Prevalence of Cannabis Lifetime Use in Iranian High School and College Students

Abstract

Cannabis is the most widely used substance in the world. This study aimed to estimate the prevalence of cannabis lifetime use (CLU) in high school and college students of Iran and also to determine factors related to changes in prevalence. A systematic review of literature on cannabis use in Iran was conducted according to MOOSE guideline. Domestic scientific databases, PubMed/Medline, ISI Web of Knowledge, and Google Scholar, relevant reference lists, and relevant journals were searched up to April, 2014. Prevalences were calculated using the variance stabilizing double arcsine transformation and confidence intervals (CIs) estimated using the Wilson method. Heterogeneity was assessed by Cochran’s Q statistic and I² index and causes of heterogeneity were evaluated using meta-regression model. In electronic database search, 4,000 citations were retrieved, producing a total of 33 studies. CLU was reported with a random effects pooled prevalence of 4.0% (95% CI = 3.0% to 5.0%). In subgroups of high school and college students, prevalences were 5.0% (95% CI = 3.0% to −7.0%) and 2.0% (95% CI = 2.0% to −3.0%), respectively. Meta-regression model indicated that prevalence is higher in college students (β = 0.089, p < .001), male gender (β = 0.017, p < .001), and is lower in studies with sampling versus census studies (β = −0.096, p < .001). This study reported that prevalence of CLU in Iranian students are lower than industrialized countries. In addition, gender, level of education, and methods of sampling are highly associated with changes in the prevalence of CLU across provinces.

Keywords

cannabis students meta-analyses meta-regression prevalence

Introduction

Cannabis (marijuana) is the most widely used substance in the world (Asbridge, Hayden, & Cartwright, 2012). Cannabis has been known to Homo sapiens in history and concern has been raised during the past centuries regarding its side effects (Anthony, 2012). Several serious side effects are enumerated for cannabis use such as suppression of rapid eye movement and diffuse slowing of background electroencephalographic activity (Greydanus, Hawver, Greydanus, & Merrick, 2013), chronic cough, sinusitis, conjunctivitis, pharyngitis, and bronchitis (Mallaret, Dal’Bo-Rohrer, & Dematteis, 2005). Increased mortality related to its side effects is reported in countries where the prevalence of cannabis use is high (Greydanus et al., 2013). Also, cannabis use has been associated with an increased risk of deadly motor vehicle crashes (Asbridge et al., 2012).

According to United Nations Office on Drugs and Crime, the majority of consumers of cannabis were young and the prevalence has been increasing in many countries (Meimandi, Nakhaee, Divsalar, & Dabiri, 2005). Meanwhile, the geographic distribution is very diverse. In Armenia, it is estimated that the prevalence is about 3.0% and in Czech Republic, 45.0% (Greydanus et al., 2013). The distribution in countries of Middle East and western part of Asia is not well studied.

It is established that two of the major cannabis production and distribution routes are in Pakistan and Afghanistan, implicating major problems for the countries around, especially for Iran (United Nations Office on Drugs and Crime, 2003). Studies in Iran have reported different prevalences across the country. In the north and northwest of the country, a lifetime prevalence of 22.2% (Mohammad Poorasl, Vahidi, Fakhari, Rostami, & Dastghiri, 2007) and a point prevalence of 3.6% (Najafi et al., 2007) have been reported.

No systematic review or meta-analyses have been conducted in Iran to estimate a combined prevalence of cannabis use according to different administrative divisions. In addition, the causes of changes in the prevalence across provinces in a country have not been studied. The aim of this study is to systematically review all cross-sectional studies conducted during the past 30 years in order to estimate cannabis lifetime use (CLU) pooled prevalence in high school and college students of Iran as well as determining factors related to changes in prevalence across provinces using meta-regression model.

Materials and Method

Search Strategy

The search was conducted for cross-sectional studies or surveys that estimated the prevalence of CLU in each province of Iran that was published since 1979. All domestic scientific databases including Iranmedex, Scientific Information Database, Magiran, Irandoc, Medlib, and IranPsych, as well as international databases, including PubMed/Medline, ISI Web of Knowledge, and Google Scholar, relevant reference lists, and relevant journals were searched up to April 2014. Communication with authors and Iranian experts was done for any additional information or unpublished studies.

Two researchers independently scanned the titles of all retrieved citations, removed duplicates, and identified potentially relevant articles for inclusion. Abstracts from selected articles were then independently reviewed by two researchers for further relevance, and full text manuscripts retrieved when appropriate. In the case of disagreement, a third assessor acted as a mediator.

