Social Inequality and Substance Use and Problematic Gambling Among Adolescents and Young Adults: A Review of Epidemiological Surveys in Germany

Transcription

1 Send Orders for Reprints to 26 Current Drug Abuse Reviews, 2016, 9, Social Inequality and Substance Use and Problematic Gambling Among Adolescents and Young Adults: A Review of Epidemiological Surveys in Germany Dieter Henkel *,1 and Uwe Zemlin 2 1 Institute of Addiction Research, Frankfurt University of Applied Sciences, Frankfurt a.m., Germany 2 AHG Klinik Wilhelmsheim, Oppenweiler, Germany Abstract: The current review provides an overview of socioepidemiological research in Germany about the prevalence of addictive behaviours (smoking, binge and hazardous drinking, consumption of cannabis and other illegal drugs, the non-medical use of prescription drugs and problematic gambling) among adolescents (11-17 years) and young adults (18-25 years), also differentiating between different socioeconomic status (SES) indicators (attended school type, family affluence, parental occupational status, parental SES, employment status) and migration background. The authors evaluated data from ten national surveys and one regional survey conducted between 2002 and 2012, which included different samples. The trends over this time frame reveal that the proportion of adolescents who smoke tobacco, show problematic patterns of alcohol consumption, use cannabis or other illegal drugs has generally declined over the investigated time span in Germany. The results nevertheless suggest that some strong associations still exist between social inequalities and the prevalence of substance use. The detailed results are summarised in twelve tables. The main results are as follows: 1) Low SES (school type, employment status) was consistently associated with more cigarette smoking, and, where such data was available, this pattern was observed in both males and females. 2) With regard to family affluence, two surveys show that boys with low and middle FAS are significantly less likely to have binge drinking experience compared to boys with high FAS. There were no significant associations between problematic alcohol use and parental SES, and not all results of the surveys show that binge drinking is more prevalent among HS-students. Employment status was associated with gender differences; problematic patterns of alcohol consumption were significantly more prevalent among young unemployed males compared to GY-students (secondary high school/grammar school) of the same age. The opposite was true for females. Another consistent finding was that among adolescents and young adults with a Turkish/Asian migration background, a problematic use of alcohol was significantly less common compared to adolescents and young adults of the same age without a migration background. 3) In terms of the consumption of cannabis, the unemployed and students with low educational level (`Hauptschule ) emerge as high-risk groups. 4) The results concerning problematic gambling significantly overlap with and reflect the findings of international research: being male, of low educational attainment, unemployed, receiving social welfare, and having a migration background significantly increased the risk of problematic gambling habits. 5) The highest lifetime prevalence rates for the consumption of illegal drugs (other than cannabis) were observed among students with low educational level. It should be noted that other SES indicators, in addition to school type, have not been examined to date. The review concludes by outlining gaps and future research areas, as well as presenting several implications for prevention initiatives. Keywords: Social inequality, SES, tobacco, alcohol, cannabis, illegal drugs, prescription drugs, problematic gambling, adolescents, young adults, Germany. 1. INTRODUCTION Social inequality is a multi-faceted phenomenon. It characterizes the unequal distribution of property and power, income and wealth, education and qualification, participation in the labour market, access to health provisions and social security, as well as opportunities to partake in social and cultural activities. The specific social situation of each individual is furthermore determined by his status as native resident or resident with a migration background. *Address correspondence to this author at the Ebertsbronn 31, D Niederstetten, Germany, Tel: (49) prof.dieter.henkel@t-online.de Social epidemiology aims to document the prevalence of addictive behaviours in different social groups. The generated insight provides important and necessary insights for targeted drug prevention activities. The international research about adolescent and young adults has focused on a specific number of factors that contribute to social inquality: socioeconomic status (SES: income, education, occupation), employment status (employed/unemployed), material deprivation/poverty, and migrationstatus. Hanson and Chen [1] produced a review summarizing the findings from 97 epidemiological studies about tobacco, alcohol and cannabis consumption in Western countries. These authors concluded that adolescents with low SES (e.g /16 $ Bentham Science Publishers

