Epidemiological trends and risk factors for tobacco, alcohol and drug use among adolescents in Scotland, 2002–13

Epidemiological trends and risk factors for tobacco, alcohol and drug use among adolescents in... Abstract Background This study estimates trends in prevalence, and patterns, of individual and multiple substance use between 2002 and 2013 amongst adolescents in Scotland. Methods The study uses data from 134 387 participants of the biennial national ‘Scottish Schools Adolescent Lifestyle and Substance Use Survey’ on smoking, alcohol and illicit drug use. Current regular use and current heavy use of smoking, alcohol, illicit drugs and multiple substances was measured. Time trends in the prevalence of each outcome were estimated using univariate and multivariate logistic regression. Results Regular smoking, alcohol, illicit drug and multiple substance use declined significantly amongst adolescents in Scotland. However, multivariate analyses that focussed upon high-risk levels of these behaviours revealed an upward linear trend in heavy alcohol (OR = 1.06; 95% CI: 1.04, 1.07) and heavy illicit drug (OR = 1.04; 95% CI: 1.00, 1.08) use (P < 0.05). Non-white pupils were more likely to be involved in individual and multiple substance use than ethnically white British pupils. In comparison to pupils from the least deprived socioeconomic quintile, pupils from the most deprived quintile had increased odds of 1.41 (95% CI: 1.02, 1.97; P < 0.05) and 1.62 (95% CI: 1.14, 2.29; P < 0.05) of being regular and heavy multiple substance users, respectively. Conclusions Further effort is required to tackle heavy alcohol and heavy illicit drug use amongst adolescents in Scotland. Prevention strategies should be informed by the risk profiles of substance misusers and evidence around the clinical and cost-effectiveness of preventive interventions. adolescents, alcohol, illicit drugs, Scotland, tobacco Introduction The adverse sequelae of tobacco, alcohol and illicit drug use are well documented.1–6 The World Health Organization (WHO) has highlighted that understanding the prevalence and role of these behavioural risk factors should play a crucial part in developing clear and effective strategies for improving global health.7 In industrialized nations, initiation of tobacco, alcohol and illicit drug use tends to occur during adolescence, a critical period of life in which risky behaviours often result in embeddedness during the remaining life course.8 Moreover, risky adolescent behaviours such as tobacco, alcohol and illicit drug use often co-occur, which in turn compounds the risk of a host of adverse health, social and economic consequences.9 The prevalence of risky adolescent behaviours during adolescence varies by behaviour and across jurisdictions. Data for the years 2000–07 from 140 WHO member states collected as part of the Global Youth Tobacco Survey revealed that ~9.5% of 13–15 year olds smoked cigarettes, with prevalence ranging from 4.9% in the Eastern Mediterranean Region to 19.2% in the European Region.10 In the United States, surveillance data collected during 2010 and 2011 revealed that 18.1% of high school students in grades 9–12 had smoked cigarettes during the 30 days before the survey, with evidence of higher prevalence amongst male (19.9%) than female (16.1%) students.11 Furthermore, data compiled across national surveys in the United States reveal that more than one half of adolescents in the United States report alcohol use, and nearly one-fourth report exposure to illicit drugs.12 In the United Kingdom (UK), recent evidence based on national surveys suggests a reduction in the prevalence of cigarette smoking, alcohol use and illicit drug use amongst adolescents in England.13 However, close scrutiny of these epidemiological data suggests mixed patterns of multiple substance use amongst adolescents that differ by sociodemographic profile, highlighting areas where future preventive efforts should be targeted.13 In contrast to other industrialized nations, there is a relative paucity of published national epidemiological evidence on substance use amongst adolescents in Scotland. Levin et al.14 analysed national data on 2692 15 year olds included in The Scottish Health Behaviour in School-aged Children Study, conducted between March and June 2010. They found that 13.6% of boys and 18.9% of girls were current smokers, broadly in accordance with data from unpublished reports based on The Scottish Schools Adolescent Lifestyle and Substance Use Survey (SALSUS).15,16 They also found that prevalence of smoking was the highest amongst those living in the second most deprived socioeconomic quintile. Recent Scotland-wide epidemiological data on alcohol and illicit drug use amongst adolescents has, to our knowledge, been restricted to unpublished reports.15–17 Since the Scottish Parliament was established in 1999, the Scottish Government has introduced a number of policy initiatives aimed at tackling substance misuse amongst