The search strategy was limited to the Persian and/or English. For the databases search, the combinations of keywords relating to CLU were used. These include the following: “Marijuana,” “Marijuana Smoking,” “Marijuana abuse,” “Hashish,” “Cannabis,” “Substance,” “Drug,” “Illicit drug,” “Adolescent,” “Youth,” “School students,” “high school students,” “student,” and “Iran.” The search strategy in PubMed was (Marijuana OR “Marijuana Smoking” OR “Marijuana abuse” OR Hashish OR Cannabis OR Substance OR Drug OR “Illicit drug”) AND (Adolescent OR Youth OR “School students”) AND (Students) AND (Iran). Also, this strategy was modified for each database. EndNote X4 software was employed to manage and scan citations.

Inclusion and Exclusion Criteria and Quality Assessment

The following exclusion criteria were applied: (a) irrelevant study design such as case-control or clinical trials; (b) inadequate reporting of results, that is, studies not reporting prevalence for CLU; (c) poor quality score (scores below five stars) based on the Newcastle−Ottawa scale (Wells et al., 2014; This scale assigns a maximum of ten stars to each study—five stars for representativeness of the sample, justification of sample size, nonrespondents rate, and ascertainment of the exposure; two stars for comparability [control for confounder and additional factors]; and three stars for ascertainment of the outcomes); (d) articles that were conducted before 1979, and (e) review articles or meta-analyses.

For inclusion into the systematic review, the studies had to meet the following criteria: (a) having clear definition of CLU; (b) having standardized diagnostic criteria for drug abuse/use using the International Classification of Diseases, Classification of Mental and Behavioral Disorders, or Diagnostic and Statistical Manual of Mental Disorders, or direct question in questionnaire and interview; (c) having prevalence proportion for CLU; and (d) place of study should be in Iran.

Data Extraction and Measures

Information on first author, year of study, method of sampling, sample size, average age, diagnostic instrument, diagnostic criteria, participation rate, geographical location, level of education, and prevalence during lifetime, 1 month later, 1 week later, and current use were extracted independently from every eligible study.

Average age categorized as ≥18 versus ≤18 years; level of education: college students and high school students; gender: male, female, and mixed; sampling methods: census, multistage, stratified, random, and haphazard sampling; geographical location: name of provinces. Other variables were considered as continuous in meta-regression analysis.

Statistical Analysis

Prevalence estimates were calculated using the variance stabilizing double arcsine transformation (Freeman & Tukey, 1950) because the use of the inverse variance weight is inappropriate when dealing with binary data with low prevalence. Also, with this method, studies with zero prevalence can be included in the meta-analyses and their weight considered in combined estimation. In addition, due to zero to very low prevalence of CLU in some studies, confidence intervals (CIs) were calculated using the Wilson method (Newcombe, 1998; Wilson, 1927). The presence of heterogeneity was assessed by Cochran’s Q statistic (p < .1), combined with I² statistic for estimates of inconsistency within the studies. A value of 0% indicates no observed heterogeneity while 100% indicates significant heterogeneity. The I² values above 75% were determined as significant heterogeneity warranting analysis with a random effect model to adjust for the observed variability (Higgins, Thompson, Deeks, & Altman, 2003). Potential source of heterogeneity was further explored through subgroup analyses and meta-regression model using likelihood ratio method. In the meta-regression model, variables were examined both in univariable and in a multiple variable models at a significance level of <.05. All variables were entered in the model (Model 2), because all factors in univariable model (Model 1) were significant. To avoid model instability, only those variables that were significant in Model 2 were entered in speared multiple regression model (Model 3). Egger’s test was employed to examine potential publication bias. All analyses were conducted using Stata version 11.2 (Stata Corp LP, College Station, TX) with “metan,” “metareg,” “metabias,” and “propcii” commands.