2 Social Inequality and Substance Use and Problematic Gambling Current Drug Abuse Reviews, 2016, Vol. 9, No family income, parental education or occupation) are much more likely to smoke than high SES adolescents. However, country-specific differences play an important role. For example, SES was a stronger predictor of smoking in northern European countries. However, smoking and SES were not significantly related in southern European countries [2, 3]. In terms of alcohol and cannabis consumption, Hanson and Chen [1] concluded that SES is not significantly linked to alcohol and cannabis consumptions during adolescence. However, problematic patterns of consumption (e.g. binge drinking, hazardous drinking) were not the focus of this research. In contrast, the meta-analysis by Lemstra and colleagues [4], summarizing findings across a sample of 200,000 adolescents in the USA, Great Britain, New Zealand, Finland, and Italy, did observe a significantly higher prevalence rate for both cannabis and alcohol risk behaviour among adolescents with low SES (parental income, parental education and occupation) in comparison to adolescents with higher SES. However, it is unclear how cannabis and alcohol risk behaviours were conceptualized. Evidence from the European literature about binge drinking shows that this pattern of alcohol use is much more common among adolescents with a low educational level, particularly in the nothern and middle parts of Europe compared to the southern parts [5]. A review of the international literature (e.g. USA, New Zealand, France, European countries) [6] concluded the following with regard to employment status: Unemployed adolescents and young adults were found to have a significantly higher rate of substance use and abuse compared to their employed counterparts. Problematic behaviours include binge and hazardous drinking, illegal drug use, cannabis dependence and smoking. Furthermore, based on twenty longitudinal studies the review [6] demonstrates that job loss/unemployment is a risk factor for young adults to develop substance use problems and substance use disorders (smoking, alcohol and cannabis use). For example, a longitudinal study from New Zealand showed that long-term unemployment significantly increases the risk of alcohol and drug abuse/dependence [76]. Another study provided further support for this as well, showing that low socioeconomic status in childhood exacerbates the risk for heavy episodic drinking and smoking [7]. And finally, a representative study covering different European countries (N=4,695) further revealed that unemployment was a discrete predictor of binge drinking among adolescents, even when controlling for age, gender and education [47]. Social and material deprivation tends to be multidimensional; various factors come into play. These include low educational status, poor employment status (precarious work, unemployment), low income/living on social welfare, precarious housing, and being socially isolated. Representative US studies link deprivation and consumption of various substances. Baumann et al. [8] found that increasing levels of deprivation made it more likey that individuals smoked tobacco, abused alcohol and used psychotropic drugs (prescribed or non-prescribed). Results from a longitudinal study show that when adolescents experienced poverty over extended periods of time during their adolescence, they were more likely to drink heavily [9]. Fettes et al. [10] further observed that substance use was higher among youths involved in some welfare scheme, specifically in terms of cannabis use over the life span and other illegal drugs. Problematic gambling was the focus of a comprehensive review of the international literature [11]. This review considers a number of sociodemographic factors by critically evaluating 27 empirical studies. The authors identified several potential risk factors for problematic gambling: being an adolescent or young adult, being male, being unemployed, receiving social welfare, having a low level of academic achievement, and having a migration background (e.g. in US studies: African-American, Hispanic or Asian background). The epidemiological research has consistently shown that substance use rates of migrant or ethnic minority groups often differ significantly in two ways. First, the prevalence rates between these groups and majority ethnic groups tend to be quite different. And second, prevalence rates also differ significantly between the migrant and ethnic groups as well. This also applies to adolescents and young adults [see, for example, 12 and 13]. As a result, factors such as migration and ethnic background represent important sociodemographic variables to be considered in ths research as they provide a complete and sufficiently differentiated picture about the type of addictive problems that are experienced by different groups in the population. Germany is one of the richest countries in the world. Nevertheless, a number of serious social problems exist. According to recent figures, around 2 million resp. 18% of all adolescents under the age of 18 are living in povertystricken families. In addition, around (6%) young adults between 18 and 25 years of age are currently unemployed. And 5.6 million individuals in this age group have a migration background (came to Germany as a child or have at least one parent who move to Germany from abroad) [14]. These circumstances highlight the importance of researching the relationship between social unequality and addictive behaviours among different groups in Germany. 2. AIMS The aim of the present review is to provide an overview of current socioepidemiological research in Germany, focusing on two questions in particular: (1) To what extent are the prevalence rates of addictive behaviours (such as tobacco smoking, the problematic consumption of alcohol, cannabis and other illegal drugs, the non-medical use of prescription drugs as well as problematic gambling) among adolescents and young adults linked to indicators of social inequality? (2) To what extent have social inequalities on substance use changed (increased or decreased) over time since 2003? To date, no other review has summarised the empirical evidence from Germany in order to address these questions on the basis of different representative surveys. In addition, the current review will specify research gaps and summarise implications for future prevention programmes. 3. METHODS The authors conducted an extensive literature search between January and March 2015 using the following electronic Databases: Pubmed; PsychIndex, Google, Google

3 28 Current Drug Abuse Reviews, 2016, Vol. 9, No. 1 Henkel and Zemlin scholar. The search covered all publications between 2000 and The keywords and related search terms were as follows: Germany/German, survey, adolescents, young adults, social inequality, socioeconomic status (SES), educational level, family income/affluence, migration background, employment status, smoking/tobacco, binge drinking, heavy/hazardous alcohol consumption, illegal drugs, cannabis, prescription drugs, problematic gambling, health behaviour, prevalence. Additional papers and surveys were identified via citations in other papers. This process was repeated until no additional publications could be located. The final selection included some papers that had not been peer-reviewed. Most of the prevalence rates reported in German surveys (Tables 2-13) were obtained from scientific research reports produced and published by research institutions rather than peer-reviewed articles. Only publications in German or English were considered. The present review draws on studies that met the following criteria: (a) The studies were conducted in Germany between 2000 and 2015; (b) they included nationally representative population-based cross-sectional samples; (c) the samples included participants between the ages of 11 to 17 (adolescents) or 18 to 25 (young adults); (d) the surveys measured prevalence rates of smoking, alcohol consumption, illegal drug/cannabis use, non-medical use of prescription drugs and problematic gambling; (e) studies on alcohol were limited to those that revealed problematic patterns of use (binge or hazardous drinking); (f) the studies included at least one indicator of social inequality (SES). Using these criteria, 10 surveys with different samples were identified. 4. SURVEYS The chosen epidemiological studies included the following surveys; these can be grouped into four categories (Table 1). 1) HBSC-surveys (`Health Behavior in School-aged Children, University of Bielefeld, Another representative HBSC-survey was conducted for the state of North Rhine-Westphalia (17.6 million inhabitants) to examine problematic gambling. This HBSC-PG-survey was also included in the present review even though it was not based on a nationally representative sample because the survey included information that was not included in other surveys. 2) DA-surveys (`Drug Affinity Studies ) from the Federal Centre for Health Promotion (BZgA, Cologne, Table 1. German socio-epidemiological surveys: sample size, age, SES-indicators/migration background, addictive behaviours. Surveys N Age SES-Indicators, Migration Background HBSC , school type family affluence migration background Addictive Behaviours S, BD S, BD, C S, BD HBSC ,274 2, school type parental occupational status family affluence S, BD, C S, BD, C S, BD, C HBSC , school type parental occupational status family affluence HBSC-PG , school type migration background S, BD, C S, BD, C S, BD, C PG PG DA ,942 3, school type migration background employment status S, BD, HC, C BD, HC, C S, BD, HC, C DA ,686 4, school type employment status migration background S, BD, HC S, BD, HC BD, HC KiGGS , school type parental SES migration background S, C S, C S, C KiGGS , parental SES S, BD, HC ESPAD , school type S, BD, HC, C, ID, PD, PG ESPAD , school type S, BD, HC, C, ID, PD ESPAD , school type S, BD, HC, C, IG, PD Note: S smoking, BD binge drinking, HC hazardous alcohol consumption, C cannabis consumption, ID illegal drug use except cannabis, PD non-medical use of prescription drugs, PG problematic gambling.

