Introduction

Globally, smoking-related diseases are considered the most preventable cause of mortality and morbidity1. Cigarette smoking by young adults, apart from the immediate adverse health consequences like addiction, accelerates the lifetime risk of chronic health problems including cancer and cardiovascular morbidity2. As per the World Health Organization (WHO), global tobacco use causes more than 8 million deaths annually andthe burden of its use is known to be significantly influenced by factors such as socioeconomic status and geography1,3.

Oman context

The ethnic Arab population in Oman, a country in the Middle East region, has experienced impressive socioeconomic and demographic transformations over the past few decades. Tobacco use remains a significant public health concern among Omani males aged > 15 years and there are limited recent studies that address the magnitude of the problem Omani adolescent4,5,6,7. Studies conducted in Oman to determine the prevalence of tobacco use among different age groups have showed that smoking rate in Omani population has been lower than in other Gulf Cooperation Council (GCC) countries, namely, Bahrain, Kuwait, Qatar, Saudi Arabia and the United Arab Emirates5,6,7,8,9.

The prevalence of tobacco use in Oman has been a topic of concern in recent years7. As per the WHO-FCTC 2020 report, the overall prevalence (male and female) of tobacco use was 10.6% in Oman with a prevalence of current smoking at 14.7% among males aged > 15 years8,10. The prevalence of exposure to secondhand smoke at school is 20.69% and student cohorts are reported to have higher rates of tobacco use than adults6,8.

Several factors have been identified as contributing to the prevalence of tobacco use in Oman. Tobacco consumption is often an acceptable sociocultural phenomenon and a symbol of social status and masculinity in Omani culture5. Peer pressure is another contributing factor for tobacco use among young adults and most current tobacco users begin using tobacco before age 18 in GCC8,11. The region has lower average prices of tobacco products compared to the Western countries12. There have been concerns that all major public health initiatives to lower tobacco use are interfered with by the tobacco industry in Oman13.

Interventions against tobacco use

Despite the global burden of tobacco use among adolescents, evidence supporting the effectiveness of community-based interventions remains inconclusive, as highlighted in a recent Cochrane review14. While programs informed by social cognitive theory and family-based approaches have shown moderate success, their impact sizes are modest, and their applicability to diverse contexts requires careful interpretation. A meta-analysis of 23 RCTs demonstrated a reduction in adolescent smoking habits through prevention programs (OR = 0.85, 95% CI = 0.77–0.93), with high-intensity school-based interventions led by trained instructors emerging as particularly effective15. The studies considered in the analysis were diverse, but the impact sizes were modest and hence, interpretation of the results must be done carefully.

School-based programs such as Education Against Tobacco (EAT) have shown promise by leveraging the social influence of the school environment and engaging students with tech tools such as photoaging app16,17. However, existing programs often lack sufficient cultural adaptation and have primarily been evaluated in high-income settings, raising questions about their generalizability to other regions such as the GCC. Sociocultural and economic factors in GCC differ significantly from Western contexts and there is a conspicuous lack of school-based interventional trials addressing adolescent tobacco use. This gap highlights the critical need for context-specific strategies that not only foster anti-smoking attitudes but also equip young individuals with the skills to resist peer pressure.

The current study (Clinical Trials NCT06483763 Date: 03-07-2024) aimed to generate actionable insights for designing tailored adolescent tobacco prevention programs by evaluating two intervention strategies: Photoaging App (PA) and School-based educational module (SBM).The primary objective was to evaluate the differences in knowledge about the health effects of tobacco use between the intervention groups. Secondary objectives included comparing attitudes towards tobacco use, and self-reported tobacco consumption behaviors across interventions.

Methods

Design

A randomized trial with two parallel arms was designed to evaluate the effectiveness of educational interventions on tobacco use knowledge and attitudes among adolescents18. We employed cluster randomization at the school level (1:1 ratio) to avoid the risk of contamination that could occur in individually randomized trials. This design protected the internal validity of the study while maintaining practical feasibility in real-life educational settings.

Participants

Participants included students from grades 9 to 12 in boys’ public schools in Sohar, North Batinah region, Oman. All boys’ public schools in Sohar, North Al-Batinah Region, from classes 9 to 12, were eligible for participation. Girls’ schools and parents and/or students who declined participation were excluded from the study. The homogeneous nature of Omani public schools—with standardized governance, funding, and resources—along with our focus on male students in grades 9–12, provided comparable clusters subject to similar curricula and cultural influences. Inclusion of only boys was guided by epidemiological data indicating significantly higher smoking rates among boys compared to girls in Oman6,10. Additionally, cultural norms and logistical considerations in Omani schools often necessitate separate interventions for boys and girls.