Results

Descriptive Statistics

In an electronic database search, 4,000 citations were retrieved, producing a total of 33 studies from 12 provinces of Iran (Ahmadi, Alishahi, & Alavi, 2004; Ahmadi & Hasani, 2003; Ahmadi & Javadpour, 2001; Ahmadi, Javadpour, & Pridmore, 2009; Ahmadi, Maharlooy, & Alishahi, 2004; Ahmadi & Yazdanfar, 2002; Attari, Asgary, Shahrokhi, Naderi, & Shariatirad, 2012; Dehghani, Zare, Dehghani, Sedghi, & Pormovahed, 2009; Farhadinasab, Allahverdipour, Bashirian, & Mahjoub, 2008; Ghavidel et al., 2012; Jodati, Shakurie, Nazari, & Raufie, 2007; Meimandi et al., 2005; Mohamadi, Shobeiri, & Mahgoob, 2012; Mohammad Poorasl et al., 2007; Mohammadkhani, 2012; Mohammadpoorasl et al., 2012; Mohtasham-Amiri, Jafari-Shakib, & Khalili-Moosavi, 2011; Momtazi & Rawson, 2010; Mortazavi Moghadam, Madarshahian, Tabiei, Pejmankhah, & Sadeghi, 2009; Mustafa et al., 2010; Najafi et al., 2007; Najafi, Zarrabi, Shirazi, Fekri, & Mohseni, 2009; Nakhaee, Ziaaddini, & Karimzadeh, 2009; Nazarzadeh, Bidel, & Carson, 2014; Sajjadi & Sajjadi, 2012; Sarajzadeh & Feizi, 2003; Shams Alizadeh, Moghadam, Mohsenpour, & Rostami Gooran, 2008; Talaei, Mokhber, Fayyazi Bordbar, Javanbakht, & Samari, 2008; Taremian, Bolhari, Peirovi, & Tabatabaei, 2005; Zardkhaneh et al., 2011; Zarrabi et al., 2009; Ziaaddini, Zare-zadeh, & Heshmati, 2006; Ziaaddini, Ziaaddini, & Nakhaee, 2013). The characteristics of each included study are reported in Tables 1 to 3 and Figure 1 outlining the details related to the selection process of studies. The studies included a total of 30,972 adolescents and youth (12,931 high school students and 18,041 college students). Fourteen articles reported prevalence in men (n = 12,788), nine included mixed gender samples (n = 9,177), and nine investigated women (n = 9,007). The average age of men (reported in 13 reports) and women (from 7 reports) were 19.6 and 20.8 years, respectively. Average age of the mixed samples was 20.3 years (from 27 studies). Overall, quality assessment of all studies produces medium quality.

Table 1.

Description and Characteristics of Included Studies (Only Male).

								Questionnaire			Prevalence (%)
First author	Year of study	Province	Age in years (M)	Students grade	Sampling methods	Sample size	Measurement tool	Validity (%) or DSM use	Reliability (%)	Response rate (%)	Current	1 week	1 month	Lifetime
Ahmadi, J.	2000	Fars	17.2	High school	Multistage	197	Q	DSM-IV	NM	95.0	3.0	—	—	7.6
Ahmadi, J.	2000	Fars	22.8	College	Multistage	59	Q	DSM-IV	NM	93.0	0.0	—	1.7	10.2
Ahmadi, J.	2000	Fars	24.3	College	Multistage	329	Q	DSM-IV	NM	90.0	4.5	—	—	20.9
Ziaaddini, H.	2001	Kerman	NM	High school	C-A	1,945	Q	NM	NM	94.8	37	11.1	9.3	11.1
Jodati, A. R.	2001	Tabriz	21.3	College	C-D	173	Q	NM	NM	79.0	—	—	—	6.0
Ahmadi, J.	2001	Fars	13.5	High school	Multistage	470	Q	DSM-IV	NM	NM	0.0	—	—	0.2
Mortazavi, G.	2003	Khorasan	Range = 20-24	College	Random	361	Q	NM	NM	87.0	—	—	—	1.9
Najafi, K.	2005	Guilan	Range = 15-20	High school	Random	1,041	Q	NM	66.0	99.0	—	—	—	3.6
Nakhaee, N.	2005	Kerman	NM	High school	Stratified	256	Q	NM	NM	NM	—	—	—	1.6
Poorasl, A.	2005	Tabriz	16.3	High school	Multistage	1,785	Q	90.0	85.0	NM	—	—	—	22.2
Mohamadi, N.	2008	Hamedan	20	College	C-D	430	Q	NM	95.0	86.0	—	—	—	4.1
Ahmadi, J.	2008	Fars	23.03	College	Multistage	150	Q	DSM-IV	NM	78.8	—	—	—	2.3
Zardkhaneh, S.	2011	Survey^a	Range = 18-35	College	Multistage	3,372	Q	74-90	NM	NM	—	—	—	3.0
Mohammadkhani, S.	2012	Survey^a	15.5	High school	Multistage	1,283	Q	90.0	NM	95.0	—	—	0.9	2.1
Nazarzadeh, M.	2012	Ilam	17.2	High school	Multistage	937	Q	90.0	85.0	93.7	—	—	—	3.3

Note. NM = not mentioned; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, fourth edition; C-D = census of dormitory; C-A = census of all student in the city; Q = questionnaire; dash (—) in prevalence column indicates not mentioned.

