Association between substance use disorder and risky behavior among students in Indonesia: a cross-sectional study

Abstract

Background

Substance use is prohibited in Indonesia; however, adolescents continue to consume these substances, creating a significant public health challenge. This study examines behavioral and demographic factors associated with general drug use, including cannabis and amphetamine use, among Indonesian adolescents, using nationally representative data.

Methods

A cross-sectional study used data from the 2023 Indonesia Global School-based Student Health Survey, which included 10,059 students aged 13 to 17 years. A two-stage cluster sampling was applied, and descriptive and regression analyses were conducted to identify factors associated with substance use.

Results

Overall, 3.03% of students reported drug use, 1.45% reported cannabis use, and 1.33% reported amphetamine use. Age and school level were not significantly associated with substance use. However, male sex, residence outside the Java-Bali region, alcohol use, and tobacco use were consistently linked to higher odds of drug, cannabis, and amphetamine use. Alcohol consumption was the strongest predictor for all types of substance use. These findings indicate that risky behaviors tend to cluster, supporting the pattern of polysubstance use among youth.

Conclusion

The study findings highlight the need for risk-specific, gender-responsive, school-based prevention programs that actively involve parents and regulate adolescents’ access to alcohol and tobacco. Enhancing life skills and character education in both school and family settings may also help students avoid risky behaviors.

Graphical Abstract

Introduction

Adolescence is a critical stage marked by identity exploration and increased vulnerability to external influences, often resulting in a higher likelihood of risk-taking behaviors influenced by peer pressure [1]. These patterns of behavior are a source of concern and unpredictability for families and societies.

Adolescents’ behaviors contribute to major global public health challenges, including increasing rates of youth delinquency and substance use, which are serious concerns for many countries. Many countries in Southeast Asia have reported a significant increase in juvenile crime [2,3], with many offenses closely related to substance abuse and excessive alcohol consumption. International research consistently shows strong associations between drug use and other high-risk behaviors, such as unprotected sexual activity, violence, and poor academic performance [4,5].

Indonesia has faced similar issues, as national data show increasing adolescent involvement in bullying, violence, group brawls, sexual harassment, and drug abuse. Act 35 of 2009 strictly prohibits substance use in Indonesia. The Indonesian Child Protection Commission reported that 2,883 minors were involved in legal cases as perpetrators of delinquency from 2016 to 2022 [6]. At the same time, the National Narcotics Board reported a steady increase in drug use among students, who are a high-risk group because of psychological vulnerability, social pressure, and environmental exposure [7].

Despite these issues, empirical research on the specific behavioral consequences and risk factors of substance use within the cultural context remains limited. In Indonesia, overall substance use, particularly among adolescents, has only been estimated using nationally representative surveys. This lack of evidence restricts public health authorities and educational institutions from developing targeted, evidence-based interventions and prevention strategies that align with the cultural needs of Indonesian youth. Therefore, this study uses data from the Indonesia Global School-based Student Health Survey (GSHS) to examine the association between substance use and specific risk behaviors among school-aged adolescents in Indonesia.

To understand the complex nature of adolescent substance use in Indonesia, we used the Socio-Ecological Model (SEM) as the theoretical framework, which highlights the interaction between individual behavior and different environmental levels. Using SEM, we analyzed the determinants of substance use at several interrelated levels: intrapersonal (individual characteristics), interpersonal (relationships with family and peers), community (geographical and social environment), and macrosystem (broader societal norms and policies) [8].

By identifying key behavioral patterns associated with substance use, this research aims to provide empirical evidence to inform the development of targeted prevention strategies, school-based interventions, and public health policies in Indonesia. A stronger understanding of these dynamics is essential for creating focused awareness campaigns and early intervention programs to reduce the adverse effects of drug-related harm among Indonesian youth.

Methods

Study design and data source

This study used a cross-sectional design with data from the 2023 GSHS. The GSHS, a collaboration between the World Health Organization (WHO) and the U.S. Centers for Disease Control and Prevention (CDC), collects data on behavioral risk factors and protective factors among adolescents aged 13 to 17 years. The survey applied a two-stage cluster sampling design to obtain a nationally representative sample of students from junior and senior high schools throughout Indonesia.