Photoaging App-based intervention (PA) The intervention involved the use of photoaging smartphone application (Smoker Face version 2.2) developed under the Education Against Tobacco (EAT) program16,17,18. The application requires users to take a selfie which is processed by software to generate images depicting the consequences of smoking (or not smoking) one pack a day over 1 to 15 years19. The application utilizes visual simulation of photoaging effects to enhance engagement and provides information specific to the cosmetic changes caused by smoking, focusing primarily on aesthetic impact.

The information regarding the application was delivered within the classrooms of enrolled schools and the participants were taught about how to use the app at home under parental supervision. The study information sheet shared with parents clarified their roles and responsibilities to make sure that the participant does not use the phone for some other activity of interest. The links to download the app from the App Store (iOS) and Play Store (Android), were shared with the parents on a social media application (WhatsApp Inc.) along with brief instructions on how to install and use the app. Selfie pictures were not shared with investigators to protect privacy.

School-based intervention (SBM) The intervention consisted of a pre-validated educational module developed in alignment with the World Health Organization Tobacco Health Hazards Awareness Toolkit20. The module was adapted to fit the Omani sociocultural context and designed by the core team of investigators to promote a tobacco- and nicotine-free environment. The adaptation process included validation by three external experts in the related fields of family medicine, child psychology and public health who reviewed the content for comprehensiveness and alignment with global tobacco prevention guidelines.

The module was delivered in 3 parts as a structured 90-minute in-person session by a team of three investigators within the classrooms of enrolled schools. In Part-1, a PowerPoint presentation in Arabic prepared based on the WHO tobacco hazard awareness content was used as an instructional tool20,21. The slides featured detailed explanations, visual aids, and key facts about the health risks of smoking and secondhand smoke exposure. It was followed by visual demonstration (Part-2) which included a gross specimen of a smoker’s lungs used to demonstrate the physical impact of smoking, particularly smoking-related lung cancer. This interactive demonstration allowed students to observe tangible representations of smoking’s detrimental effects on the lungs. Part-3 consisted of a brief video clip of an experiment to illustrate how cigarette smoke gets deposited in lung tissue. This experiment simulated how smoking leads to tar accumulation in the lungs, reinforcing the physiological consequences of tobacco use.

Participants were encouraged to ask questions and share their thoughts during the session to foster active engagement and critical thinking. Printed handouts summarizing the key educational points were distributed to participants for reference. These materials included infographics, fact sheets, and actionable steps for promoting a tobacco-free lifestyle. To extend the learning experience beyond the classroom, printed educational materials were also provided for home use.

The primary intended difference between the intervention arms was the content focus—photoaging-related consequences versus health hazards of tobacco use. Both interventions incorporated visual elements and interactive components, though these differed in modality (digital vs. in-class demonstration). The inclusion of the lung model and video demonstrations aimed to achieve a similar level of engagement while leveraging in-person, classroom-based delivery.

Procedure

Both interventions in the study were implemented by a team of three medical students from the College of Medicine and Health Sciences in the National University of Science and Technology, Sohar, Oman. They underwent a one-day in-person workshop facilitated by the lead investigators to achieve standardized delivery of the intervention and follow-up schedules. To ensure accuracy, surveys were completed in classrooms under the supervision of the team, who provided clear instructions and monitored participants to prevent discussion of responses. The trial concluded after completing the follow-up assessments.

Assents were obtained from participants after providing them with an age-appropriate explanation of the study. Written informed consent was obtained from the parent or legal guardian of participants. To ensure confidentiality, unique alphanumeric codes were assigned to each participant, and these codes were used for all data collection and analysis. Personal identifiers were stored separately from survey data in a secure, password-protected database accessible only to authorized members of the research team.

Outcomes

The outcomes in this study were assessed at two time-points: baseline (pre-intervention) and 3 months post-intervention. These assessments were conducted using the Global Youth Tobacco Survey (GYTS), a standardized component of the Global Tobacco Surveillance System which evaluates various dimensions of tobacco use, including prevalence, initiation, cessation, knowledge, attitudes, and exposure to secondhand smoke22.

