assessed based on the designed algorithm. Furthermore, the adolescent’s awareness of changes occurred in their physical, academic and social status due to internet addiction was evaluated. Subsequently, the adolescent’s level of expectations was assessed based on predefined criteria to determine whether these expectations were appropriate, inappropriate, excessive, reasonable or insufficient.

Step 2. desired goal setting

In this step, specific goals were set and prioritized in collaboration with the adolescents, based on the needs assessment and results from the previous stage, as well as their self-expectations, in order to minimize the gap between their current performance and expected levels.

Step 3. planning

In this step, a care plan was developed using the established goals to address the adolescents’ problems based on self-management empowerment model’s domains.

It is worth noting that the first three steps of the intervention (needs assessment, goal setting, and planning) were conducted in two in-person sessions, each lasting approximately 30 min.

Step 4. adjusting physical, psychological and social structures

The empowerment intervention was implemented based on the results of assessments, needs analyses and developed plan using self-management empowerment model’s domains (e.g., attending classes to address educational difficulties, learning new sports and artistic skills to make personal goals more meaningful, reconnecting with family and friends, maintaining physical activity and weight control, recording daily internet usage and striving to reduce it, time management). To make structural adjustments and implement the proposed strategies, the adolescents were given 6 weeks, with first 2 weeks dedicated to self-awareness assessment, goal setting and planning, resulting in a total of 8 weeks for the intervention. During this period, the researcher contacted the adolescents once a week via phone call to assess and record the progress of the intervention. The ambiguities or issues, if present, were addressed. To facilitate the implementation of the intervention, an educational booklet containing key empowerment points based on the concepts of sense of coherence and self-efficacy was provided to the adolescents in the intervention group.

Step 5. evaluation

The evaluation process was conducted at all steps of the intervention. Evaluation was carried out during the empowerment sessions and throughout the follow-up period. Inquiries were made to set goals and proposed intervention plans. All follow-ups were aligned with the agreed-upon objectives. To ensure standardization of the intervention, an algorithm grounded in an empowerment model, focusing on self-efficacy and sense of coherence, was developed with contributions from experts in nursing and public health. This algorithm outlined the core objectives and the minimum essential actions required to ensure intervention fidelity. Based on this, a checklist was developed, and participants who completed fewer than 40% of the prescribed actions were excluded from the study. Finally, after eight weeks of follow-up with the intervention group, the participants in both groups completed all the three questionnaires once again. During the course of the intervention, no adverse events occurred that would hinder continuation of the study, and participants demonstrated strong engagement and enthusiasm during the face-to-face sessions and subsequent follow-up activities. A limited number of follow-ups on general, non-intervention-related topics were held for the control group to minimize the Hawthorne effect, and participants completed the questionnaires again two months later.

Data collection

The data collection tools included a demographic form, Young’s internet addiction test (IAT), Antonovsky’s sense of coherence (SOC-13) questionnaire and Sherer’s general self-efficacy scale (GSES), which were distributed and collected by the researcher. The questionnaires were completed by the participants of both groups at the beginning of the intervention (before the interview), at the end of the intervention in the intervention group and two months after the pre-test in the control group.

Demographic information included items such as age, educational level, parents’ education and occupation, birth order, weekly pocket money, average hours of digital device use per week, average number of nights per week preferring digital device use over sleep, and other related factors.

Internet addiction was assessed using the Young’s IAT, which contains 20 items. Each item is scored based on a 6-point Likert scale, including never, rarely, sometimes, usually, often and always, with scores ranging from 0 to 5. The total score ranges from 0 to 100, with a score of 0–20 indicating a typical user, 21–49 indicating mild internet addiction, 50–79 representing moderate internet addiction and 80–100 indicating severe internet addiction. The best clinical cutoff point for this questionnaire was 46 [15]. The Young’s IAT showed good psychometric properties in the Iranian population and could be used in psychological and psychiatric research to screen typical internet users from addicts (Cronbach’s alpha = 0.95) [14].

Antonovsky’s SOC-13 questionnaire, which consists of 13 items, was used to assess sense of coherence and its dimensions: Comprehensibility (5 items), meaningfulness (4 items) and manageability (4 items). Each item includes seven options, ranging from 1 to 7. The questionnaire demonstrated good validity and reliability (Cronbach’s alpha = 0.96) [43, 44]. This questionnaire has been extensively validated and used cross-culturally among both adults and adolescents [23].