adolescents. This has included, but not limited to, a ban on tobacco advertising in 2002, an increase in the age for tobacco sales from 16 to 18 years in 2007, a ban on the display of cigarettes for sale in shops and self-service sales from automatic vending machines in 2010, and national frameworks aimed at mitigating the damaging impacts that alcohol and drug misuse have on families and communities, including young people. The objective of this study was to estimate trends in the prevalence, and patterns, of individual and multiple substance use amongst adolescents in Scotland against this policy background. Methods Data sources Data from SALSUS formed the basis of this empirical investigation. SALSUS is a continuation of a series of biennial national surveys on smoking, alcohol and illicit drug use among young people that were carried out jointly in Scotland and England between 1982 and 2000.15 From 2002, Scotland has carried out its own national survey, namely SALSUS, which generates epidemiological data on substance use among adolescents, and provides a vehicle for monitoring progress towards Scottish Government targets on smoking, alcohol and illicit drug use. The SALSUS surveys were carried out biennially between 2002 and 2010 (2002: n = 23 090; 2004: n = 7 062; 2006: n = 23 180; 2008: n = 10 063; 2010: n = 37 307) and subsequently after a three year period in 2013 (n = 33 685). SALSUS takes the form of a confidential, self-completed questionnaire completed by secondary school second year (S2) (average age of 13 years) and secondary school fourth year S4 (average age of 15 years) pupils in school settings. In each survey year, the Scottish Government schools database was used as the sampling frame and included all state funded, grant-maintained and independent secondary schools across the country, but excluded schools dedicated to children with additional support needs. Each survey adopted a multistage sample design that determined the probability of being a selected school, and the probability of being a selected class within that school. With the exception of the 2004 survey, weighting for school type and age group non-response was also applied within local authority strata to ensure that the samples were representative both at a national and a local authority level; the weighting system applied in 2004 ensured that the sample was nationally representative. Consent to participate was provided both by schools and the pupils and their parents. The overall response rate, calculated from the school, class and pupil response rates, varied between 57% in 2006 and 65% in 2002. Detailed methodology for each of the SALSUS surveys, including survey design, sampling strategy, questionnaire design and consent procedures is described in the appendices of the SALSUS annual reports (http://www.scotpho.org.uk/publications/overview-of-key-data-sources/surveys-cross-sectional/scottish-schools-adolescent-lifestyle-a-substance-use-survey). SALSUS data are publicly available and were downloaded from the UK Data Archive on 16 May 2016 (http://www.data-archive.ac.uk/). Outcome measures All substance use measures were self-reported by the adolescents as part of questionnaires completed under exam conditions. Each questionnaire was returned to the class teacher in a sealed envelope without the reporting of names to ensure confidentiality. The outcome variables were defined separately for two levels of substance misuse: current regular use and current heavy use. Current regular smoking was defined as usually smoking one cigarette a week or more; current regular alcohol use was defined as drinking once a week or more on average; and current illicit drug use was defined as having taken any illicit drugs in the last month. These definitions were broadly consistent with those applied in previous studies of adolescent substance use in the UK.13 Current regular multiple substance use was defined as engaging simultaneously in all these behaviours. Using these definitions, data were available for all study years with the exception of 2004. With regards to current heavy substance use, the report of smoking at least 60 cigarettes in last week, drinking at least 21 units of alcohol in the last week and taking illicit drugs most days were considered measures of heavy smoking, alcohol and illicit drug use, respectively. In keeping with the operational definition adopted by SALSUS, heavy multiple substance use was defined as engaging in at least two out of three of these behaviours.15,16 Using these definitions, data were available for all study years with the exception of 2002 for heavy smoking, 2002, 2004 and 2006 for heavy alcohol use, and 2004 for heavy illicit drug use. Sociodemographic data incorporated in the SALSUS surveys included gender (male, female), school year (S2, S4; indicative of age) and ethnicity (Scottish/white British, white other, other ethnicity, do not know/refused to answer). It also included socioeconomic quintile