Study population and sample

The target population included students aged 13 to 17 years attending grades 7 to 12 in junior high schools, Islamic junior high schools, senior high schools, Islamic senior high schools, and vocational high schools. The sampling frame comprised public and private schools listed by the Ministry of Education. In the first stage, 79 schools were selected using probability proportional to enrollment size. In the second stage, classes within these schools were randomly chosen from all eligible grades. All students in the selected classes were invited to participate in the survey. Students who were absent, unwell, or unable to complete the questionnaire were excluded. The school response rate was 100%, and the student response rate was 84.6%. The final analysis included 10,059 observations with complete responses to key variables.

Study setting and data collection procedure

The survey was conducted in selected provinces across three main regions of Indonesia: Sumatra, Java-Bali, and other islands beyond Java and Sumatra. This method ensured national representation and reflected the country’s geographic and cultural diversity. Data were collected from 7 August to 1 September 2023, using a standardized, self-administered questionnaire developed by WHO, CDC, and UNICEF, and adapted for Indonesia. The GSHS questionnaire contains 10 modules on health-related topics, such as dietary habits, tobacco use, physical activity, sexual behaviors, alcohol consumption, and substance use (drugs, amphetamine, and cannabis).

The Indonesia GSHS questionnaire was developed through collaboration among the national GSHS team at the National Research and Innovation Agency (BRIN), the Ministry of Health, the Ministry of Education, and the WHO Country Office in Jakarta. The 2023 version included additional items on social media use. After development, the questionnaire was submitted to WHO Headquarters and the WHO Regional Office for South-East Asia (SEARO) for review. It was translated into Bahasa Indonesia, then back-translated into English by an independent translator. Experts at the CDC Atlanta and WHO Headquarters reviewed the back-translated version to ensure consistency and equivalence with the original. Final approval was granted when no significant differences were found between the Bahasa and English versions. The translation and validation process took more than 2 months.

Survey administrators collected data during regular class hours in selected schools. Student participation was anonymous and voluntary, with informed consent from parents or guardians and agreement from students. Administrators also collected anthropometric data, including students’ height and weight, using standardized measurement protocols.

Variables

The analysis examined variables related to adolescent substance use and associated behavioral and social factors. Lifetime use of three substances was assessed using self-report items: general illicit drug use (including cannabis/marijuana, glue, heroin, amphetamines, or other intoxicating substances), cannabis use (marijuana), and amphetamine use (locally known as “shabu”). Each item was coded as yes or no to indicate whether the respondent had ever used the substance.

The study included several sociodemographic characteristics: age group (<16 years and ≥16 years), sex (male or female), grade level (junior high school, grades 7–9; senior high school, grades 10–11), and school area (Java-Bali region, Sumatra region, and others). The study also examined two behavioral factors commonly associated with substance use, such as tobacco use and alcohol consumption, both measured by reported use in the past 30 days. While the GSHS included several protective factors, this study analyzed only parental communication. Parental communication served as a proxy for perceived emotional support and connectedness between adolescents and their parents, based on evidence that positive parental communication creates strong family bonds, which are associated with greater self-efficacy among adolescents in refusing drug offers from peers [9]. All variables used in this study are described in Supplement 1.

Data analysis

We performed analyses using Stata ver. 15.0 (Stata Corp.). Descriptive statistics summarized the prevalence of substance use and participant characteristics. We used chi-square tests to assess bivariate associations between independent variables and substance use. We analyzed three outcomes: general drug use (a composite variable including any use of marijuana, glue, heroin, methamphetamines, or other intoxicating substances), cannabis use, and amphetamine use. We did not model other substance-specific items separately because their proportions were very low. We constructed separate multivariable logistic regression models for each substance use outcome, adjusting for all independent variables. The multivariate analysis included gender, region of schools, tobacco use, alcohol consumption, and communication with parents or guardians. Results are reported as adjusted odds ratios (AORs) with 95% confidence intervals (CIs). All analyses used survey weights to account for the sampling design. A P-value of 0.05 indicated statistical significance. Before analysis, the data set was cleaned for inconsistencies, and missing values were imputed using auxiliary variables correlated with the incomplete variables.