Self-efficacy was assessed using Sherer’s general self-efficacy scale (GSES), which includes 17 items rated on a 5-point Likert scale ranging from strongly disagree to strongly agree (from 1 to 5). The questionnaire demonstrated good validity and reliability (Cronbach’s alpha = 0.86) [45].

The study’s primary outcomes included both sense of coherence and self-efficacy, while the secondary outcome focused on the reduction of internet addiction. The reliability of the measures in our study was examined using Cronbach’s alpha. The Sense of Coherence scale and the General Self-Efficacy Scale both showed good internal consistency at pretest (α = 0.69 and 0.78, respectively) and remained reliable at post-test (α = 0.86 and 0.89, respectively).

Statistical analysis

The collected data from the questionnaires were coded, entered into the computer and, then, analyzed using SPSS 26.0. In data analysis, in addition to presenting descriptive statistics (mean, standard deviation, frequency and percentage of frequency), Chi-square test was used to compare the distribution of qualitative demographic variables between the two groups before the intervention. Changes in Sense of Coherence (SOC) and General Self-Efficacy Scale (GSES) scores were initially examined using repeated-measures analysis of covariance (ANCOVA), adjusting for age and having a personal room, with measurements at pretest and posttest. Before performing ANCOVA, the assumptions of normality, homogeneity of variances, and homogeneity of regression slopes were checked. Normality of residuals was generally satisfied, except for the post-test sense of coherence score in the intervention group (Shapiro-Wilk p < 0.05). Levene’s test indicated homogeneity of variance for sense of coherence, but not for self-efficacy. Homogeneity of regression slopes was satisfied for both outcomes at pre- and post-test. Since Linear Mixed Models (LMM) were used to account for the clustered structure of the data, these minor violations of ANCOVA assumptions were appropriately addressed. To account for the hierarchical structure of the data, participants nested within schools and repeated measurements per participant, linear mixed-effects models (LMMs) were subsequently applied. For each LMM, fixed effects included group (intervention vs. control), time (pretest vs. posttest), having a personal room, centered age, and the group × time interaction, with participant-level random intercepts. Age was centered around the sample mean in the LMM to improve interpretability of the intercept (representing the estimated outcome for a participant of average age) and to reduce multi-collinearity with other fixed effects. School-level random effects were initially considered but were not estimable due to the small number of clusters, so the final model included only participant-level random intercepts. Estimated marginal means (EMMeans) for each group at each time point were calculated, adjusted for covariates. Restricted maximum likelihood (REML) estimation and Type III sums of squares were used to test fixed effects. The EMMeans were used to compute adjusted mean changes, adjusted mean differences with 95% confidence intervals, and partial eta squared effect sizes. In all analyses, the significance level of 0.05 was considered. Some results from this data set have been reported in Persian Language before [46]; and we focus on other outcomes including sense of coherence and self-efficacy in this paper.

Ethical considerations

To ensure compliance with ethical research principles and gain access to the study setting, ethical approval was obtained from the Research Council and the Ethics Committee of the Vice-Chancellor of Research at Bushehr University of Medical Sciences. Subsequently, permission was secured from the Bushehr Education Department and the relevant school administrators. Participants were informed that their involvement in the study was entirely voluntary and that all information provided would be kept strictly confidential. It was clearly emphasized that participants could withdraw from the study at any time without any academic or other consequences. Finally, written informed consent was obtained from the parents, and oral informed consent was obtained from the adolescent participants.

Results

Among the 100 adolescent girls who were invited to participate in this study, 90 met the inclusion criteria. The participants were assigned to the control group (n = 44) and the intervention group (n = 46). During the study, four adolescent girls in the control group and six in the intervention group withdrew due to loss to follow-up or unwillingness to continue. Consequently, the analysis was conducted on 40 adolescent girls in each group (see Fig. 1).

Fig. 1.

CONSORT flow diagram displaying the progress of all participants through the trial

The mean age of the participants in intervention and control groups was 13.52 ± 0.90 and 13.02 ± 0.83, respectively. There were no statistically significant differences between the two groups in terms of mean monthly pocket money (p = 0.38), daily time spent on the Internet (p = 0.66), or the number of nights spent awake using the Internet (p = 0.31). Other demographic characteristics of the participants in both groups are detailed in Table 1.