derived from Scottish Index of Multiple Deprivation (SIMD)18 ranks that were themselves derived from postcodes for home addresses reported by the pupils. Socioeconomic data were only available from 2006 onwards. Statistical analysis The prevalence of individual and multiple regular and heavy substance use behaviours was calculated for the total sample in each survey year and separately by gender, school year, ethnicity and socioeconomic quintile within each survey year. Statistical analysis provided a description of the time trends for each outcome measure in two alternative ways: (i) percentage change between the first and last survey year available and its 95% confidence interval (CI), calculated using univariate logistic regression with survey year as the only independent variable considered on a nominal scale; and (ii) annual change adjusted for the independent variables (gender, school year, ethnicity and socioeconomic quintile) using multivariate logistic regression with survey year treated as a continuous independent variable. The rationale for these two different methods of time trend calculation was to evaluate the effects of restricting the trend to be linear on a logarithmic scale with survey year as a continuous predictor variable and of adjusting for the independent variables other than survey year. The univariate logistic regressions were used to calculate the marginal distributions of the survey year estimates (via the post-estimation command ‘margins’ in STATA). Subsequently, the differences between the first versus last survey year estimates were divided by their respective time spans to generate average annual changes. The latter were then multiplied by 100 to generate average annual percentage point changes. In addition, estimates of marginal odds and associated standard errors within the univariate regressions were converted into odds ratios (ORs) and corresponding 95% CIs, and presented graphically as time trends. Interactions between the independent variables were also investigated. In keeping with previous similar studies with large samples,13 missing data were handled by using listwise deletion, except for known confounding effects of non-response such as refusing to answer or answering ‘don’t know’ to the question on respondent’s ethnicity. The latter may be an indicator of social vulnerability (e.g. for immigrants) and correlated with higher risk of substance use. All regression analyses accounted for clustering of respondents within primary sampling units and the varying probability of selecting an individual from the target population. The distribution of multivariate regression residuals was examined to verify whether the assumptions for model estimation had been met. In addition, two types of sensitivity analysis were performed: (i) by adding independent variables related to family circumstances (free school meal entitlement, household composition, maternal and paternal knowledge of children’s activities, amount of pocket money) in order to verify the stability of time trends for substance use; and (ii) by applying more conservative definitions of current regular use of tobacco and alcohol covering the past month rather than the past week in order to enhance comparability with some international surveys.10,19 All analyses were performed using STATA software (StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP). Results Prevalence of substance use Table 1 summarizes the weighted prevalence of adolescent substance use by type of substance and heaviness of use for the overall samples in each survey year. The weighted prevalence of regular smoking, alcohol, illicit drug and multiple substance use declined steadily between 2002 and 2013 from 13.06 to 5.22%, 27.09 to 6.78%, 14.27 to 5.61% and 5.48 to 1.44%, respectively. Uninterrupted declines in regular substance use were also observed in each sociodemographic subgroup with the exception of all ethnicity subgroups for which the prevalence of regular smoking, alcohol and multiple substance use peaked in 2006, and all socioeconomic subgroups for which the prevalence of regular smoking, illicit drug and multiple substance use peaked in 2008 (data not reported). Table 1 Number (%)* of adolescents using substances by type of substance, heaviness of use and year of survey   2002  2004  2006  2008  2010  2013  Total surveyed (N)  23 090  7062  23 180  10 063  37 307  33 685  Regular substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  2 949 (13.06)  –  2 118 (9.67)  891 (9.54)  2 857 (7.80)  1 670 (5.22)   Drinking  6 166 (27.09)  –  4 472 (20.12)  1 579 (16.76)  4 645 (12.90)  2 236 (6.78)   Drug use  3 165 (14.27)  –  1 787 (8.31)  769 (8.13)  2 502 (6.95)  1 849 (5.61)   Multiple substances  1 222 (5.48)  –  719 (3.40)  332 (3.53)  961 (2.59)  476 (1.44)  Heavy substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  –  171 (2.42)  389 (1.89)  339 (3.76)  658 (1.81)  327 (1.08)   Drinking  –  –  –  511 (5.32)  1 587 (4.29)  840 (2.55)   