4 Social Inequality and Substance Use and Problematic Gambling Current Drug Abuse Reviews, 2016, Vol. 9, No ) ESPAD-surveys (`The European School Survey Project on Alcohol and other Drugs`, Institute for Therapy Research IFT, Munich, 4) KiGGS-Surveys 2006 and 2012 (`Children and Adolescents Health Survey, Robert Koch-Institut RKI, Berlin, Definitions of Substance Use and Problematic Gambling Overall, the surveys examined the following specific characteristics of substance use and problematic gambling (assessed via self-report in all surveys): 1) Regular smoking, daily smoking, heavy smoking (>10/>20 cigarettes per day), age at which individuals first started smoking, electronic cigarette smoking (ecigarettes with liquids containing nicotine) and second-hand smoking among non-smokers (based on individuals spending time at least several times a week in facilities where others smoke). 2) In terms of alcohol consumption, the present review considered two patterns of consumption which can be viewed as clearly problematic: (a) binge drinking (heavy episodic drinking) was examined in all surveys (except the data collected by KiGGS-survey 2006) using the same international standard [5] to define binge drinking as behaviour that leads to intoxication due to the consumption of five (or six, KiGGs-survey 2012) or more units in one sitting; (b) hazardous consumption of alcohol (alcohol consumption that leads to long-term health risks) as indicated by an average daily intake of more than 12g/20g for girls or 24g/30g of pure alcohol for boys (according to DA-surveys/ESPAD-surveys). The authors of the KiGGS-survey 2012 [8] assessed hazardous consumption by using the internationally established brief screening test AUDIT-C [15]. 3) In terms of cannabis consumption, the frequency of consumption was examined in terms of three measures: prevalence over the lifespan, use during the last 12 months and/or the last 30 days. The 2007 and 2011 ESPAD-surveys also included the six-item Cannabis Abuse Screening Test (CAST) by Legleye et al. [16]. Based on this measure, when individuals score four or more out of six points (one for each question), their cannabis consumption is considered problematic as it may result in addiction. Problems listed in the CAST include: (a) consuming cannabis on one s own, (b) consuming cannabis in the morning, (c) experiencing problems due to consuming cannabis, (d) particulary memory-related problems, (e) advice from friends or family to reduce the consumption of cannabis, and (f) trying and failing to reduce one s consumption. 4) Problematic gambling always involves financial stakes and risks being taken. The 2002 HBSC-PGsurvey defined adolescents as problematic gamblers if their behaviour met the criteria outlined in the DSM- IV-MR-J according to Fisher [17]. These criteria were adapted to reflect gambling-specific addiction, as outlined by the DSM-IV [18]. The 2011 ESPADsurvey recorded only the frequencies with which participants engaged in different types of gambling activities during the last 12 months (e.g., card games, gambling machines, sport bets). 5) The ESPAD-surveys also asked students as to whether or not they had ever consumed any illegal drugs (not including cannabis), such as amphetamines, ecstasy, LSD, cocaine, crack, heroin and psychotropic mushrooms. The present review describes the frequency of consumption of any of these drugs over the lifespan. 6) The ESPAD-surveys also asked pupils about their non-medical use of prescription drugs, specifically, whether they had ever consumed tranquilizer/sedatives Indicators of Social Inequality/SES The current surveys consider different criteria of social inequality. These criteria are labelled indicators of socioeconomic status (SES) in the present review. These indicators were self-reported and include, in addition to school type (as most surveys were conducted in school settings), the following variables: 1) Family affluence, as measured by the Family Affluence Scale (FAS) [19]. The FAS consists of four different items: a) family ownership of one or more automobiles, b) number of family holidays taken over the last 12 months, c) separate bedrooms for children and adolescents (yes/no) and d) number of computers in the home. In the HBSC-surveys, the FAS was used to differentiate between low, middle and high levels of family affluence [2]. There were several studies of FAS that suggest acceptable validity [20-22]. Furthermore, there is strong consistency in the associations found between FAS and health outcomes across countries and between survey cycles [2]. 2) Type of school attended at the secondary level. This is often seen as an indicator of the students current and future SES [23]. The German school system comprises four different school types: a) `Hauptschule (abbreviated as HS) = low educational level, b) `Realschule (RS) = middle level, c) `Gymnasium (GY) (secondary high school/grammar school) = high level; d) comprehensive schools (`Gesamtschule ) were also included in the surveys. However, the latter do not fit clearly in with the hierarchical structure of education associated with HS-, RS- and GY-schools. As a result, this school type was not included in the present review. 3) Parental occupational status. 4) Parental socioeconomic status, measured by income, education and occupation. 5) Employment status (employed/unemployed). 6) Migration background. The surveys also asked adolescents whether or not one or both of their parents had emigrated from another country/region to Germany (e.g., Turkey/Asia) or, more generally, whether they had