Table 1.

Comparison of demographic characteristics of the intervention and control groups of adolescent girls with internet addiction, Bushehr 2022

Variable		Groups				p-value *
		Intervention		Control
		Frequency	%	Frequency	%
Birth rank	First	19	47.5	19	47.5	0.17
	Second	15	37.5	11	40
	Third	2	5	5	12.5
	Forth	4	10	0	0
Living with	Mother	3	7.5	1	2.5	0.49
	Father	1	2.5	3	7.5
	Both	36	90	35	87.5
	With other	0	0	1	2.5
Father’s education	Below diploma	13	32.5	18	45	0.29
	Diploma	5	12.5	7	17.5
	University	22	55	15	37.5
Mother’s education	Below diploma	15	37.5	23	57.5	0.18
	Diploma	12	30	7	17.5
	University	13	32.5	10	25
Repeating grade	Yes	0	0	2	5	0.49
	No	40	100	38	95
Grade	7	19	47.5	12	30	0.14
	8	14	35	14	35
	9	7	17.5	14	35
Level of parental control (From the adolescents’ perspective)	Low	11	27.5	15	37.5	0.40
	Medium	24	60	18	45
	High	5	12.5	7	17.5
Mother’s job	Freelance	3	7.5	6	15	0.08
	Employee	3	7.5	9	22.5
	Housewife	34	85	25	62.5
Father’s job	Freelance	19	47.5	19	47.5	1.000
	Employee	21	52.5	21	52.5

*A significance level of less than 0.05 is considered

The results of the chi-square and independent t-tests for qualitative and quantitative variables indicated that the two groups did not differ significantly in most demographic variables (p > 0.05); however, there were significant differences between the groups in terms of age (p = 0.012) and having a personal room at home (p = 0.033). Furthermore, nearly 90% of the participants in both groups had a moderate level of internet addiction, while the remaining participants had a severe level of addiction. At baseline, the two groups did not differ in the level of internet addiction [46].

The results of a repeated measure ANCOVA using a general linear model (GLM), adjusting for age and having a personal room, indicated that there was a significant time × group interaction and the mean changes in sense of coherence over time remained significantly different between the two groups (p < 0.001, partial ƞ²=0.79). The mean adjusted change in the sense of coherence score after the intervention was + 25.57 ± 1.1 and − 2.67 ± 1.1 in the intervention and control groups, respectively. Among the covariates, having a personal room was a significant predictor (p = 0.02), but age was not (p = 0.11) (Table 2).

Table 2.

Repeated measures ANCOVA for sense of coherence and self-efficacy (adjusted for age and having a personal room) in adolescent girls with internet addiction, Bushehr 2022

Outcome	Group	Before* Adjusted mean ± SE	After* Adjusted mean ± SE	Adjusted mean change (After-Before) ± SE	Adjusted mean difference (95% CI)	Time × Group
						F	p-value	Partial ƞ2 (95% CI)
Sense of coherence	Control	40.85 ± 1.63	38.18 ± 1.61	−2.67 ± 1.1	28.24 (23.77, 32.71)	302.50	< 0.001	0.799 (0.723, 0.852)
	Intervention	34.17 ± 1.63	59.74 ± 1.61	+ 25.57 ± 1.1
Self-efficacy	Control	52.83 ± 1.48	50.06 ± 1.39	−2.77 ± 0.76	19.82 (18.31, 21.33)	317.13	< 0.001	0.807 (0.744, 0.851)
	Intervention	50.21 ± 1.48	67.26 ± 1.39	+ 17.05 ± 0.76

*Values are adjusted means ± standard errors

Repeated measure ANCOVA using a general linear model (GLM), adjusting for age and having a personal room, demonstrated that there was a significant time × group interaction and the mean changes in self-efficacy over time remained significantly different between the two groups (p < 0.001, partial ƞ²=0.80). The mean adjusted change in the self-efficacy score after the intervention was + 17.05 ± 0.76 in the intervention group and − 2.77 ± 0.76 in the control group. None of the covariates, age (p = 0.82) and having a personal room (p = 0.21), was a significant predictor (Table 2).