Drug use  416 (1.93)  –  203 (0.95)  85 (0.83)  323 (0.85)  294 (0.88)   Multiple substances  –  –  44 (0.25)  124 (1.32)  410 (1.08)  200 (0.60)    2002  2004  2006  2008  2010  2013  Total surveyed (N)  23 090  7062  23 180  10 063  37 307  33 685  Regular substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  2 949 (13.06)  –  2 118 (9.67)  891 (9.54)  2 857 (7.80)  1 670 (5.22)   Drinking  6 166 (27.09)  –  4 472 (20.12)  1 579 (16.76)  4 645 (12.90)  2 236 (6.78)   Drug use  3 165 (14.27)  –  1 787 (8.31)  769 (8.13)  2 502 (6.95)  1 849 (5.61)   Multiple substances  1 222 (5.48)  –  719 (3.40)  332 (3.53)  961 (2.59)  476 (1.44)  Heavy substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  –  171 (2.42)  389 (1.89)  339 (3.76)  658 (1.81)  327 (1.08)   Drinking  –  –  –  511 (5.32)  1 587 (4.29)  840 (2.55)   Drug use  416 (1.93)  –  203 (0.95)  85 (0.83)  323 (0.85)  294 (0.88)   Multiple substances  –  –  44 (0.25)  124 (1.32)  410 (1.08)  200 (0.60)  *Note: Survey weights were not applied for the number of respondents using a substance (n) and the total surveyed (N). This differed from the respective percentages in parentheses, which were multiplied by survey weights, i.e. (n/N) × (survey weight) ×100. View Large Table 1 Number (%)* of adolescents using substances by type of substance, heaviness of use and year of survey   2002  2004  2006  2008  2010  2013  Total surveyed (N)  23 090  7062  23 180  10 063  37 307  33 685  Regular substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  2 949 (13.06)  –  2 118 (9.67)  891 (9.54)  2 857 (7.80)  1 670 (5.22)   Drinking  6 166 (27.09)  –  4 472 (20.12)  1 579 (16.76)  4 645 (12.90)  2 236 (6.78)   Drug use  3 165 (14.27)  –  1 787 (8.31)  769 (8.13)  2 502 (6.95)  1 849 (5.61)   Multiple substances  1 222 (5.48)  –  719 (3.40)  332 (3.53)  961 (2.59)  476 (1.44)  Heavy substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  –  171 (2.42)  389 (1.89)  339 (3.76)  658 (1.81)  327 (1.08)   Drinking  –  –  –  511 (5.32)  1 587 (4.29)  840 (2.55)   Drug use  416 (1.93)  –  203 (0.95)  85 (0.83)  323 (0.85)  294 (0.88)   Multiple substances  –  –  44 (0.25)  124 (1.32)  410 (1.08)  200 (0.60)    2002  2004  2006  2008  2010  2013  Total surveyed (N)  23 090  7062  23 180  10 063  37 307  33 685  Regular substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  2 949 (13.06)  –  2 118 (9.67)  891 (9.54)  2 857 (7.80)  1 670 (5.22)   Drinking  6 166 (27.09)  –  4 472 (20.12)  1 579 (16.76)  4 645 (12.90)  2 236 (6.78)   Drug use  3 165 (14.27)  –  1 787 (8.31)  769 (8.13)  2 502 (6.95)  1 849 (5.61)   Multiple substances  1 222 (5.48)  –  719 (3.40)  332 (3.53)  961 (2.59)  476 (1.44)  Heavy substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  –  171 (2.42)  389 (1.89)  339 (3.76)  658 (1.81)  327 (1.08)   Drinking  –  –  –  511 (5.32)  1 587 (4.29)  840 (2.55)   Drug use  416 (1.93)  –  203 (0.95)  85 (0.83)  323 (0.85)  294 (0.88)   Multiple substances  –  –  44 (0.25)  124 (1.32)  410 (1.08)  200 (0.60)  *Note: Survey weights were not applied for the number of respondents using a substance (n) and the total surveyed (N). This differed from the respective percentages in parentheses, which were multiplied by survey weights, i.e. (n/N) × (survey weight) ×100. View Large In contrast to the pattern for regular substance use, the weighted prevalence of heavy smoking, alcohol and multiple substance use peaked in 2008 and declined thereafter. Moreover, this pattern was observed for all gender, school year (age), ethnicity and socioeconomic subgroups. Trends in substance use: univariate regressions Table 2 summarizes the results of the univariate logistic regressions estimating time trends in substance use by type of substance and heaviness of use with survey year as the only independent variable. Average annual percentage point declines of 0.65, 1.69, 0.72 and 0.34% were estimated for regular smoking, alcohol, illicit drug and multiple substance use, respectively, between 2002 and 2013 (P < 0.0001). Lower average annual percentage point declines of 0.14, 0.46 and 0.09% were estimated for heavy smoking, alcohol and illicit drug use, respectively, between the first and last years surveyed for each of these substances (P < 0.0001). Notably, however, an average annual percentage point increase of 0.04% was estimated for heavy multiple substance use between the first and last years surveyed (P < 0.0001). Graphical representations of the time trends in probabilities of regular and heavy substance use, by type of substance, are presented in Fig. 1. Table 2 Average annual average percentage point change in substance use between first and last year surveyed; all adolescents Type of substance use  Time span (complete calendar years)a  Annual average % point changeb  SE  95% CI  P-value  Regular substance use   Smoking  12  −0.653  0.007  (−0.535, −0.771)  <0.0001   Drinking  12  −1.692  0.008  (−1.559, −1.825)  <0.0001   Drug use  12  −0.722  0.006  (−0.621, −0.823)  <0.0001   Multiple substances  12  −0.336  0.004  (−0.276, −0.396)  <0.0001  Heavy substance use   Smoking  10  −0.135  0.002  (−0.092, −0.177)  <0.0001   Drinking  6  −0.462  0.003  (−0.367, −0.558)  <0.0001   Drug use  12  −0.087  0.001  (−0.064, −0.111)  <0.0001   Multiple substances  8  0.043  0.001  (0.023, 0.064)  <0.0001  Type of substance use  Time span (complete calendar years)a  Annual average % point changeb  SE  95% CI  P-value  Regular substance use   Smoking  12  −0.653  0.007  (−0.535, −0.771)  <0.0001   Drinking  12  −1.692  0.008  (−1.559, −1.825)  <0.0001   Drug use  12  −0.722  0.006  (−0.621, −0.823)  <0.0001   Multiple substances  12  −0.336  0.004  (−0.276, −0.396)  <0.0001  Heavy substance use   Smoking  10  −0.135  0.002  (−0.092, −0.177)  <0.0001   Drinking  6  −0.462  0.003  (−0.367, −0.558)  <0.0001   Drug use  12  −0.087  0.001  (−0.064, −0.111)  <0.0001   Multiple substances  8  0.043  0.001  (0.023, 0.064)  <0.0001  SE, standard error; CI, confidence interval. aTime span (years) between first and last year surveyed. bAnnual percentage point change in substance use between first and last year surveyed. View Large Table 2 Average annual average percentage point change in substance use between first and last year surveyed; all adolescents Type of substance use  Time span (complete calendar years)a  Annual average % point changeb  SE  95% CI  P-value  Regular substance use   Smoking  12  −0.653  0.007  (−0.535, −0.771)  <0.0001   Drinking  12  −1.692  0.008  (−1.559, −1.825)  <0.0001   Drug use  12  −0.722  0.006  (−0.621, −0.823)  <0.0001   Multiple substances  12  −0.336  0.004  (−0.276, −0.396)  <0.0001  Heavy substance use   Smoking  10  −0.135  0.002  (−0.092, −0.177)  <0.0001   Drinking  6  −0.462  0.003  (−0.367, −0.558)  <0.0001   Drug use  12  −0.087  0.001  (−0.064, −0.111)  <0.0001   Multiple substances  8  0.043  0.001  (0.023, 0.064)  <0.0001  Type of substance use  Time span (complete calendar years)a  Annual average % point changeb  SE  95% CI  P-value  Regular substance use   Smoking  12  −0.653  0.007  (−0.535, −0.771)  <0.0001   Drinking  12  −1.692  0.008  (−1.559, −1.825)  <0.0001   Drug use  12  −0.722  0.006  (−0.621, −0.823)  <0.0001   Multiple substances  12  −0.336  0.004  (−0.276, −0.396)  <0.0001  Heavy substance use   Smoking  10  −0.135  0.002  (−0.092, −0.177)  <0.0001   Drinking  6  −0.462  0.003  (−0.367, −0.558)  <0.0001   Drug use  12  −0.087  0.001  (−0.064, −0.111)  <0.0001   Multiple substances  8  0.043  0.001  (0.023, 0.064)  <0.0001  SE, standard error; CI, confidence interval. aTime span (years) between first and last year surveyed. bAnnual percentage point change in substance use between first and last year surveyed. View Large Fig. 1 View largeDownload slide Time trends in probability of regular and heavy substance use; all adolescents. Solid lines represent regular substance use; dashed lines represent heavy substance use. Fig. 1 View largeDownload slide Time trends in probability of regular and heavy substance use; all adolescents. Solid lines represent regular substance use; dashed lines represent heavy substance use. Appendices 1–4 summarize the results of the univariate logistic regressions estimating time trends in substance use for each type of substance and heaviness of use, by sociodemographic subgroup. For each gender (Appendix 1) and school year (Appendix 2) subgroup, there were statistically significant average annual percentage point decreases in regular smoking, alcohol, illicit drug and multiple substance use and in heavy smoking, alcohol and illicit drug use, but also a statistically significant increase in heavy multiple substance use. A similar pattern was observed for the ethnicity subgroups with the exception of regular smoking, for which a non-statistically significant decline was estimated in all the ethnicity subgroups, and heavy illicit drug use, for which an increase in use was estimated in all the ethnicity subgroups (Appendix 3). Similarly, when the analyses were replicated by socioeconomic quintile, a temporal increase in heavy drug use was estimated within each quintile, although the estimated average annual percentage point increase was only statistically significant (P = 0.039) in the most deprived socioeconomic quintile (Appendix 4). Trends and patterns in substance use: multivariate regressions The multivariate regressions revealed significant downward linear trends over time for all forms of regular substance use: ORs (95% CIs) of 0.92 (0.91, 0.94), 0.84 (0.83, 0.85), 0.96 (0.95, 0.97) and 0.91 (0.87, 0.93) for regular smoking, alcohol, illicit drug and multiple substance use, respectively (P < 0.001) (Table 3). In contrast, the multivariate regressions revealed significant upward linear trends over time for heavy alcohol (OR = 1.06; 95% CI: 1.04, 1.07) and illicit drug (OR = 1.04; 95% CI: 1.00, 1.08) use (P < 0.05). Table 3 Factors predicting individual risk behaviours amongst all adolescents (n = 