Survey includes five universities from whole of Iran: Tehran University, Razi University of Kermanshah, Ferdowsi University of Mashhad, Isfahan University of Technology, and Shahid Bahonar University of Kerman.

Table 2.

Description and Characteristics of Included Studies (Only Female).

								Questionnaire			Prevalence (%)
First author	Year of study	Province	Age in years (M)	Students grade	Sampling methods	Sample size	Measurement tool	Validity (%) or DSM	Reliability (%)	Response rate (%)	Current	1 week	1 month	Lifetime
Ahmadi, J.	2000	Fars	15	High school	Multistage	200	Q	DSM-IV	NM	94.0	0.0	—	—	0.0
Ahmadi, J.	2000	Fars	20	College	Multistage	341	Q	DSM-IV	NM	93.0	0.3	—	—	0.9
Ahmadi, J.	2000	Fars	22	College	Multistage	172	Q	DSM-IV	NM	90.0	—	—	—	2.3
Ziaaddini, H.	2001	Kerman	NM	High school	C-A	1373	Q	NM	NM	95.0	15.4	23.1	7.7	23.1
Mortazavi, G.	2003	Khorasan	22	College	Random	509	Q	NM	NM	87.0	—	—	—	0.4
Najafi, K.	2005	Guilan	18	High school	Random	886	Q	NM	66.0	99.0	—	—	—	0.2
Nakhaee, N.	2005	Kerman	NM	High school	Stratified	396	Q	NM	NM	NM	—	—	—	2.6
Ahmadi, J.	2008	Fars	23	College	Multistage	150	Q	DSM-IV	NM	79.0	1.2	—	—	4.7
Zardkhaneh, S.	2011	Survey^a	27	College	Multistage	4,980	Q	74.0-90.0	NM	NM	—	—	—	0.4

Note. NM = not mentioned; DSM = Diagnostic and Statistical Manual of Mental Disorders, fourth edition; C-A = census of all student in the city; Q = questionnaire; dash (—) in prevalence column indicates not mentioned.

Survey include five universities from whole of Iran: Tehran University, Razi University of Kermanshah, Ferdosi University of Mashhad, Isfahan University of Technology, and Shahid Bahonar University of Kerman.

Table 3.

Description and Characteristics of Included Studies (Studies in Which Men and Women Were Not Separated).

								Questionnaire			Prevalence (%)
First author	Year of study	Province	Age in year (M)	Students Grade	Sampling methods	Sample size	Measurement tool	Validity (%) or DSM use	Reliability (%)	Response rate (%)	Current	1 week	1 month	Lifetime
Ahmadi, J.	2000	Fars	24	College	Multistage	346	Q	DSM-IV	NM	73.0	—	—	—	4.6
Mortazavi, G.	2003	Khorasan	22	College	Multistage	509	Q	NM	NM	87.0	—	—	—	0.4
Mohtasham-Amiri, Z.	2005	Guilan	22	College	Multistage	1,800	Q	NM	NM	94.0	—	—	—	2.4
Attari, M. A.	2006	Esfahan	17	High school	Multistage	537	ELISA	NM	NM	NM	—	—	—	3.2
Zarrabi, H.	2006	Guilan	22	College	Haphazard	827	Q	NM	NM	98.0	—	—	0.4	2.8
Taremian, F.	2006	Tehran	NM	College	Multistage	2,999	Q	NM	NM	99.0	—	—	0.8	2.2
Sajjadi, S. S.	2008	Fars	NM	High school	Multistage	1,225	Q	NM	74.0-83.0	NM	—	—	—	1.4
Ghavidel, N.	2008	Alborz	17	High school	Multistage	400	Q	NM	72.0	93.0	—	—	0.3	0.3
Dehghani, K.	2009	Yazd	22	College	Multistage	534	Q	NM	86.0	99.0	—	—	—	0.2

Note. NM = not mentioned; Q = questionnaire; DSM = Diagnostic and Statistical Manual of Mental Disorders; ELISA = Enzyme–Linked Immunosorbent Assay; em dash (—) in prevalence column indicates not mentioned.

Figure 1.

Flowchart of the reviewing process for prevalence of cannabis lifetime use in Iran.