5 30 Current Drug Abuse Reviews, 2016, Vol. 9, No. 1 Henkel and Zemlin any migration background (HBSC-surveys, KiGGSsurveys). It is important to be mindful of the social situation of migrants in Germany; a migration background does not only reflect cultural and ethnic differences. It is also associated with hierarchical social inequality. Low SES status is less prevalent (22% versus 54%) among native Germans than German residents with a migration background [23] Trend Data The ESPAD-surveys provide trend data for a time span of nine years ( ) that show how the prevalence of smoking, binge and hazardous drinking, cannabis use, illegal drug use (except cannabis), and non-medical use of prescription drugs changes over time. The survey data also allows to differentiate the results for different school types. Although the data was collected at different points, the same indicators of social inequality and also the same substance use measures were applied (in contrast to the HBSC- and KiGGS-surveys). 5. RESULTS Given that different substance use behaviours and problematic gambling may have different associations with the above mentioned indicators of social inequality, the relationships between substance use and SES indicators were examined separately for each individual behaviour, resulting in five sections on the consumption of tobacco, alcohol, cannabis, other substances, and problematic gambling Tobacco Germany is one of the countries where SES and smoking tend to be closely linked: 1) A consistent trend emerges, regardless of the period when the survey data was collected and which particular smoking characteristic were examined: Smoking rates are significantly higher among the unemployed and adolescents who attended schools with low educational level (HS). For example, the propensity to take up smoking tends to several times higher among HS-students (OR 4.6 for boys, 3.4 for girls) than among GS-students (reference OR 1.0), adjusted for age, migration background, and parental SES (KiGGs 2006, Table 2: OR). In contrast, no consistent relationship was observed between smoking, family affluence, and parental occupational status. Only the KiGGS-survey 2012 shows a significant relationship between parental SES and smoking (Table 2). 4) The proportion of heavy smokers (>10/>20 cigarettes per day) is significantly higher amongst adolescents with low SES (HS-students, unemployed) compared to GY-students (the reference group) of the same age (Tables 2 and 4). The prevalence rates of the employed young adults are also significantly higher, but considerably lower compared to the unemployed (DA-surveys 2010/2012, see also the results for daily smoking, Table 2). 5) Low SES adolescents started to smoke considerably earlier, for example, 32% of the HS-students started before they reached 12 years of age, compared with 24% of the RS-students and 14% of the GY-students [24]. 6) Unemployed young adults were significantly more likely to smoke e-cigarettes, although the prevalence rate of 4.5% observed for a time period of 30 days is low (Table 2). 7) Non-smoking adolescents with low SES (school type, parental SES) are significantly more often affected by second-hand smoking than GY-students (KiGGSsurvey 2006, Table 2). The prevalence rates for RS students tends to lie between the rate for HS- and GYstudents. 8) With regard to the KiGGS-survey 2006, adolescents with migration background were significantly less likely to smoke (OR 0.5, Table 2). But the 2010 HBSC findings show the opposite. A differentiated examination is not available due to the lack of information about countries or regions of origin. 9) Several analyses were conducted to assess whether or not the prevalence rates actually correlated with SES and the extent to which this relationship may or may not have been confounded by potential covariates such as gender and age (Table 2). These logistic regression analyses revealed that each of the two variables (school type, employment status) was consistently related to smoking, independently from the covariates mentioned above (OR, Table 2) Trends ( ) The proportion of smokers among adolescents in most Western European countries, including Germany, has significantly declined in the first decade of the 21st century [25]. According to the DA-surveys of the BZgA, which first started to examine tobacco consumption using representative samples of German adolescents in 1979, the proportion of smokers fell considerably among the 12 to 17 year olds. The decisive change came about in From 2001 to 2011, the smoking rate declined from 30% to 11% and 12% respectively among boys and girls [36]. These are, by far, the lowest percentages ever obtained across three decades of data collection. The findings of the ESPAD-surveys [24], the HBSC-surveys [28], and the KiGGS-suveys [27] confirm this trend. The causes for the change in smoking rates have not yet been examined empirically, in Germany or in other countries. Yet, at least in Germany, this reduction may be traced back to several measures that were introduced in Germany since These include: 1) The incremental, but overall very pronounced, increase in tax rates for tobacco, resulting in an almost three-fold price increase from 1980 to ) Legal changes that raised the minimum age for buying and consuming tobacco from 16 to 18 years of old.