The primary outcomes assessed participants’ perceptions of harm associated with tobacco use, specifically their views on the harmfulness of both tobacco use and secondhand smoke, measured using an ordinal scale that categorizes perceptions as ‘definitely harmful,’ ‘probably harmful,’ or ‘definitely not harmful.’ Additionally, susceptibility to tobacco use was assessed through a pre-validated susceptibility score ranging from 0 to 22, with higher scores indicating a greater likelihood of future tobacco use (Supplementary Table 1). Susceptibility scores will be categorized into three levels: Low (0–7), Moderate (8–14), and High (15–22). These measures allowed for quantification of the intervention’s impact on reducing susceptibility to tobacco use.

Secondary outcomes included participants’ attitudes toward tobacco use, evaluated through their support for smoking bans in both enclosed and outdoor spaces, along with the prevalence of non-smoking practices among participants.

Sample size

The sample size calculation was designed to determine the number of groups required per condition to detect the hypothesized intervention effect of 0.80 with a Type I error rate of 0.05 and a power of 0.80. The primary outcome is assumed to be a count variable with an event rate of 0.01 and no overdispersion. An intraclass correlation coefficient (ICC) of 0.05 is used to account for clustering effects23. Considering the typical class enrollments in Omani public schools, the expected number of participants per cluster is set at 25, with a coefficient of variation of group sizes of 0.97 reflecting moderate variability in class sizes, based on a pilot study. Considering these parameters, it was determined that a total of 6 clusters (3 per condition) were required24. To account for participant attrition, the number of samples within each cluster allowed an increase of 20%. This adjustment ensured that the study retained sufficient power and precision despite possible dropouts.

Randomization

Schools were stratified by grade levels (Classes 9–10 and 11–12), to ensure balanced representation of younger and older adolescents across intervention groups.

Within each stratum, schools were randomized to one of the two intervention arms using a computer-based random number generator by an independent research assistant. The randomization process ensured balance between intervention arms and minimizing bias. All students in each school received the same intervention to avoid contamination and ensure homogeneity of exposure within clusters.

While blinding participants and intervention administrators was not feasible due to the nature of the interventions, outcome assessors and data analysts remained blinded to group assignments to avoid bias during data analysis. Though schools were not explicitly matched prior to randomization, stratification by grade levels ensured balanced representation across intervention arms. Additionally, demographic and baseline characteristics of participants were assessed and adjusted for in the statistical analyses to account for any potential differences between groups.

Statistical methods

In this study, continuous variables were presented as mean, median, standard deviation, and interquartile range, while categorical variables were presented as frequency and percentage. Normality of continuous variables was assessed using the Shapiro-Wilk test, and homogeneity of variances was evaluated using Levene’s test. Transformations were applied where necessary to meet the assumptions of parametric tests.

For comparisons of mean susceptibility scores (continuous data) within-group pre-post comparisons were performed using paired samples t-tests. Between-group post-intervention comparisons, adjusting for baseline scores, were analyzed using analysis of covariance (ANCOVA). This allowed us to account for potential baseline differences in susceptibility scores while evaluating intervention effects. For categorical data, variables with multiple categories were dichotomized based on theoretical relevance and distribution patterns. Pre-post changes within groups were analyzed using McNemar’s test for paired categorical data.

To account for the clustered design and hierarchical data structure, a mixed-effects logistic regression model was employed for categorical outcomes, considering both fixed effects (intervention group, baseline characteristics) and random effects (school clusters)25. A random intercept was included for each school to account for intra-cluster correlation. Model fit was evaluated using the Akaike Information Criterion (AIC) and likelihood ratio tests. For dichotomized susceptibility scores (low vs. moderate-to-high), a generalized estimating equations (GEE) model was used to evaluate changes over time and between groups while addressing intra-cluster correlation and repeated measures. All analysis were conducted following the intention-to-treat principle A P-value < 0.05 was considered statistically significant and SPSS software (IBM SPSS Statistics for Windows, Version 29.0. Armonk, NY) was used for analysis.

Results

Baseline characteristics

The participants’ flow diagram is shown in Fig. 1. The study included a total of 227 adolescents, with 105 in the school-based educational module group and 122 in the photoaging app group. The baseline demographics of the two groups were similar, as shown in Table 1. The mean age of participants was 16.8 years (SD 1.3), with a similar distribution across the age groups between the two study arms. The participants were in grades 9–12, with a comparable distribution across the grades. In agreement with CONSORT guidelines, baseline differences were reported descriptively to avoid over-interpretation of baseline statistical differences26. The final sample sizes in the intervention groups differed slightly due to student absences during the enrollment and allocation phases. Absences were attributed to factors such as illness, school activities, or other external commitments, which are common in real-world school settings. A statistically significant baseline difference between the groups in weekly expenditure on personal needs (p = 0.004) was accounted for in the adjusted analyses.