Ethical considerations

Ethical approval for the 2023 Indonesia GSHS was granted by the Ethics Commission for Health of the BRIN (No. 014/KE.03/SK/12/2022). The survey also received ethical approval from the SEARO Research Ethics Review Committee (SEARO_ERC) with unique proposal ID: 2023.10.NW, approved on 20 April 2024. Verbal consent to participate was obtained from parents the day before data collection. The questionnaire and answer sheet did not include personal identifiers, ensuring anonymity. Students could withdraw from the study at any time before or during data collection and could decline to answer any questions.

Results

As shown in Table 1, the prevalence of drug use among students was 3.03%, while cannabis and amphetamine use were 1.45% and 1.33%, respectively. The highest prevalence of all three types of substance use occurred among boys, students at higher school levels, and those living in regions outside Java-Bali and Sumatra (Figure 1). In contrast to general drug and cannabis use, amphetamine use was more common among younger students, as shown in Figure 1. Bivariate analysis found a statistically significant relationship between the use of all three substances and tobacco use, alcohol consumption, and parental communication (P<0.05). Other demographic factors, such as gender and school region, were also independently associated with the use of general illicit drugs, cannabis, and amphetamine.

Table 1.

Prevalence of drug use, cannabis use, and amphetamine use by characteristics, GSHS 2023

Variable	Drug use (n=9,599)			Cannabis use (n=9,596)			Amphetamine use (n=9,427)
	No. (%)	OR (95% CI)	P-value	No. (%)	OR (95% CI)	P-value	No. (%)	OR (95% CI)	P-value
Age group (y)			0.459			0.109			0.226
<16	176 (2.98)	Ref		130 (1.34)	Ref		123 (1.35)	Ref
≥16	13 (3.85)	1.29 (0.64–2.56)		9 (2.78)	2.01 (0.84–4.76)		3 (0.61)	0.44 (0.11–1.68)
Sex			<0.001			<0.001			<0.001
Female	50 (1.05)	Ref		21 (0.45)	Ref		18 (0.39)	Ref
Male	241 (4.94)	4.88 (2.53–9.40)		118 (2.42)	5.46 (2.51–11.86)		107 (2.26)	5.9 (2.50–13.93)
Grade			0.183			<0.001			0.789
Junior high school	282 (2.99)	Ref		133 (1.41)	Ref		123 (1.32)	Ref
Senior high school	9 (4.82)	1.64 (0.78–3.43)		6 (3.28)	2.36 (0.79–7.08)		3 (1.52)	1.15 (0.39–3.36)
Region
Java-Bali	87 (1.64)	Ref		24 (0.46)	Ref		22 (0.42)	Ref
Sumatra	84 (3.54)	2.200 (1.04–4.66)	0.039	53 (2.25)	5.01 (1.76–14.20)	0.010	34 (1.49)	3.57 (0.98–13.00)	0.053
Rest of the country	120 (6.26)	3.99 (1.61–9.87)	0.004	62 (3.22)	7.28 (2.32–22.83)	0.003	69 (3.67)	8.99 (2.15–37.51)	0.003
Used tobacco product			<0.001			<0.001			<0.001
Never	75 (0.96)	Ref		13 (0.17)	Ref		12 (0.15)	Ref
Yes, ever used	216 (11.85)	13.80 (9.39–20.27)		126 (6.95)	45.05 (19.81–102.44)		114 (6.62)	47.96 (21.90–101.67)
Drunk alcohol			<0.001			<0.001			<0.001
Never	156 (1.70)	Ref		51 (0.56)	Ref		41 (0.46)	Ref
Yes, ever drunk	135 (29.09)	23.65 (15.86–35.27)		88 (19.00)	41.51 (25.46–67.67)		84 (20.38)	55.74 (30.98–100.29)
Communication with parents/guardians
Most of the time/always	63 (2.11)	Ref		25 (0.85)	Ref		28 (0.97)	Ref
Rarely/sometimes	156 (3.41)	1.63 (0.97–2.73)	0.060	79 (1.73)	2.06 (1.13–3.77)	0.019	66 (1.46)	1.51 (0.95–2.38)	0.074
Never	72 (3.49)	1.67 (1.03–2.71)	0.036	35 (1.70)	2.00 (1.36–2.98)	0.001	31 (1.56)	1.61 (1.08–2.40)	0.019
Total	289 (3.01)			139 (1.45)			126 (1.33)