Pooled Prevalence and Subgroup Analysis

A potential publication bias was detected for CLU (Egger’s test β₀ = −1.4; p < .001). CLU was reported in 33 surveys with a random effects pooled prevalence of 4.0% (95% CI = 3.0% to 5.0%; Figure 2). Estimates ranged from 0.0% to 23.0% with substantial heterogeneity (χ² = 1,424.5, p < .001, I² = 97.8%). Higher prevalences were identified in smaller studies with smaller weight (Figure 3). Moreover, subgroup analyses based on gender produce evidence of significant difference between three subgroups (Figure 4). In mixed, male, and female subgroups, prevalence were 2.0% (95% CI = 1.0% to 3.0%), 6.0% (95% CI = 4.0% to 9.0%), and 3.0% (95% CI = 2.0% to 6.0%), respectively. Further exploration according to level of education and method of sampling were conducted. For high school and college students subgroups prevalences were 5.0% (95% CI = 3.0% to 7.0%), and 2.0% (95% CI = 2.0% to 3.0%), respectively (Figure 2) and in multistage, random, stratified sampling and census, pooled prevalences were 4.0% (95% CI = 3.0% to 5.0%), 1.0% (95% CI = 0.0% to 3.0%), 2.0% (95% CI = 0.0% to 4.0%), and 11.0% (95% CI = 3.0% to 19.0%), respectively (Figure 5).

Figure 2.

Forest plot of studies related to prevalence of cannabis lifetime use in young Iranian population. Data are presented separately for college students and high school students.

Figure 3.

Forest plot of studies related to prevalence of cannabis lifetime use in young Iranian population. Studies are sorted by increasing weight in a random effects model.

Figure 4.

Forest plot of studies related to prevalence of cannabis lifetime use in young Iranian population. Data are presented separately for male, female, and mixed.

Figure 5.

Forest plot of studies related to prevalence of cannabis lifetime use in young Iranian population. Data are presented separately for sampling methods.

Current, last week, and last month prevalences reported only in few surveys, consequently pooled prevalences were not calculated for them. (Tables 1 -3).

Meta-Regression Model

In univariable meta-regression analyses, newer studies had higher prevalences of CLU (β for study date [year] = 0.0008, p < .001), and where the average age was ≥18 years (vs. ≤18), the prevalence was higher (β = 0.007, p < .001). The college versus high school students had higher prevalence (β = 0.008, p < .001). Also in category of ≥1,000 sample size versus <1,000, prevalence was higher (β = 0.009, p < .001), but where gender was male versus female and mixed, and in multistage, stratified, and random sampling versus census, prevalences were lower (β = −0.003, p < .001; β = −0.111, p < .001), respectively (Table 4, Model 1). In a multiple variable meta-regression model, all of the variables were significant except average age (Table 4, Model 2). Consequently, average age was excluded from the Model 3. Model 3 has identified that the prevalence is significantly higher in college students (β = 0.089, p < .001), and in male gender (β = 0.017, p < .001), and is lower in studies with sampling versus census studies (β = −0.096, p < .001).

Table 4.

Meta-Regression Analysis^a of Cannabis Prevalence Heterogeneity Suspected Variables.

	Model 1^b			Model 2^c			Model 3^d
Variables	β	SE	p	β	SE	P	β	SE	p
Study date (year)	0.0008	0.00006	<.001	−0.0003	0.0001	.04	−0.0002	0.0001	.17
Average age (≥18 vs. ≤18 years Ref)	0.007	0.0006	<.001	−0.008	0.006	.14	—	—	—
Level of education (College students vs. high school students Ref)	0.008	0.0006	<.001	0.017	0.006	.005	0.008	0.001	<.001
Gender (Male vs. female and mixed Ref)	−0.003	0.0008	<.001	0.019	0.001	<.001	0.017	0.001	<.001
Sampling methods (Multistage, stratified, and random sampling vs. census Ref)	−0.111	0.004	<.001	−0.019	0.008	.02	−0.096	0.005	< .001
Sample size	0.009	0.001	<.001	0.009	0.001	<.001	0.009	0.001	.09

Note. Significant associations (p ≤ 0.05) are indicated in bold typeface; Ref = reference group.SE = standard error

Likelihood ratio method.

Univariable model.

Multiple variable model.

Variables entered into the model if p ≤ 0.05.

Discussion

This systematic review of CLU prevalence in Iranian youth identified 33 surveys including 30,972 subjects. There are four main findings: (a) prevalence of CLU in young Iranian people was 4.0%, (b) higher prevalence in male compared with female and in college students versus high school students, (c) higher prevalence in census based studies versus sampling based studies, (d) little changes to prevalence during past 30 years.