96721)a 2002–13   Regular substance use  Heavy substance use  Smoking  Drinking  Drug use  Multiple  Smoking  Drinking  Drug use  Multiple  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Year  0.92* (0.91, 0.94)  0.84* (0.83, 0.85)  0.96* (0.95, 0.97)  0.91* (0.87, 0.93)  0.91* (0.89, 0.93)  1.06** (1.04, 1.07)  1.04** (1.00, 1.08)  1.02 (0.99, 1.05)  Gender   Maleb  1.00                 Female  1.28* (1.19, 1.37)  0.96 (0.92, 1.01)  0.75* (0.69, 0.80)  0.91* (0.89, 0.93)  0.99 (0.88, 1.10)  0.68* (0.63, 0.74)  0.32* (0.25, 0.40)  0.57* (0.48, 0.68)  School year   S2 (13 years old)b  1.00                 S4 (15 years old)  4.95* (4.36, 5.63)  4.49* (4.09, 4.92)  4.62* (4.22, 5.07)  1.01 (0.90, 1.13)  6.15* (5.06, 7.49)  4.02* (3.56, 4.54)  3.61* (2.97, 4.37)  5.98* (4.62, 7.74)  Ethnicity   Scottish/White Britishb  1.00     White Other  1.12 (0.90, 1.39)  0.99 (0.83, 1.18)  1.38* (1.12, 1.69)  1.11 (0.70, 1.78)  0.90 (0.65, 1.26)  1.21 (0.96, 1.53)  1.80** (1.08, 3.00)  1.28 (0.70, 2.34)   Other Ethnicity  1.20** (1.00, 1.43)  1.04 (0.85, 1.27)  1.70* (1.41, 2.06)  1.39* (1.09, 1.77)  1.40** (1.00, 1.96)  1.49* (1.17, 1.90)  5.00* (3.96, 6.30)  2.72* (1.92, 3.85)   Don’t know/refused  2.86* (2.45, 3.35)  2.19* (1.93, 2.48)  1.79* (1.41, 2.27)  2.12* (1.67, 2.69)  4.21* (2.96, 5.98)  2.44* (1.93, 3.10)  6.84* (4.78, 9.79)  3.54* (2.42, 5.17)  Socioeconomic quintile   Fifth (least deprived)b  1.00  1.00  1.00  1.00  1.00  1.00  1.00  1.00   Fourth  1.21* (1.07, 1.37)  1.18* (1.10, 1.28)  1.03 (0.94, 1.13)  0.98 (0.83, 1.16)  0.98 (0.73, 1.32)  1.11 (0.94, 1.30)  0.73 (0.53, 1.01)  0.67** (0.50, 0.89)   Third  1.41* (1.25, 1.59)  1.32* (1.21, 1.45)  1.21* (1.06, 1.39)  1.15 (0.94, 1.39)  1.38** (1.06, 1.78)  1.41* (1.17, 1.70)  0.91 (0.67, 1.25)  1.13 (0.79, 1.61)   Second  1.70* (1.49, 1.94)  1.32* (1.17, 1.49)  1.33* (1.15, 1.53)  1.25* (1.01, 1.54)  1.86* (1.48, 2.32)  1.53* (1.31, 1.79)  1.03 (0.62, 1.71)  1.32 (0.91, 1.89)   First (most deprived)  1.96* (1.58, 2.44)  1.40* (1.18, 1.66)  1.54* (1.30, 1.82)  1.41* (1.02, 1.97)  2.27* (1.68, 3.09)  1.56* (1.22, 1.99)  1.11 (0.89, 1.38)  1.62** (1.14, 2.29)    Regular substance use  Heavy substance use  Smoking  Drinking  Drug use  Multiple  Smoking  Drinking  Drug use  Multiple  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Year  0.92* (0.91, 0.94)  0.84* (0.83, 0.85)  0.96* (0.95, 0.97)  0.91* (0.87, 0.93)  0.91* (0.89, 0.93)  1.06** (1.04, 1.07)  1.04** (1.00, 1.08)  1.02 (0.99, 1.05)  Gender   Maleb  1.00                 Female  1.28* (1.19, 1.37)  0.96 (0.92, 1.01)  0.75* (0.69, 0.80)  0.91* (0.89, 0.93)  0.99 (0.88, 1.10)  0.68* (0.63, 0.74)  0.32* (0.25, 0.40)  0.57* (0.48, 0.68)  School year   S2 (13 years old)b  1.00                 S4 (15 years old)  4.95* (4.36, 5.63)  4.49* (4.09, 4.92)  4.62* (4.22, 5.07)  1.01 (0.90, 1.13)  6.15* (5.06, 7.49)  4.02* (3.56, 4.54)  3.61* (2.97, 4.37)  5.98* (4.62, 7.74)  Ethnicity   Scottish/White Britishb  1.00     White Other  1.12 (0.90, 1.39)  0.99 (0.83, 1.18)  1.38* (1.12, 1.69)  1.11 (0.70, 1.78)  0.90 (0.65, 1.26)  1.21 (0.96, 1.53)  1.80** (1.08, 3.00)  1.28 (0.70, 2.34)   Other Ethnicity  1.20** (1.00, 1.43)  1.04 (0.85, 1.27)  1.70* (1.41, 2.06)  1.39* (1.09, 1.77)  1.40** (1.00, 1.96)  1.49* (1.17, 1.90)  5.00* (3.96, 6.30)  2.72* (1.92, 3.85)   Don’t know/refused  2.86* (2.45, 3.35)  2.19* (1.93, 2.48)  1.79* (1.41, 2.27)  2.12* (1.67, 2.69)  4.21* (2.96, 5.98)  2.44* (1.93, 3.10)  6.84* (4.78, 9.79)  3.54* (2.42, 5.17)  Socioeconomic quintile   Fifth (least deprived)b  1.00  1.00  1.00  1.00  1.00  1.00  1.00  1.00   Fourth  1.21* (1.07, 1.37)  1.18* (1.10, 1.28)  1.03 (0.94, 1.13)  0.98 (0.83, 1.16)  0.98 (0.73, 1.32)  1.11 (0.94, 1.30)  0.73 (0.53, 1.01)  0.67** (0.50, 0.89)   Third  1.41* (1.25, 1.59)  1.32* (1.21, 1.45)  1.21* (1.06, 1.39)  1.15 (0.94, 1.39)  1.38** (1.06, 1.78)  1.41* (1.17, 1.70)  0.91 (0.67, 1.25)  1.13 (0.79, 1.61)   Second  1.70* (1.49, 1.94)  1.32* (1.17, 1.49)  1.33* (1.15, 1.53)  1.25* (1.01, 1.54)  1.86* (1.48, 2.32)  1.53* (1.31, 1.79)  1.03 (0.62, 1.71)  1.32 (0.91, 1.89)   First (most deprived)  1.96* (1.58, 2.44)  1.40* (1.18, 1.66)  1.54* (1.30, 1.82)  1.41* (1.02, 1.97)  2.27* (1.68, 3.09)  1.56* (1.22, 1.99)  1.11 (0.89, 1.38)  1.62** (1.14, 2.29)  aMultivariate analysis based on sample with complete data for outcomes and all covariates. bReference category. OR denotes odds ratio; CI denotes confidence interval. *P < 0.001; **P < 0.05. View Large Table 3 Factors predicting individual risk behaviours amongst all adolescents (n = 96721)a 2002–13   Regular substance use  Heavy substance use  Smoking  Drinking  Drug use  Multiple  Smoking  Drinking  Drug use  Multiple  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Year  0.92* (0.91, 0.94)  0.84* (0.83, 0.85)  0.96* (0.95, 0.97)  0.91* (0.87, 0.93)  0.91* (0.89, 0.93)  1.06** (1.04, 1.07)  1.04** (1.00, 1.08)  1.02 (0.99, 1.05)  Gender   Maleb  1.00                 Female  1.28* (1.19, 1.37)  0.96 (0.92, 1.01)  