6 Social Inequality and Substance Use and Problematic Gambling Current Drug Abuse Reviews, 2016, Vol. 9, No Table 2. Prevalence rates (%) and Odds ratios (OR) for smoking among adolescents and young adults by SES and migration background. Surveys, Age, Refs. Tobacco Consumption Characteristics SES-Indicators, Migration Background P % P % OR OR OR adj. for HBSC 2010 daily smoking family affluence low ns 2.8 *) a, m [68] [69] middle ns 1.0 ns high (ref) HBSC 2006 smoking 1x/week family affluence low ns 1.4 ns a, o, s [53] middle ns 1.2 ns high (ref) HBSC 2002 smoking 1x/week family affluence low ns 1.0 ns a, ö, s [64] middle ns 1.0 ns high (ref) HBSC 2006 regular smoking parental occupational status low ns 0.9 ns a, g, m [53] middle ns 0.8 ns high (ref) HBSC 2002 smoking 1x/week parental occupational status low ns 1.2 ns a, f, s [64] middle ns 1.1 ns high (ref) KiGGS 2006 current smoking parental SES low ns 1.8 ns a, m, s [65] middle ns 1.5 ns high (ref) KiGGS 2012 daily smoking parental SES low *) 7.0 *) a [38] middle *) 5.4 *) high (ref) HBSC 2010 regular smoking school type HS 12.2 *) 15.5 *) [7ß] RS 6.7 ns 7.6 ns GY (ref) HBSC 2006 regular smoking school type all others *) 3.0 *) a, f, o [53] GY (ref) HBSC 2002 smoking 1/week school type all others *) 2.5 *) a, f, o [64] GY (ref)

7 32 Current Drug Abuse Reviews, 2016, Vol. 9, No. 1 Henkel and Zemlin (Table 2) contd.. Surveys, Age, Refs. Tobacco Consumption Characteristics SES-Indicators, Migration Background P % P % OR OR OR adj. for DA 2012 daily smoking school type HS 10.2 *) OR np a, g [40] RS 4.0 *) OR np GY (ref) DA 2010 daily smoking school type HS 7.5 *) OR np a, g [72] RS 3.6 ns OR np GY (ref) DA 2012 >20 cigarettes/daily school type HS 4.2 *) OR np a, g [40] RS 0.2 ns OR np GY (ref) DA 2010 >10 cigarettes/daily school type HS 5.0 *) OR np a, g [72] RS 0.9 ns OR np GY (ref) KiGGS 2006 current smoking school type HS *) 3.4 *) a, m, p [65] RS *) 1.7 *) GY (ref) DA 2012 daily smoking employment status unemployed 41.9 *) OR np [40] employed 29.1 *) OR np GY (ref) DA 2010 daily smoking employment status unemployed 49.0 *) OR np a, g [72] employed 30.1 ns OR np GY (ref) DA 2012 >20 cigarettes/daily employment status unemployed 18.0 *) OR np a, g [40] employed 6.6 *) OR np GY (ref.) DA 2010 >20 cigarettes/daily employment status unemployed 17.0 *) OR np a, g [72] employed 6.8 *) OR np GY (ref.) DA 2012 regular e-cigarette smoking last 30 days employment status unemployed 4.5 *) OR np a, g [40] employed 1.6 ns OR np GY (ref.)

8 Social Inequality and Substance Use and Problematic Gambling Current Drug Abuse Reviews, 2016, Vol. 9, No (Table 2) contd.. Surveys, Age, Refs. Tobacco Consumption Characteristics SES-Indicators, Migration Background P % P % OR OR OR adj. for KiGGS 2006 second-hand smoking parental SES low *) 3.2 *) a, m [73] middle *) 2.2 *) high (ref) KiGGS 2006 second-hand smoking school type HS *) 1.5 *) a, m [27] RS *) 1.7 *) GY (ref) KiGGS 2006 current smoking migration background yes *) 0.5 *) a [65] no (ref) KiGGS 2006 current smoking migration background both parents 12.8 ns [23] one parent 16.5 ns no (ref) 16.9 HBSC 2010 smoking 1x/week migration background both parents *) 1.0 ns a, f [68] [69] one parent *) 0.7 ns no (ref) Note: P % prevalence rate, *) p <0.05, ns p >0.05, chi2-test/or logistic regression adjusted for covariates, a (age), g (gender), f (FAS), m (migration background), o (parental occupational status), p (parental SES), s (school type), (ref) reference, nss no test for statistical significance was conducted, np OR-value not published. 3) Tobacco advertising restrictions. 4) The implementation of a widespread smoking ban in all public buildings, on all means of transport, schools, restaurants, pubs and so on since the introduction of a number of laws banning smoking in These measures were supported by awareness campaigns involving mass media and led to greater public recognition of tobacco smoking as a problematic behaviour. One can assume that these campaigns also led to a significant change in attitudes about smoking among children and adolescents [27, 28]. The question remains as to whether these measures have also changed the smoking rates for different social groups. The results from the surveys can be summarized as follows: The developments can be differentiated according to different school types (Tables 3 and 4). Some noticeable general trends emerged, indicating a decline. However, these trends were influenced by school type. The ESPAD-surveys [24] show that the proportion of pupils who smoke declined in all three school types since 2003 (Table 3). However, by 2011, the HS-smoking rate was back to 19.2 percentage points above the rates observed for GY-students). And since 2003, the difference between the prevalence rate of heavy smoking (>10 cigarettes per day) among smoking HSstudents and GY-students increased from 7.4 percentage points in 2003 to 16.5 points in 2011 (Table 4) Alcohol A review of the results outlined in Tables 5-7 show that the relationship between SES and alcohol consumption is less clear compared to the relationship between SES and smoking: 1) With regard to family affluence (FAS), two HBSCsurveys (2002, 2006) show that boys with low and middle FAS are significantly less likely to have binge drinking experience compared to boys with high FAS. Differences in parental occupational status and parental SES, on the other hand, were not associated with different prevalence rates observed for binge drinking or hazardous alcohol consumption (Table 5). 2) With regards to European findings, a negative relationship exists between the prevalence of binge drinking and the level of education. That is, the prevalence rates rise as the level of education declines [5]. The results of German surveys do not provide similarly consistent findings (Tables 5-7). According to the DA- surveys, no significant school type differences exist in terms of binge drinking prevalence rates (Table 5). In contrast to this, the 2002, 2006 and 2010 HBSC-surveys and the ESPADsurveys reveal substantially lower prevalence rates for binge drinking among GY-students in comparison to HS-students resp. to HS- plus RS-students. The 2006 HBSC-survey also shows gender differences (Tables