Much research has been conducted to confirm high use of cannabis among adolescents and young adults across communities. The 2007 European School Survey Project on Alcohol and other Drugs had reported that CLU among students aged 15 to 16 years in Europe is 19.0% in average (among 35 countries) and ranged from 3.0% in Armenia to 45.0% in the Czech Republic (Greydanus et al., 2013). In addition, in Canadian college students, the reported prevalence of cannabis abuse was approximately 30.0% (Fischer et al., 2013). According to the United States Centers for Disease and Prevention’s Youth Risk Behavioral Surveillance reports, CLU prevalence among high school students was 31.3% in 1991, 47.2% in 1999, and 36.8% in 2009 (Eaton et al., 2010). Compared with these international reports, the prevalence of CLU was considerably less in Iran. A primary explanation for this small prevalence might be legal limitations of cannabis and other substances in Iran. In addition, because of the historical background of opium and derivatives use, the prevalence of opium use is higher than other substances such as cannabis in Iran (Siassi & Fozouni, 1980). On the other hand, most of the included studies in these meta-analyses assessed CLU using questionnaires or direct questions. Consequently, false negative answers can be considered as potential low value for pooled prevalence.

This study’s finding that the sampling method of studies was significantly associated with prevalences suggests that researchers need to interpret their cross-sectional findings with regard to their methodology specifically with regard to sampling method for which we reported some underestimation. One of the interesting results of this study were the differences in subgroup analyses and meta-regression results on prevalence between high school and college students. Subgroup analysis indicated that prevalence in high school students is higher than for college students but the meta-regression model had indicated contrast result. It seems possible that these contrasts are due to the confounding role of other variables that were not considered in subgroup analysis. After adjustment for all possible confounders in meta-regression, it was revealed that the prevalence is significantly higher in college students versus high school students. Consequently, the results for subgroup analysis in meta-analyses studies should be interpreted with consideration of other possible confounders. One finding of this review is that future research in prevalence of any drug should provide clear information about study questionnaires and their validity and reliability.

This review is important for enhancing our methodological knowledge of the epidemiology of CLU in adolescents and young adults and provides a more definitive statement on the direction of prevalence changes. Also, it provides better estimation for prevalences in national studies such as the burden of diseases.

This study does have several limitations, as all pooled analyses contained significant heterogeneity and subsequently should be interpreted with caution. Most studies failed to report questionnaire validity, reliability, and response rate. These characteristics may be a significant source of heterogeneity (Fazel, Khosla, Doll, & Geddes, 2008). Moreover, a lack of comprehensive coverage for gray literature containing university databases and research projects, further limit the search.

In conclusion, the results of these meta-analyses indicate that prevalence of CLU in Iranian high school and college students are lower than for industrialized countries. In addition, gender, level of education, and methods of sampling are highly associated with variation in estimation of CLU prevalence.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was financially supported by the Ilam University of Medical Sciences, and a research grant (Grant No. 909230) from the Prevention of Psychosocial Injuries Research Center.

References

Ahmadi

Alishahi

Alavi

(2004). Substance use disorders in a sample of Iranian secondary school students. Social Indicators Research, 65, 355-360. doi:10.1023/B:SOCI.0000003547.60431.d0

Ahmadi

Hasani

(2003). Prevalence of substance use among Iranian high school students. Addictive Behaviors, 28, 375-379.

Ahmadi

Javadpour

(2001). Assessing substance use among Iranian healthcare students. Journal of Substance Use, 6, 196-198. doi:10.1080/14659890152558813

Ahmadi

Javadpour

Pridmore

(2009). Epidemiology of substance use among dentistry students. Iranian Journal of Psychiatry and Behavioral Sciences, 3, 25-28.

Ahmadi

Maharlooy

Alishahi

(2004). Substance abuse: Prevalence in a sample of nursing students. Journal of Clinical Nursing, 13, 60-64.

Ahmadi

Yazdanfar

(2002). Current substance abuse among Iranian university students. Addictive Disorders & Their Treatment, 1, 61-64.

Anthony

J. C.

(2012). Steppingstone and gateway ideas: A discussion of origins, research challenges, and promising lines of research for the future. Drug and Alcohol Dependence, 123(Suppl. 1), S99-S104. doi:10.1016/j.drugalcdep.2012.04.006

Asbridge

Hayden

J. A.

Cartwright

J. L.

(2012). Acute cannabis consumption and motor vehicle collision risk: Systematic review of observational studies and meta-analysis. British Medical Journal, 344, e536. doi:10.1136/bmj.e536bmj.e536

Attari

M. A.

Asgary

Shahrokhi

Naderi

G. A.

Shariatirad

(2012). Cannabis and opium abuse patterns and their associated complications in a sample of young Iranians. Journal of Child & Adolescent Substance Abuse, 21, 238-246. doi:10.1080/1067828x.2012.689806

10.