0.75* (0.69, 0.80)  0.91* (0.89, 0.93)  0.99 (0.88, 1.10)  0.68* (0.63, 0.74)  0.32* (0.25, 0.40)  0.57* (0.48, 0.68)  School year   S2 (13 years old)b  1.00                 S4 (15 years old)  4.95* (4.36, 5.63)  4.49* (4.09, 4.92)  4.62* (4.22, 5.07)  1.01 (0.90, 1.13)  6.15* (5.06, 7.49)  4.02* (3.56, 4.54)  3.61* (2.97, 4.37)  5.98* (4.62, 7.74)  Ethnicity   Scottish/White Britishb  1.00     White Other  1.12 (0.90, 1.39)  0.99 (0.83, 1.18)  1.38* (1.12, 1.69)  1.11 (0.70, 1.78)  0.90 (0.65, 1.26)  1.21 (0.96, 1.53)  1.80** (1.08, 3.00)  1.28 (0.70, 2.34)   Other Ethnicity  1.20** (1.00, 1.43)  1.04 (0.85, 1.27)  1.70* (1.41, 2.06)  1.39* (1.09, 1.77)  1.40** (1.00, 1.96)  1.49* (1.17, 1.90)  5.00* (3.96, 6.30)  2.72* (1.92, 3.85)   Don’t know/refused  2.86* (2.45, 3.35)  2.19* (1.93, 2.48)  1.79* (1.41, 2.27)  2.12* (1.67, 2.69)  4.21* (2.96, 5.98)  2.44* (1.93, 3.10)  6.84* (4.78, 9.79)  3.54* (2.42, 5.17)  Socioeconomic quintile   Fifth (least deprived)b  1.00  1.00  1.00  1.00  1.00  1.00  1.00  1.00   Fourth  1.21* (1.07, 1.37)  1.18* (1.10, 1.28)  1.03 (0.94, 1.13)  0.98 (0.83, 1.16)  0.98 (0.73, 1.32)  1.11 (0.94, 1.30)  0.73 (0.53, 1.01)  0.67** (0.50, 0.89)   Third  1.41* (1.25, 1.59)  1.32* (1.21, 1.45)  1.21* (1.06, 1.39)  1.15 (0.94, 1.39)  1.38** (1.06, 1.78)  1.41* (1.17, 1.70)  0.91 (0.67, 1.25)  1.13 (0.79, 1.61)   Second  1.70* (1.49, 1.94)  1.32* (1.17, 1.49)  1.33* (1.15, 1.53)  1.25* (1.01, 1.54)  1.86* (1.48, 2.32)  1.53* (1.31, 1.79)  1.03 (0.62, 1.71)  1.32 (0.91, 1.89)   First (most deprived)  1.96* (1.58, 2.44)  1.40* (1.18, 1.66)  1.54* (1.30, 1.82)  1.41* (1.02, 1.97)  2.27* (1.68, 3.09)  1.56* (1.22, 1.99)  1.11 (0.89, 1.38)  1.62** (1.14, 2.29)    Regular substance use  Heavy substance use  Smoking  Drinking  Drug use  Multiple  Smoking  Drinking  Drug use  Multiple  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Year  0.92* (0.91, 0.94)  0.84* (0.83, 0.85)  0.96* (0.95, 0.97)  0.91* (0.87, 0.93)  0.91* (0.89, 0.93)  1.06** (1.04, 1.07)  1.04** (1.00, 1.08)  1.02 (0.99, 1.05)  Gender   Maleb  1.00                 Female  1.28* (1.19, 1.37)  0.96 (0.92, 1.01)  0.75* (0.69, 0.80)  0.91* (0.89, 0.93)  0.99 (0.88, 1.10)  0.68* (0.63, 0.74)  0.32* (0.25, 0.40)  0.57* (0.48, 0.68)  School year   S2 (13 years old)b  1.00                 S4 (15 years old)  4.95* (4.36, 5.63)  4.49* (4.09, 4.92)  4.62* (4.22, 5.07)  1.01 (0.90, 1.13)  6.15* (5.06, 7.49)  4.02* (3.56, 4.54)  3.61* (2.97, 4.37)  5.98* (4.62, 7.74)  Ethnicity   Scottish/White Britishb  1.00     White Other  1.12 (0.90, 1.39)  0.99 (0.83, 1.18)  1.38* (1.12, 1.69)  1.11 (0.70, 1.78)  0.90 (0.65, 1.26)  1.21 (0.96, 1.53)  1.80** (1.08, 3.00)  1.28 (0.70, 2.34)   Other Ethnicity  1.20** (1.00, 1.43)  1.04 (0.85, 1.27)  1.70* (1.41, 2.06)  1.39* (1.09, 1.77)  1.40** (1.00, 1.96)  1.49* (1.17, 1.90)  5.00* (3.96, 6.30)  2.72* (1.92, 3.85)   Don’t know/refused  2.86* (2.45, 3.35)  2.19* (1.93, 2.48)  1.79* (1.41, 2.27)  2.12* (1.67, 2.69)  4.21* (2.96, 5.98)  2.44* (1.93, 3.10)  6.84* (4.78, 9.79)  3.54* (2.42, 5.17)  Socioeconomic quintile   Fifth (least deprived)b  1.00  1.00  1.00  1.00  1.00  1.00  1.00  1.00   Fourth  1.21* (1.07, 1.37)  1.18* (1.10, 1.28)  1.03 (0.94, 1.13)  0.98 (0.83, 1.16)  0.98 (0.73, 1.32)  1.11 (0.94, 1.30)  0.73 (0.53, 1.01)  0.67** (0.50, 0.89)   Third  1.41* (1.25, 1.59)  1.32* (1.21, 1.45)  1.21* (1.06, 1.39)  1.15 (0.94, 1.39)  1.38** (1.06, 1.78)  1.41* (1.17, 1.70)  0.91 (0.67, 1.25)  1.13 (0.79, 1.61)   Second  1.70* (1.49, 1.94)  1.32* (1.17, 1.49)  1.33* (1.15, 1.53)  1.25* (1.01, 1.54)  1.86* (1.48, 2.32)  1.53* (1.31, 1.79)  1.03 (0.62, 1.71)  1.32 (0.91, 1.89)   First (most deprived)  1.96* (1.58, 2.44)  1.40* (1.18, 1.66)  1.54* (1.30, 1.82)  1.41* (1.02, 1.97)  2.27* (1.68, 3.09)  1.56* (1.22, 1.99)  1.11 (0.89, 1.38)  1.62** (1.14, 2.29)  aMultivariate analysis based on sample with complete data for outcomes and all covariates. bReference category. OR denotes odds ratio; CI denotes confidence interval. *P < 0.001; **P < 0.05. View Large When sociodemographic risk factors were considered, girls had a significantly increased odds of being a regular smoker than boys (OR = 1.28; 95% CI: 1.19, 1.37), but also had a significantly decreased odds of regular use of illicit drugs (OR = 0.75; 95% CI: 0.69, 0.80) and multiple substances (OR = 0.91; 95% CI: 0.89, 0.93) and heavy use of alcohol (OR = 0.68; 95% CI: 0.63, 0.74), illicit drugs (OR = 0.32; 95% CI: 0.25, 0.40) and multiple substances (OR = 0.57; 95% CI: 0.48, 0.68) (P < 0.001). With the exception of regular multiple substance use, S4 (average age of 15 years) pupils had a significantly increased odds of all forms of regular and heavy substance use than S2 (average age of 13 years) pupils. Ethnically white non-British pupils had a significantly increased odds of being a regular (OR = 1.38; 95% CI: 1.12, 1.69; P < 0.001) and heavy (OR = 1.80; 95% CI: 1.08, 3.00; P < 0.05) illicit drug user than ethnically white British pupils. With the