GSHS, Indonesia Global School-based Student Health Survey; OR, odds ratio; CI, confidence interval; Ref, reference.

Figure. 1.

Demographic and other characteristics of drug, cannabis, and amphetamine users among students.

Table 2 summarizes risk factors linked to all three types of substance use among students. In the multivariate analysis, after controlling for other variables, gender, tobacco use, and alcohol consumption were significantly associated with all three types of substance use. Alcohol use was the strongest predictor in the model. Alcohol consumption increased the risk of drug use by almost 8 times, cannabis use by 10 times, and amphetamine use by approximately 13 times (Table 2). Although tobacco use and alcohol consumption were associated with a notably increased risk of substance use, interpretation should be cautious because the wide CI may indicate variability or potential instability in the estimate. Regional differences were positively associated only with cannabis use, with students in the Sumatra region having a 4 times greater risk and those in regions other than Java-Bali and Sumatra having a 3 times greater risk.

Table 2.

Risk factors for all types of substance use among Indonesian students, GSHS 2023

Variable	Drug use		Cannabis use		Amphetamine use
	AOR (95% CI)	P-value	AOR (95% CI)	P-value	AOR (95% CI)	P-value
Sex
Female	Ref		Ref		Ref
Male	2.89 (1.55–5.38)	0.001	2.57 (1.17–5.66)	0.020	2.92 (1.28–6.62)	0.011
Region
Java-Bali	Ref		Ref		Ref
Sumatra	1.85 (0.96–3.55)	0.052	4.16 (1.73–9.99)	0.002	2.74 (0.84–8.95)	0.091
Rest of the country	2.18 (1.05–4.55)	0.048	3.12 (1.23–7.89)	0.017	4.13 (1.15–14.79)	0.030
Used tobacco product
Never	Ref		Ref		Ref
Yes, ever used	6.16 (4.37–8.69)	<0.001	17.15 (6.33–46.43)	<0.001	15.84 (6.37–39.41)	<0.001
Drunk alcohol
Never	Ref		Ref		Ref
Yes, ever drunk	7.74 (5.34–11.20)	<0.001	10.24 (5.95–17.63)	<0.001	13.03 (6.96–24.38)	<0.001

GSHS, Indonesia Global School-based Student Health Survey; AOR, adjusted odds ratio; CI, confidence interval; Ref, reference.

Discussion

Substance use among adolescents remains a critical concern in Indonesia and globally because it is associated with several risk factors, including sociodemographic characteristics, smoking, alcohol consumption, and lack of parental attachment. Evidence from this study shows significant differences in substance use by sex, region, and other risky behaviors, such as smoking and alcohol use.

Gender differences

At the individual and interpersonal levels, boys reported higher lifetime use of drugs, cannabis, and amphetamines compared to girls, indicating gender-specific risk profiles. These sex differences may result from distinct developmental pathways: boys tend to show higher sensation-seeking at the individual level, while at the interpersonal level, peer pressure and gendered social norms contribute to these patterns [10,11]. Other studies have also found that gender patterns are often associated with social norms, gender expectations, and risk tolerance [12]. Together, these findings indicate that masculinity norms and peer influence are important for understanding gender differences in substance use in Indonesia.