Dehghani

Zare

Dehghani

Sedghi

Pormovahed

(2009). Prevalence and related factors of substances abuse among students of Shahid Sadoughi University of Medical Sciences. Journal of Shahid Sadoughi University of Medical Sciences, 18, 164-169.

11.

Eaton

D. K.

Kann

Kinchen

Shanklin

Flint

K. H.

Hawkins

. . . Wechsler

(2010). Youth risk behavior surveillance (YRBS) US 2009 (Vol. 59, No. SS-5). Retrieved from http://www.cdc.gov/mmwr/pdf/ss/ss5905.pdf

12.

Farhadinasab

Allahverdipour

Bashirian

Mahjoub

(2008). Lifetime pattern of substance abuse, parental support, religiosity, and locus of control in adolescent and young male users. Iranian Journal of Public Health, 37, 88-95.

13.

Fazel

Khosla

Doll

Geddes

(2008). The prevalence of mental disorders among the homeless in western countries: Systematic review and meta-regression analysis. PLoS Medicine, 5, e225. doi:10.1371/journal.pmed.005022508

14.

Fischer

Dawe

McGuire

Shuper

P. A.

Capler

Bilsker

. . . Rehm

(2013). Feasibility and impact of brief interventions for frequent cannabis users in Canada. Journal of Substance Abuse Treatment, 44(1), 132-138. doi:10.1016/j.jsat.2012.03.006

15.

Freeman

Tukey

(1950). Transformations related to the angular and the square root. Annals of Mathematical Statistics, 21, 607-611.

16.

Ghavidel

Samadi

Kharmanbiz

Asadi

Feyzi

Ahmadi

. . . Fatemeh

(2012). Investigation of substance use prevalence and the interrelated factors involved through third-year high school students in Nazarabad city from January 2008 to June 2008. Razi Journal of Medical Sciences, 19(97), 28-36.

17.

Greydanus

D. E.

Hawver

E. K.

Greydanus

M. M.

Merrick

(2013). Marijuana: Current concepts. Frontiers in Public Health, 1, 42. doi:10.3389/fpubh.2013.00042

18.

Higgins

Thompson

Deeks

Altman

(2003). Measuring inconsistency in meta-analyses. British Medical Journal, 327, 557-560.

19.

Jodati

A. R.

Shakurie

S. K.

Nazari

Raufie

M. B.

(2007). Students’ attitudes and practices towards drug and alcohol use at Tabriz University of Medical Sciences. Eastern Mediterranean Health Journal, 13, 967-971.

20.

Mallaret

Dal’Bo-Rohrer

Dematteis

(2005). Adverse effects of marijuana. La Revue du Praticien, 55(1), 41-49.

21.

Meimandi

M. S.

Nakhaee

Divsalar

Dabiri

(2005). Estimating the prevalence of cannabinoid use urine testing: A preliminary study in Kerman, Iran. Addictive Behaviors, 30, 1464-1467. doi:10.1016/j.addbeh.2005.01.016

22.

Mohamadi

Shobeiri

Mahgoob

(2012). The prevalence of substance abuse in students of Hamadan University of Medical Sciences. Zahedan Journal of Research in Medical Sciences, 14, 86-89.

23.

Mohammad Poorasl

Vahidi

Fakhari

Rostami

Dastghiri

(2007). Substance abuse in Iranian high school students. Addictive Behaviors, 32, 622-627. doi:10.1016/j.addbeh.2006.05.008

24.

Mohammadkhani

(2012). Prevalence of cigarette smoking, alcohol drinking and illegal drugs use among Iranian adolescents. Journal of Kerman University of Medical Sciences, 19(1), 32-48.

25.

Mohammadpoorasl

Nedjat

Fakhari

Yazdani

Foroushani

A. R.

Fotouhi

(2012). Substance abuse in high school students in association with socio-demographic variables in northwest of Iran. Iranian Journal of Public Health, 41(12), 40-46.

26.

Mohtasham-Amiri

Jafari-Shakib

Khalili-Moosavi

(2011). Prevalence and factors associated with ecstasy use among college undergraduates in north of Iran—2005. Asian Journal of Psychiatry, 4, 31-34. doi:10.1016/j.ajp.2011.01.004

27.