9 34 Current Drug Abuse Reviews, 2016, Vol. 9, No. 1 Henkel and Zemlin Table 3. Trend data , 30-days-prevalence rates (%) for smoking among adolescents (age years) by SES (school type). ESPAD- Surveys HS Prevalence Rate % RS Prevalence Rate % GY Prevalence Rate % Difference PP HS-GY *) 50.4 *) 37.2 *) 18.6 nss ns 38.9 ns 29.9 ns 12.4 nss 2011 (ref) nss Difference PP Note: *) p <0.05, ns p >0.05, OR logistic regression adjusted for age, gender, OR-value not published, (ref) reference, nss no test for statistical significance was conducted, PP percentage points [24]. Table 4. Trend data , 30-days-prevalence rates (%) for smoking >10 cigarettes per day among smoking adolescents (age years) by SES (school type). ESPAD-Surveys HS Prevalence Rate % RS Prevalence Rate % GY Prevalence Rate Difference PP HS-GY nss nss nss Difference PP nss 10.3 nss 6.7 nss Note: nss no test for statistical significance was conducted, PP percentage points [24] [74] [33]. 5 and 6). No systematic relationship appears to exist between hazardous alcohol use and school type (Tables 5 and 7). 3) HS-students reported more severe problems associated with alcohol consumption than their GYcounterparts (ESPAD 2011) [24]. The problems ranged from exhibiting aggressive behaviour, engaging in unprotected sexual intercourse, to poor school performance, personal experience of violence as well as problems with the police. These problems are typical consequences of binge drinking [5]. 4) No significant relationships were found between the employment status of 18 to 25 year olds, their rate of binge drinking and risky alcohol consumption (Table 5). However, the relationships between these variables become more distinctive when gender differences are also considered. The rate of high risky alcohol consumption (hazardous alcohol use plus binge drinking 1x/last 30 days) is significantly higher among young unemployed men (BZgA-survey 2010) [29], while the opposite is true for women (Table 5). 5) Results related to migration background show the following consistent trends: Problematic patterns of alcohol use are less frequently observed among adolescents and young adults who have Turkish, Asian or Middle Eastern/Northern African backgrounds. These findings apply to binge drinking, hazardous as well as exceedingly risky alcohol consumption (the latter includes both hazardous and binge drinking, Table 5). This pattern, where examined to date, appears to be similar for both males and females Trends ( ) The mass media in Germany (TV, newspapers, and Internet outlets) often discusses the issue of alcohol consumption among adolescents in very dramatic terms. These reports are based on the number of children and adolescents admitted to hospital due to acute alcohol intoxication. The number of 10 to 20 year olds being admitted to hospital for this reason has more than doubled between 2000 and 2013 from around 9,500 cases to more than 23,000 cases per year [30]. Still, this relatively small and special group of adolescents may not be representative of the alcohol consumption patterns in adolescents as a whole. Representative data from the ESPAD- and DA-surveys show, in fact, a drop in hazardous alcohol use among adolescents overall, as well as a declining or unchanged prevalence rate in terms of binge drinking [24, 29]. This is observed for both genders. The evaluations of the 2002, 2006 and 2010 HBSC-surveys confirm these findings [28]. Similarly trends were observed at an international level. The proportion of adolescents who consumed alcohol significantly declined in various European and North American countries between 2002 and 2010 [31], including Germany. Unfortunately it was not analysed to what extent this trend differed among various SES groups [31]. With regard to social inequality (attended school type), the ESPAD-surveys revealed the following: The prevalence rates associated with hazardous alcohol consumption significantly declined among HS- and RS-students between 2003 and 2011 respectively, and between 2007 and 2011 among GY-students (Table 7). The differences in prevalence rates between the three school types also diminished over

10 Social Inequality and Substance Use and Problematic Gambling Current Drug Abuse Reviews, 2016, Vol. 9, No Table 5. Prevalence rates (%) and Odds Ratios (OR) for binge drinking and hazardous alcohol consumption among adolescents and young adults by SES and migration background. Surveys, Age, Refs. Alcohol Consumption Characteristics SES-Indicators, Migration Background P % P % OR OR OR adj. for HBSC 2010 binge drinking 4x/lifetime family affluence low 8.1 nss 7.4 nss [68] middle 6.6 nss 4.5 nss high (ref) 6.3 nss 4.9 nss HBSC 2006 binge drinking 2x/lifetime family affluence low *) 0.9 ns a, o, s [53] middle ns 1.0 ns high (ref) HBSC 2002 binge drinking 2x/lifetime family affluence low ns 0.9 ns a, o, s [64] middle *) 1.0 ns HBSC 2006 binge drinking 2x/lifetime parental occupational status high (ref) low ns 0.9 ns a, f, s [53] middle ns 0.9 ns HBSC 2002 binge drinking 2x/lifetime parental occupational status high (ref) low ns 1.0 ns a, f, s [64] middle ns 1.2 ns KiGGS 2012 binge drinking 1x/last month high (ref) parental SES low 12.9 ns 8.0 ns [39] middle 13.8 ns 11.3 ns KiGGS 2012 hazardous alcohol consumption AUDIT-C Score 5/4 boys/girls high (ref) parental SES low ns 0.8 ns a [39] middle ns 1.1 ns high (ref) HBSC 2010 binge drinking 2x/lifetime school type HS 22.3 *) 16.5 *) [70] RS 14.4 ns 12.5 ns GY (ref) HBSC 2006 binge drinking 2x/lifetime school type all others *) 1.5 ns a, f, o [53] GY (ref) HBSC 2002 binge drinking 2x/lifetime school type all others *) 1.4 *) a, f, o [64] GY (ref) ns 1.0