Demographic stratification by age and school level did not significantly differentiate substance use patterns; instead, psychosocial and environmental factors were more influential determinants. This suggests that SEM interpersonal and community-level factors are more important than basic demographic characteristics. This finding aligns with previous studies indicating that psychosocial and environmental factors may play a more central role in shaping substance use behaviors than demographic attributes alone [13]. Notably, across all demographic categories, alcohol and tobacco use showed the strongest associations with other substance use, indicating that polysubstance involvement may result from shared behavioral or neurobiological pathways rather than age-related progression [14].

Gateway effects of alcohol and tobacco

Two complementary SEM frameworks explain these associations. At the individual and biological level, the gateway hypothesis proposes that early legal substance use lowers psychological and neurobiological barriers to illicit drug experimentation by impairing decision-making and increasing craving sensitivity. At the interpersonal and community levels, the common liability model suggests that shared underlying vulnerabilities, such as sensation-seeking tendencies, peer influence, and permissive social norms, predispose adolescents to engage in multiple risk behaviors at the same time [15,16]. This neuropsychological overlap between alcohol and stimulant abuse may increase risk-taking behavior and reduce inhibition through shared neurotransmitter systems.

In contrast, the common liability model highlights that underlying psychosocial vulnerabilities and environmental influences explain concurrent substance use. Within this model, peer influence and social norms at the interpersonal level, substance availability and normalization at the community level, and sensation-seeking and behavioral dysregulation at the individual level together contribute to simultaneous use of multiple substances [17]. Tobacco and alcohol can serve as gateways in some cases; in others, common liabilities such as psychosocial vulnerabilities and environmental influences explain the co-occurrence of substance use and risky behaviors. Students who used alcohol or tobacco had a significantly higher likelihood of using other substances, supporting the clustering of risky behaviors during adolescence [18,19].

Prior studies have reported a strong comorbidity between nicotine dependence and stimulant abuse [12]. This association may result from shared neurobiological mechanisms or behavioral pathways that increase the risk of using both substances. In the context of Indonesia, tobacco use is the most common initial substance among adolescents, influenced by cultural normalization, low cost, and the widespread availability of kretek (clove cigarettes) [20,21]. Although public health campaigns have increased awareness, smoking increases in frequency and dependence during adolescence, which may reduce barriers to using more dangerous substances [22]. New products, such as e-cigarettes, indicate changing patterns of polysubstance initiation, showing that community-level environments continue to shape adolescent substance exposure [23].

Cannabis, often called marijuana and officially referred to as “ganja” in Indonesia, remains a public health issue because of its euphoric effects and potential for dependency. It is the most widely used illegal narcotic in Indonesia and is commonly chosen by adolescents experimenting with substances. Cannabis use is linked to higher risk of psychological distress, poor academic outcomes, and, in some cases, suicidal ideation [24].

Regional disparities

At the community and policy levels, regional differences significantly affect substance use patterns. For example, cannabis use shows clear geographic clustering; it is most prevalent in the northern part of the island of Sumatra, especially in the Aceh region. In Aceh, cannabis use is linked to cultural and traditional practices, with lower rates of abuse, as local communities use cannabis for culinary and medicinal purposes. In contrast, in other Sumatran regions, such as Bengkulu, Lampung, and Mandailing Natal, illicit commercial cultivation increases supply and leads to higher rates of abuse among adolescents [25]. Local reports indicate that cannabis is also widely cultivated in several areas of Java and the Kalimantan region. The high rate of cannabis abuse is closely related to its easy accessibility and lower price compared to other drugs [26].