Momtazi

Rawson

(2010). Substance abuse among Iranian high school students. Current Opinion in Psychiatry, 23, 221-226. doi:10.1097/YCO.0b013e328338630d

28.

Mortazavi Moghadam

G. R.

Madarshahian

Tabiei

Pejmankhah

Sadeghi

M. H.

(2009). Prevalence of drug abuse among the student of Birjand Universities (2003). Modern Care Journal, 6, 60-64.

29.

Mustafa

Borujerdi

Motalebi

Asheri

Abbasinejad

Motamedi

(2010). The prevalence of cigarette smoking, alcohol consumption, psychostimulant and cannabinoid drugs abuse among 15 to 35 years old Tehranis. Iranian Journal of Psychiatry and Behavioral Sciences, 4(2), 26-30.

30.

Najafi

Fekri

Mohseni

Zarabi

Najafi

Faghirpour

Shirazi

(2007). Survey of the prevalence of substances use among high school students in Guilan 2005-2006. Journal of Guilan University of Medical Science, 16(62), 67-79.

31.

Najafi

Zarrabi

Shirazi

Fekri

Mohseni

(2009). Prevalence of substance use among Iranian high school students in 2005-2006. Kuwait Medical Journal, 41, 20-25.

32.

Nakhaee

Ziaaddini

Karimzadeh

(2009). Epidemiologic study on drug abuse among first and second grade high school students in Kerman. Addiction and Health, 1(1), 31-38.

33.

Nazarzadeh

Bidel

Carson

K. V.

(2014). The association between tramadol hydrochloride misuse and other substances use in an adolescent population: Phase I of a prospective survey. Addictive Behaviors, 39, 333-337. doi:10.1016/j.addbeh.2013.09.013

34.

Newcombe

R. G.

(1998). Two-sided confidence intervals for the single proportion: Comparison of seven methods. Statistics in Medicine, 17, 857-872.

35.

Sajjadi

(2012). A study on prevalence of drug usage among teenagers in the Lamerd and Mohr cities. Asian Journal of Medical and Pharmaceutical Researches, 2, 01-04.

36.

Sarajzadeh

S. H.

Feizi

(2003). Drug and alcohol use among college students. Social Welfare, 6(25), 86-110.

37.

Shams Alizadeh

Moghadam

Mohsenpour

Rostami Gooran

(2008). Prevalence of substance abuse in medical students of Kurdistan University. Scientific Journal of Kurdistan University of Medical Sciences, 13(2), 18-26.

38.

Siassi

Fozouni

(1980). Distribution of opium coupons to addicts in Iran: Policies and problems. Chemical Dependencies, 4, 127-133.

39.

Talaei

Mokhber

Fayyazi Bordbar

M. R.

Javanbakht

Samari

A. A.

(2008). Patterns and correlates of substance use among university students in Iran. Iranian Journal of Psychiatry and Behavioral Sciences, 2(2), 15.

40.

Taremian

Bolhari

Peirovi

Tabatabaei

(2005). The prevalence of drug abuse among students of Tehran city. Iranian Journal of Psychiatry and Clinical Psychology, 13, 335-342.

41.

United Nations Office on Drugs and Crime. (2003). Global illicit drug trends: Chapter 1. Trends: Consumption cannabis. Retrieved from https://www.unodc.org/pdf/trends2003_www_E.pdf

42.

Wells

G. A.

Shea

O’Connell

Peterson

Welch

Losos

Tugwell

(2014). The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Retrieved from http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp

43.

Wilson

(1927). Probable inference, the law of succession, and statistical inference. Journal of the American Statistical Association, 29, 209-212.

44.

Zardkhaneh

Sohrabi

Taraghija

Porsharifi

Taremian

Peyrovi

Falsafinejad

Yaghubi

Ramzani

(2011). Demographic characteristics and substance use among students in national universities. Psychology and Education Studies, 7(12), 1-22.

45.

Zarrabi

Najafi

Shirazi

Borna

Sabahi

Nazifi

(2009). Prevalence of substance use among students of Guilan University of Medical Sciences 2005-2006. Journal of Qazvin University of Medical Sciences, 12(4), 69-74.

46.

Ziaaddini

Zare-zadeh

Heshmati

(2006). The prevalence rate of substance abuse and addiction and some relevant factors among junior and senior high school students in Kerman city (2000-2001). Journal of Kerman University of Medical Sciences, 2, 84-94.

47.

Ziaaddini

Nakhaee

(2013). Pattern and trend of substance abuse in eastern rural Iran: A household survey in a rural community. Journal of Addiction, 2013, 297378. doi:10.1155/2013/297378