As reported previously [46], the chi-square test showed no significant difference in the severity of internet addiction between the intervention and control groups at baseline (p = 1.000), whereas a significant difference emerged after the intervention (p = 0.03). Following the program, the proportion of participants with severe addiction in the intervention group dropped to zero, with 10% shifting to the mild category, while no change was observed in the control group. Accordingly, the mean change in the internet addiction score after the intervention was − 12.85 ± 0.79 in the intervention group and + 1.70 ± 0.28 in the control group (p < 0.001).

The effect size estimates indicated that the intervention had a very strong influence on both outcomes. For sense of coherence, the interaction between group and time explained approximately 79% of the variance in change (η²p = .79), and for self-efficacy, the corresponding interaction accounted for about 81% of the variance (η²p = .81). The values obtained for both variables exceed the conventional threshold for a large effect (η²p ≥ .14), thereby indicating that the intervention produced substantial improvements in these psychological measures. In comparison, the effect of having a personal room on sense of coherence was of medium size (η²p = .07), while age showed no significant or meaningful contribution. (Table 2).

The LMM confirmed ANCOVA results. The intervention significantly improved both Sense of Coherence (SOC) and self-efficacy in adolescent girls with internet addiction. Linear mixed-effects models, adjusted for age and having a personal room, revealed significant group × time interactions for both outcomes (SOC: F = 335.48, p < 0.001, partial η² = 0.811; self-efficacy: F = 369.27, p < 0.001, partial η² = 0.821). The control group showed minimal or negative changes over time, whereas the intervention group showed substantial improvements. Adjusted mean changes, adjusted mean differences, and 95% confidence intervals are presented in Table 3.

Table 3.

Linear mixed model for sense of coherence and self-efficacy (adjusted for age, having a personal room and participant ID) in adolescent girls with internet addiction, Bushehr 2022

Outcome	Group	Before* mean ± SE	After* mean ± SE	Adjusted mean change (After-Before) ± SE	Adjusted mean difference (95% CI)	Time × Group
						F	p-value	Partial ƞ2 (95% CI)
Sense of coherence	Control	39.89 ± 1.59	37.39 ± 1.59	−2.50 ± 1.07	27.90 (24.87, 30.93)	335.48	< 0.001	0.811 (0.738, 0.859)
	Intervention	33.38 ± 1.71	58.78 ± 1.71	+ 25.40 ± 1.07
Self-efficacy	Control	52.38 ± 1.41	49.68 ± 1.41	−2.70 ± 0.72	19.67 (17.64, 21.72)	369.27	< 0.001	0.821 (0.752, 0.865)
	Intervention	49.83 ± 1.52	66.80 ± 1.52	+ 16.97 ± 0.72

*Values are estimated marginal means ± SE (95% CI) derived from linear mixed models with participant ID as a random intercept. Fixed effects included time (pretest/posttest), group (intervention/control), group×time interaction, age, and having a personal room. Adjusted mean differences between groups were computed from the model estimates

Discussion

This study evaluated the impact of a self-management empowerment-based educational intervention on sense of coherence and self-efficacy in adolescent girls with internet addiction in Bushehr, Iran. Most participants reported concerns regarding excessive internet use and its adverse impacts on their education, social life, sleep, and mental health. The results showed that implementing the intervention significantly increased the sense of coherence and self-efficacy of the intervention group. There was no statistically significant difference in most demographic information between the two groups, except for age and access to a personal room. After adjusting for these factors using ANCOVA, it was revealed that the intervention group experienced a significantly greater improvement in sense of coherence and self-efficacy than the control group.

Although the effect sizes observed for both variables were large, these findings should be interpreted with caution. Such magnitudes may partly reflect contextual factors, such as the relative homogeneity of participants, the limited follow-up period, or the motivational impact of educational sessions. Therefore, further replication in more heterogeneous and longer-term studies is warranted to confirm the stability and real-world applicability of these results.