Fig. 1
figure 1

The participants’ flow diagram.

Full size image
Table 1 Demographic characteristics of the study population.
Full size table

Pre-intervention knowledge, attitude, and practices

Baseline characteristics of the study participants are shown in Table 2. The majority, 89.0% (n = 202), were not currently smoking (i.e., smoked for at least 1 day during the past 30 days). Among the participants categorized as current smokers, 56% (n = 14) expressed a willingness to quit. Participants reported having discussed the harmful effects of tobacco use with their families (68.3%), and/or schools (49.8%) while 31.7% (n = 72) reported no such discussions. Most participants believed secondhand smoke is harmful, with 39% perceiving it as definitely harmful and 33.5% perceiving it as probably harmful. About 80% of participants (n = 181) favored banning smoking in outdoor public places.

Table 2 Baseline knowledge, attitude and practices of participants about tobacco and its use.
Full size table

Participants in the two intervention groups were comparable in their perceptions of the attractiveness of tobacco use, willingness to use tobacco if offered by a best friend, or support for banning smoking in public places. There was no significant difference between the groups in the perception of harm from tobacco use or from secondhand smoke. A higher proportion of participants in the PA group (43.4%) believed that smoking tobacco would help people feel more comfortable in social gatherings and agreed that they might enjoy smoking (16.4%), compared to the SBM group (24.8% and 8.6% respectively). Furthermore, a higher proportion of participants in the PA group (83.6%; n = 102) believed that their peers use tobacco compared to the SBM group (p = 0.004) and the same group had a higher proportion of current smokers (10.6%) who did not want to quit, compared to the SBM group (5.7%). The clusters did not differ significantly in the number of cigarettes smoked daily, attempts to quit smoking, or the duration of experiencing strong cravings after smoking.

Effectiveness of the SBM and photoaging app

A comparison of post-intervention knowledge, beliefs, and practices related to tobacco use between the interventions is provided in the Supplementary Table 2. Regarding the willingness to use tobacco, a significantly higher proportion of participants in the PA group reported that they would “definitely” not use tobacco if offered by a best friend compared to the SBM group (95.7% vs. 83.5%, p = 0.031). The SBM group showed higher rates of support for banning smoking in both enclosed (89.4% vs. 84.9%, p = 0.335) and outdoor (88.2% vs. 81.7%, p = 0.228) public places compared to the PA group. There were no significant differences between the groups in the willingness to stop smoking or the willingness to use tobacco in the next 12 months.

The results of mixed effects logistic regression analysis of post-intervention outcomes related to knowledge, susceptibility, and attitudes toward tobacco use while accounting for clustering effects and controlling for important covariates are shown in Table 3. Participants in the PA group demonstrated a higher likelihood of recognizing tobacco as harmful compared to those in the SBM group. Within-group and between-group pre-post comparisons of key parameters are shown in Table 4. Post-intervention susceptibility scores for tobacco use were 7.81 (SD 3.21) for SBM group and 7.92 (SD 2.78) for PA group with adjusted mean difference 0.11 (CI: – 0.53, 0.75) and p-value of 0.732. The results of pre-post comparison in the GEE analysis indicated that participants in the PA group were 21.2% less likely to show moderate to high susceptibility to tobacco use compared to those in the SBM group (OR 0.788, 95% CI: 0.525–1.184). However, this difference was not statistically significant (p = 0.252).

Table 3 Adjusted analysis of post-intervention outcomes with mixed effects logistic regression*.
Full size table
Table 4 Adjusted analysis of post-intervention comparisons of key outcomes.
Full size table

Discussion

The current study makes several unique contributions to tobacco prevention literature. First, it compares two active interventions—a standalone photoaging app versus a culturally tailored educational module—rather than testing against no-intervention control. Second, it is the first trial in the GCC region to examine educational interventions addressing tobacco use in school children. Third, unlike previous trials that combined photoaging with the EAT curriculum, we tested the independent effectiveness of the photoaging app16,17.