11 36 Current Drug Abuse Reviews, 2016, Vol. 9, No. 1 Henkel and Zemlin (Table 5) contd.. Surveys, Age, Refs. Alcohol Consumption Characteristics SES-Indicators, Migration Background P % P % OR OR OR adj. for DA 2012 binge drinking 1x/last month school type HS 14.3 ns OR np a, g [75] RS 12.1 ns OR np GY (ref) DA 2012 hazardous alcohol consumption school type HS 5.2 ns OR np a, g [75] RS 3.3 ns OR np GY (ref) DA 2010 hazardous alcohol consumption + binge drinking 1x/last month school type HS ns OR np ns OR np a, e, m [29] RS ns OR np ns OR np GY (ref) DA 2012 hazardous alcohol consumption employment status unemployed 18.7 ns OR np [75] employed 14.8 ns OR np GY (ref) DA 2012 binge drinking 1x/last month employment status unemployed 41.7 ns OR np a, g [75] employed 42.3 ns OR np GY (ref) DA 2010 hazardous alcohol consumption + binge drinking 1x/last month employment status unemployed *) OR np *) OR np a, m. s [29] employed ns OR np ns OR np GY (ref) DA 2012 hazardous alcohol consumption migration background Western Europe 8.6 ns OR np a, e, s [75] Eastern Europe 2.5 ns OR np Turkey, Asia 0.1 *) OR np no (ref) DA 2012 hazardous alcohol consumption migration background Western Europe 16.0 ns OR np a, e, s [75] Eastern Europe 13.0 ns OR np Turkey, Asia 5.3 *) OR np no (ref) DA 2012 binge drinking 1x/last month migration background Western Europe 14.7 ns OR np a, e, s [75] Eastern Europe 12.5 ns OR np Turkey, Asia 2.6 *) OR np no (ref)

12 Social Inequality and Substance Use and Problematic Gambling Current Drug Abuse Reviews, 2016, Vol. 9, No (Table 5) contd.. Surveys, Age, Refs. Alcohol Consumption Characteristics SES-Indicators, Migration Background P % P % OR OR OR adj. for DA 2012 binge drinking 1x/last month migration background Western Europe 29.1 ns OR np a, e, s [75] Eastern Europe 36.1 ns OR np Turkey, Asia 16.0 *) OR np no (ref) DA 2010 hazardous alcohol consumption + binge drinking 1x/last month migration background Turkey/Asia *) OR np *) OR np a, e, s [29] Eastern Europe *) OR np ns OR np Western Europe ns OR np *) OR np no (ref) DA 2010 hazardous alcohol consumption + binge drinking 1x/last month migration background Turkey/Asia *) OR np *) OR np a, e, s [29] Eastern Europe ns OR np ns OR np Western Europe ns OR np ns OR np no (ref) HBSC 2010 binge drinking 4x/lifetime migration background both parents 7.6 nss 4.1 nss - [68] one parent 11.6 nss 4.5 nss no (ref) Note: P % prevalence rate, *) p <0.05, ns p >0.05, chi2-test/ OR logistic regression adjusted for covariates, a (age), g (gender),s (school type), e (employment status), f (FAS), m (migration background), o (parental occupational status), s (school type), (ref) reference, nss no test for statistical significance was conducted, np OR-value not published. time. With regard to binge drinking, a remarkable difference (11.4 percentage points) in prevalence rates persists in 2011 for HS-students compared to GY-students (Table 6). Unfortunately no test for statistical significance was conducted Cannabis The situation in Germany is provided in Tables They show the following results. 1) Adolescents use of cannabis does not systematically correlate with family affluence, socioeconomic or occupational status of the parents (Table 8). See also [32]. 2) Migration background does not appear to have a significant effect on cannabis consumption among the 12 to 17 year olds or the 18 to 25 year olds (Table 8). However, information about gender differences was missing. 3) No significant differences emerged in terms of cannabis consumption and school type (Table 8). But Table 6. Trend data , prevalence rates (%) for binge drinking 3x/last month among adolescents (age years) by SES (school type). ESPAD-surveys HS Prevalence Rate % RS Prevalence Rate % GY Prevalence Rate % Difference PP HS-GY ns 32.0 ns 18.6 ns 18.4 nss ns 34.9 *) 23.5 ns 12.9 nss 2011 (ref) nss Difference PP Note: *) p <0.05, ns p >0.05, OR logistic regression adjusted for age, gender, OR-value not published, (ref) reference, nss no test for statistical significance was conducted, PP percentage points [24]