At the community level, adolescents outside Java-Bali, especially in Central and Eastern Indonesia, showed higher amphetamine risk because of regional differences in substance availability, recreational use norms, and enforcement patterns. Amphetamine’s stimulant properties may increase its recreational appeal in cultural contexts that value alertness or work capacity. According to the National Drugs report, this substance is widely used for recreational purposes, as it acts as a stimulant that can accelerate the central nervous system at high doses, altering perception of reality and causing hallucinations [26]. The use of amphetamine, commonly known as shabu, is increasing among teenagers. Within the SEM framework, regional inequalities that may lead to differences in substance use at the community level include variations in social norms, socioeconomic conditions, and the availability and access to substances [27]. Regional disparities may reflect variations in the availability and accessibility of substances, as well as differences in law enforcement, education, and socioeconomic conditions across regions [28]. Understanding these community-level disparities identifies intervention targets, including addressing regional variation in substance accessibility, improving law enforcement consistency, and adapting prevention education to local contexts. In addition to community factors, individual and interpersonal protective factors also require examination.

Protective factors

At the interpersonal level, parental communication is traditionally considered a key protective factor against adolescent risk behaviors. However, in this study, its direct effect was not statistically significant, indicating that its protective influence may operate indirectly. This finding suggests that the effect of parental communication is likely mediated by other factors, particularly the selection and behavior of peer groups [29]. There is a possible causal pathway in which strong and positive parental communication supports resilience and social skills, enabling adolescents to form healthier and lower-risk peer relationships. This pathway may differ by gender: weak family functioning and low parental control are associated with higher substance use, particularly among male adolescents, potentially because failures in parental communication remove an important influence that counters strong peer pressure and sensation-seeking behaviors described earlier [30].

Study limitations

Although our study was nationally representative, several limitations should be noted. First, the analysis was limited to the variables available in the data set. Mental health measures, such as loneliness, suicidal ideation, and other depressive symptoms, were excluded to prevent over-adjustment, as these factors may act as mediators and confound the association between substance use and the main risk factors (alcohol consumption and smoking). In addition, socioeconomic variables (e.g., parental education, occupation, or wealth level) were not collected in the 2023 Indonesia GSHS data set. Consequently, our results did not adjust for students’ socioeconomic status; instead, we included region as an aggregate-level proxy for socioeconomic context. Second, because the GSHS is school-based, adolescents who are out of school and more likely to be socioeconomically disadvantaged were not represented, which may limit the generalizability of the findings to all adolescents. Third, we focused on parental communication because it is modifiable, theory-based, and well measured in the GSHS. Including additional similar protective variables (e.g., parental supervision, peer support) would reduce the clarity and stability of the findings. Fourth, although the GSHS questions reference multiple drugs (e.g., inhalants, cocaine, heroin, injecting), our analyses focused on cannabis and amphetamines because the frequencies of other substances were low. Finally, because data were self-reported, there is a potential for recall and social desirability bias, which may have led to underreporting.

Conclusion

This study provides empirical evidence on factors linked to adolescent substance use in Indonesia, using nationally representative data from the 2023 GSHS. Gender, school region, tobacco use, and alcohol consumption were consistently associated with the use of general drugs, cannabis, and amphetamines. The results also show that alcohol and tobacco use are closely related to the use of other substances, indicating a common pattern of polysubstance use among youth.

Recommendations

These results highlight the need for risk-specific and gender-responsive intervention strategies. Although substance abuse is often viewed as a men’s issue, providing equal treatment for women and men is essential. Gender-sensitive messaging, parental engagement, and school-based programs should be integrated with efforts to regulate student access to tobacco and alcohol. Enforcing smoking and alcohol bans in schools may involve enacting prohibitive regulations or age restrictions, conducting awareness campaigns on the dangers of smoking and alcohol consumption, and monitoring the school environment. This monitoring should include oversight of vendors near schools who sell alcohol and tobacco products, with support from community police officers or ‘Kamtibmas’ as territorial defense authorities.

Early prevention approaches should be implemented at the family level, including supporting positive character development in adolescents so they can avoid risky social relationships and refuse drug offers from peers. In addition, local authorities should strengthen their involvement by developing alternative positive activities for teenagers in coordination with community leaders and local youth organizations, including implementing community-based rehabilitation programs.

Supplementary materials

Supplementary materials can be found via https://doi.org/10.4082/kjfm.25.0237.

Supplement 1.

Description of the variables in the study.