Our findings contribute to the growing evidence base for school-based smoking prevention programs and are in alignment with the recent meta-analysis demonstrating modest reductions in adolescent smoking behaviors through such interventions (OR = 0.85, 95% CI = 0.77–0.93)15. The PA intervention in this RCT mirrors similar strategies observed in a Brazilian trial which showed a significant mitigation of smoking prevalence observed over 12 months17. This reduction in tobacco susceptibility and a positive shift in attitudes toward smoking, underscoring the relevance of such technology-based interventions in adolescents. Compared to the SBM intervention, PA was particularly effective in enhancing perceptions of the harmful effects of smoking, reducing tobacco susceptibility by 21.2%. In a trial by Lisboa et al., where smoking prevalence in the group using photoaging app increased only marginally from 14.1 to 15.6% over 12 months, compared to a significant rise in the control group (11.0–20.9%; P < 0.01)17.

Statistical analysis adjusted for baseline differences demonstrated that the photoaging app was significantly more effective than SBM in enhancing adolescents’ perceptions of the harmful effects of tobacco (p = 0.002) and secondhand smoke (p = 0.014). In contrast, the SBM showed a relative advantage in promoting support for smoking bans in enclosed spaces (p = 0.065) and public places (p = 0.083), highlighting this approach’s strength in fostering community-oriented attitudes toward tobacco control27. These findings suggest that PA exceled in delivering personalized health messages that resonate at the individual level, particularly by emphasizing the direct health consequences of smoking. This finding could potentially reflect the complexity of behavioral change interventions and highlights the likely benefits of hybrid programs that integrate the strengths of both traditional and tech-based approaches28. Supplementary modules to address broader tobacco-related issues, including the environmental and societal impacts of tobacco use could enhance the effectiveness of the photoaging app.

While prior research suggested that photoaging interventions might be less effective for males compared to females due to differences in self-concept and responsiveness to visual cues, our findings challenge this notion16. In our trial, which exclusively recruited male participants, the app demonstrated a significant reduction in tobacco susceptibility and positive shifts in attitudes toward smoking. These findings highlight the potential of digital tools to appeal to adolescents’ interests in appearance and leverage their familiarity with smartphones, offering a scalable, cost-effective intervention29,30.

Despite the known importance of anti-tobacco education in facilitating adolescent quit attempts, many countries lack adequate youth-focused cessation programs31. A recent behavioral modification intervention among secondary school students in Ghana failed to show significant effects on smoking behavior at 2 weeks follow-up32. While social norms and short-term interventions for substance use among adolescents have shown limited success, culturally adapted, context-specific strategies can overcome some of these challenges32,33,34. The efficacy of our interventions among adolescent boys at 12 weeks follow-up highlights the potential of gender-focused strategies, especially in sociocultural contexts where gender norms may influence receptivity to health messaging35,36. This is particularly important as much of the existing evidence comes from high-income, developed countries and may not be directly applicable to regions with differing sociocultural and economic contexts, such as Oman and the broader GCC.

Strengths and limitations

The strengths of the study include its randomized design, robust statistical methods, and use of GYTS, a standardized assessment tool to measure knowledge, attitudes, and practices. Social desirability bias in self-reported measures of smoking behaviors and attitudes is one of the potential limitations of the study. The study sample consisted of adolescents from public schools, which may limit the generalizability to adolescents from private school settings. The differences between intervention designs in terms of instruction modality, exposure frequency, and parental supervision are other potential limitations of the study. However, these could highlight the complexity of real-world implementation of educational interventions. Participant attrition is a limitation in this study. Although retention strategies such as school-based follow-ups and parental reminders were employed, a proportion of participants did not complete post-intervention assessments.

Conclusion

This trial demonstrates the effectiveness of both traditional and technology-based approaches in tobacco prevention among adolescent males. The photoaging app proved particularly effective in enhancing personal risk perception, while the school-based module showed strengths in promoting community-level tobacco control awareness. The study results could have immediate implications for public health practice in Oman and similar cultural contexts36. The standalone effectiveness of the photoaging app suggests a scalable, resource-efficient alternative to traditional curriculum-based programs, particularly in settings where extensive classroom time may not be available. However, the differential impacts of each intervention suggest that a hybrid approach might be optimal.

Our findings provide a foundation for evidence-based tobacco prevention strategies that are both culturally appropriate and technologically innovative, potentially transforming how we approach adolescent tobacco prevention in the region. Future research should consider integrated interventions that combine digital and traditional approaches and long-term follow-up to assess sustained behavioral outcomes. A multi-method outcome assessment incorporating qualitative interviews, and parental reports will further improve the strength of the research findings.