13 38 Current Drug Abuse Reviews, 2016, Vol. 9, No. 1 Henkel and Zemlin Table 7. Trend data , prevalence rates (%) for hazardous alcohol consumption among adolescents (age years) by SES (school type). ESPAD-surveys HS Prevalence Rate % RS Prevalence Rate % GY Prevalence Rate % Difference PP HS-GY *) 15.1 *) 9.8 ns 3.3 nss ns 14.7 *) 10.6 *) 1.8 nss 2011 (ref) nss Difference PP Note: hazardous consumption = average daily intake >12g/>24g pure alcohol (boys/girls), *) p <0.05, ns p >0.05, OR logistic regression adjusted for age, gender, OR-value not published, (ref) reference, nss no test for statistical significance was conducted, PP percentage points [24] again, no information was available about gender differences. 4) In an effort to support prevention, it is important to identify risky forms of cannabis consumption as well as those groups who may be at risk. The 2007 and 2011 ESPAD-surveys [24, 33] are the only ones that examined risky cannabis consumption with the help of the Cannabis Abuse Screening Test (CAST) by Legleye et al. [16]. According to the 2011 results, the problematic consumption (CAST 4) is obviously unequally distributed (Tables 8-10). The prevalence rate is highest among HS-students (2.9%). With respect to past-year use, the percentage of problematic consumers among cannabis consuming HS-students was 17%, almost four times higher than the percentage of problematic consumers among cannabis consuming GY-students (Tables 8 and 9). In other words: when HS-students start consuming cannabis, they are much more likely to develop a problematic pattern of cannabis use, which may then lead to later cannabis addiction. 5) In addition to HS-students, another risk group includes the 18 to 25 year olds unemployed (DAsurvey 2012, Table 8). Their 30 days prevalence rate of 14.5% is significantly higher than the prevalence rate of 3.1% reported for GY-students of similar age, and the prevalence rate is three times higher for those who consume cannabis on a regular basis (that is, weekly), even after adjusting for age and gender (Table 8) Trends ( ) According to an international study by Hublet et al. [25], the proportion of adolescents consuming cannabis declined significantly in Western European countries between the years of 2002 and The consumption of cannabis among adolescents and young adults in Germany has also declined for both males and females since 1997 [34]. While the 1997 prevalence rate among 12 to 17 year olds was around 5% (based on consumption of cannabis over the course of 30 days), this rate had declined to 2% by A similar decline from around 10% to 6% was observed for 18 to 25 year olds. Evaluations based on the HSBC-surveys confirm this trend [28]. Whether or not these changes are similar for the different SES groups can only be answered in relation to school type. The ESPAD-surveys [24] indicate a significant and substantial decline for the 30 days prevalence rates for pupils from all three school types from 2003 to At the same time, the differences in prevalence rates between the school types shrank almost entirely over the same time frame (Table 9). The problematic cannabis consumption as assessed by CAST reveals a contrasting development over time. Table 10 shows little change in prevalence rates among RS- and GYstudents in 2007 and 2011 (no data was available for 2003). However, the rate increases significantly for HS-students. When considering the change in prevalence rates across all HS-students, an overall increase is observed (from 0.7% to 2.9%). This increase appears to be most pronounced among those HS-students who had consumed cannabis over 12 months (rate increased from 5.0% to 17.0%). Examining and following this noteworthy trend will be an important task in future research Other Substances With regard to the consumption of illegal drugs (excluding cannabis), a number of knowledge gaps exist that may need to be addressed in future work. The only indicator of social inequality that has been examined so far is school type. The results of the 2003 ESPAD-survey suggest that the prevalence rate of illegal drug use other than cannabis across the life span of HS- and RS-students was around 15%, double the rate observed for GY-students (Table 11). In 2011 the same difference persists for HS-students compared to GY-students. The prevalence rate has only declined significantly for RS-students between 2003 and 2011, but not in other student groups. To date, only the ESPAD-surveys examined the nonmedical use of prescription drugs. As is shown in Table 12, the prevalence rate across the lifespan for the use of tranquilizers or sedatives increased significantly among HSstudents between 2003 and The difference in prevalence rates observed for HS-students compared to GYstudents also appears to become more pronounced over the same time frame Problematic Gambling The results from the 2002 HBSC-PG-survey [35] and the 2011 ESPAD-survey [24] provide informations about

14 Social Inequality and Substance Use and Problematic Gambling Current Drug Abuse Reviews, 2016, Vol. 9, No Table 8. Prevalence rates (%) and Odds ratios (OR) for cannabis consumption among adolescents and young adults by SES and migration backround. Surveys, Age, Refs. Cannabis Consumption Characteristics SES-Indicator, Migration Background P % P % OR OR OR adj. for HBSC x/last month family affluence low 6.7 ns 5.9 ns [68] middle 4.9 ns 2.5 ns high (ref) HBSC x/last 12 months family affluence low ns a, o, s [53] middle ns high (ref) HBSC x/last 12 months family affluence low ns 0.6 ns a, o, s [64] middle ns 0.9 ns high (ref) HBSC x/last 12 months parental occupational status low ns a, f, s [53] middle ns high (ref) 2.7 HBSC x/last 12 months parental occupational status low ns 0.8 ns a, f, s [64] middle ns 0.9 ns high (ref) KiGGS x/last 12 months parental SES low ns 0.8 ns a, m, s [65] middle ns 0.9 ns high (ref) KiGGS x/last 12 months school type HS ns 1.5 ns a, m, p [65] RS ns 1.3 ns GY (ref) HBSC x/last 12 months school type all others ns a, f, o [53] GY (ref) HBSC x/last 12 months school type all others ns 1.2 ns a, f, o [64] GY (ref) KiGGS x/last 12 months school type HS ns 1.4 ns a, m, p [65] RS GY (ref)