In line with our findings, Navabi et al. indicated that implementing an intervention based on the self-management empowerment model led to positive changes in the sense of coherence and self-efficacy among adolescents with motor vehicle injuries in Darab [42]. Similarly, Ebrahimi and Parand conducted a study on 385 junior and senior high school students in Tehran and showed that increasing the sense of coherence reduced internet addiction among adolescents [21]. Skonieczna et al. [47] and Mortezaei and Rahiminezhad [22] demonstrated that sense of coherence inversely predicts internet addiction. In other words, a weak sense of coherence plays a significant role in predicting internet addiction. Regarding the similarity of these findings with the present study, it could be stated that sense of coherence helped adolescents engage with and understand stressful situations, rather than avoiding them and resorting to tools like the internet. By managing these challenges, they could give meaning to their lives. In sum, comparable international studies have also highlighted the benefits of empowerment-oriented or cognitive-behavioral-based programs on adolescents’ control over internet use. For instance, a systematic review of ten randomized controlled trials found that Cognitive Behavioral Therapy (CBT) and Electro-Acupuncture (EA) effectively reduce internet addiction symptoms in youth, with school-based and brief CBT interventions showing additional promise [48]. Persistent conceptual and diagnostic inconsistencies highlight the need for standardized, multidisciplinary approaches to advance evidence-based treatment. However, while these trials often included both genders and various cultural contexts, the current study focused on a specific group of Iranian adolescent girls, suggesting that the self-management empowerment model may be adaptable across settings but that several factors require consideration when generalizing results.

The present study showed after the two-month self-management empowerment intervention, self-efficacy scores and their mean changes significantly increased in the intervention group, even after adjusting for age and having a personal room, and decreased in the control group. These findings were consistent with the study by Hourzad et al., who demonstrated that the self-management empowerment model improved self-efficacy and sense of coherence among retired elderly individuals with chronic diseases [49]. Results of the systematic review by Dehkordi showed the empowerment and self-efficacy model could improve the quality of life in patients with chronic diseases. Therefore, in addition to the benefits of empowerment for patients with chronic diseases in previous studies, it has been recommended to be implemented for patients with other diseases as well [50]. Additionally, Keshavarz et al. [51] and Bozorgkhoo et al. [30] demonstrated that individuals with internet addiction had lower self-efficacy compared to regular users. Chen et al. conducted a cross-sectional study on 451 fifth- and sixth-grade adolescents in Taiwan and found that low self-efficacy was a contributing factor to internet addiction among adolescents [52]. In relation to the similarity between these findings and the present study, it can be stated that, self-regulation is an outcome of self-efficacy, implying that individuals with higher self-efficacy exert greater control over their thoughts, emotions, motivation, and behavior. Consequently, they tend to demonstrate greater perseverance, achieve more success, and maintain better health. The substantial effect sizes observed following the intervention may also be attributable to the intensity or adaptation of the intervention protocol, or to specific contextual factors characteristic of the present sample.

Moreover, international evidence concerning empowerment-based and cognitive-behavioral interventions aimed at reducing adolescent internet addiction through self-efficacy enhancement remains inconsistent. A systematic review by Morton et al. (2013) reported no significant effects of empowerment programs on self-efficacy (N = 167), citing insufficient evidence [53]. Similarly, Romero Saletti et al. (2021) found that although several interventions showed individual benefits, their overall effectiveness disappeared in meta-analysis [54]. Malinauskas et al. (2019) observed potential positive trends but highlighted the need for more rigorous RCTs. Collectively, the findings indicate promising yet inconclusive evidence regarding the efficacy of self-efficacy–based interventions for adolescent Internet addiction [55]. Individuals with internet addiction, due to low self-efficacy, tend to turn to the internet when faced with life challenges, while individuals with high self-efficacy strive to control the events that affect their lives, such as internet usage. Bringsvor et al. (2018) investigated the effect of a health-promoting self-management intervention on domains related to self-management, self-efficacy and sense of coherence among COPD patients. They found that the intervention did not significantly affect general self-efficacy and sense of coherence [56]. Discrepancy in the results could be attributed to the point that the post-test scores were assessed shortly after the intervention period by three nurses with specific salutogenic orientations. Additionally, due to the relatively stable nature of self-efficacy and sense of coherence as personal traits, these changes may require more time to develop. Therefore, the timeframe for detecting changes in these aspects may have been too short in their study.

With all these interpretations, it is also important to consider possible sources of bias that may have influenced our findings. Selection bias could have occurred due to voluntary participation, meaning that more motivated adolescents may have been overrepresented. Social desirability bias might have led participants to report lower internet use or higher self-efficacy scores. Observer bias was possible because the same researcher facilitated and evaluated the sessions. In addition, the interactive format of the sessions could have temporarily increased participants’ perceived control and self-efficacy during the intervention phase, inflating post-test outcomes without ensuring sustained behavioral change.

In this study, efforts were made to empower participants through a five-step design. This research focused on two psychological domains of self-efficacy and SOC, which encompass various physical, mental and social aspects. It appears that the adolescent girls who participated in the study successfully reduced their internet addiction following the intervention. This improvement may be attributed to their comprehensive understanding of personal problems and the negative effects of excessive internet use on health, social relationships, and other aspects of life. Moreover, through receiving guidance on corrective strategies and participating in multiple follow-up sessions, they developed a sense of self-efficacy that enabled them to continue applying both the suggested solutions and those tailored to their individual circumstances. Overall, high SOC and self-efficacy enable individuals to use health-related strategies when faced with common challenges, rather than relying on tools like the internet, they giving their lives a sense of meaning.

Limitations

This study included several limitations. The absence of blinding constitutes a potential source of bias, possibly affecting the study’s internal validity. Moreover, individual participant characteristics, including mental and emotional conditions, as well as cultural and environmental factors influencing how participants perceived the steps of the intervention, limit the generalizability of our findings. Additionally, the intervention was limited to adolescent girls (grades 7–9) due to time constraints and the preference for same-gender participants and interviewer. The follow-up period was only two months, preventing the assessment of long-term effectiveness. A minor limitation is that the sense of coherence questionnaire showed only moderate reliability at baseline, although its internal consistency improved to a good level at post-test. Although both ANCOVA and LMM analyses showed robust improvements in sense of coherence and self-efficacy, the results must be interpreted cautiously. The small number of clusters limits the precision of the school-level variance estimates, and the absence of an attention control group restricts our ability to separate intervention-specific effects from general contact effects. In addition to these limitations, potential biases such as selection, observer, and social desirability bias should be considered when interpreting the results. These factors may have magnified short-term improvements and limited external validity. Future research should include blinded evaluators, longer follow-up durations, and mixed-gender samples from varied cultural backgrounds to enhance generalizability.

Despite these limitations, this study demonstrated several strengths, including a randomized controlled design, structured implementation of the self-management empowerment model, and the use of validated psychometric tools. Its focus on a specific group of adolescent girls adds valuable insight for developing context-sensitive preventive strategies against internet addiction.

Conclusion

While the intervention demonstrated a substantial positive effect on sense of coherence and self-efficacy, these outcomes should be interpreted with caution due to the several potential biases, short follow-up period and the homogeneity of participants. Empowering adolescent girls through the development of self-management skills appears to be a particularly promising strategy for strengthening core psychological resources, promoting cognitive awareness, and enhancing emotional regulation. This empowerment facilitates a shift from passive internet dependency and addiction toward autonomous and adaptive digital behavior, thereby improving self-regulation, sense of coherence, and self-efficacy. Rather than focusing solely on restricting internet use, future interventions should prioritize strengthening adolescents’ self-management and self-efficacy skills to promote responsible and balanced digital engagement with potential implications for broader educational and clinical practices.

Abbreviations

ANCOVA

Analysis of covariance

ADHD

Attention-deficit/hyperactivity disorder

COPD

Chronic obstructive pulmonary disease

CONSORT

Consolidated Standards of Reporting Trials

COVID-19

Coronavirus disease 2019

GLM

General linear model

GSES

General self-efficacy scale

IAT

Internet addiction test

RCT

Randomized controlled trial

SOC

Sense Of coherence questionnaire

Authors’ contributions

The contribution of the authors to the research was as follows: Research design by MR, SB, NM, MB; data collection by SB; research execution by MR, SB, NM, MB; data analysis by NM; documentation by MR, SB, NM, MB; and primary responsibility for the final content by MR, SB, NM, MB. All the authors have read and approved of the final manuscript.

Funding

This work was supported by the Research Council of Bushehr University of Medical Sciences (grant number: IR.BPUMS.REC.1401.078). The Research Council of Bushehr University of Medical Sciences had no role in conducting the study, data analysis or interpretation, writing the manuscript, or decisions about submitting the script for publication.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the ethics committee (code: IR.BPUMS.REC. 1401.078) of Bushehr University of Medical Sciences, Bushehr, Iran; and registered in the Iranian Clinical Trial Registration System (IRCT20130422013092N11). Informed consent was obtained from all the participants involved in the study or their legal guardians. The authors do not have any conflict of interest to declare.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.