exception of regular alcohol use, non-white pupils and pupils who were unaware or refused to identify their ethnicity had a significantly increased odds of taking all forms of regular and heavy substance use compared to ethnically white British pupils. With regard to socioeconomic status, significantly increased odds of regular smoking and alcohol use were estimated with increasing levels of socioeconomic deprivation. Finally, in comparison to pupils from the least deprived socioeconomic quintile, pupils from the most deprived quintile had increased odds of 1.41 (95% CI: 1.02, 1.97; P < 0.05) and 1.62 (95% CI: 1.14, 2.29; P < 0.05) of being regular and heavy multiple substance users, respectively. Additional analyses The distribution of multivariate logistic regression residuals was centred around zero and approximately normal. Also, the residuals’ correlation with the main exposure variable (survey years) was close to zero (details not shown), suggesting that key assumptions for the parameter estimation had been met. The sensitivity analyses revealed that estimates of temporal trends in substance use remained robust to the incorporation of independent variables related to family circumstances (Appendix 5) and application of more conservative definitions of current regular use (Appendix 6). Discussion Main findings of this study This study revealed that regular smoking, alcohol, illicit drug and multiple substance use declined significantly amongst adolescents in Scotland over the period 2002–13. However, multivariate analyses that focussed upon high-risk levels of these behaviours revealed an upward trend over this time horizon in heavy alcohol and illicit drug use. Sociodemographic patterns within the study data suggest complex gender profiles with girls more likely to be regular smokers, but boys more likely to be use alcohol, illicit drugs and multiple substances in risky ways. Older adolescents were significantly more likely to use individual substances either regularly or in risky ways than younger adolescents. Our results also suggest that non-white pupils and those who were unaware or refused to identify their ethnicity were more likely to be involved in individual and multiple substance use. Furthermore, we observed an association between socioeconomic deprivation and an increased likelihood of being involved in all types of individual and multiple substance use. What is already known on this topic? A number of large cross-sectional surveys have revealed high levels of risky behaviours in adolescents that vary by behaviour and jurisdiction.10–12,14 The findings of this study affirm analyses of national representative data from several industrialized nations, which previously suggested that adolescent substance use has been declining since the turn of the 21st century.13,20,21 Furthermore, the sociodemographic predictors of individual and multiple substance use, by heaviness of use, revealed by this study are broadly consistent with the previous literature.9,13,14,22–24 A number of theoretical and small observational studies have identified socialization, cultural and environmental mechanisms for the initiation and sustenance of adolescent substance use,25–28 but data on these factors are largely absent from national cross-sectional and longitudinal surveys. What this study adds Our study findings provide a transparent, nuanced account of recent declines in the prevalence of regular substance misuse amongst adolescents in Scotland against concerted policy initiatives aimed at its prevention. However, in contrast to recent epidemiological evidence from England, which showed a significant downward linear trend for a combination of risky alcohol use (either heavy regular drinking or binge drinking), regular smoking and regular illicit drug use amongst 11–15 year olds between 1998 and 2009 (OR = 0.90; 95% CI: 0.88, 0.93; P < 0.001),13 we did not observe a decline in the prevalence of multiple heavy substance use. Differences between our results and those from England may be explained by a number of factors including differences in the time horizons of the underpinning data, categorization of individual and multiple exposures, and covariates incorporated into each set of models. Nevertheless, we cannot discount the possibility that patterns of behavioural risk factors differ between adolescents in the two nations. Our findings highlight the need for implementation of effective prevention strategies that particularly target heavy alcohol and heavy illicit drug use amongst adolescents in Scotland. Randomized controlled trials of family-based or school-based interventions aimed at preventing adolescents misusing tobacco,29 alcohol30 or illicit drugs31 have been carried out to good effect. However, less is known about the effectiveness of prevention programmes targeting high-risk behaviours in adolescence, nor about the common antecedents to multiple risk factors that should be the focus of future prevention efforts.32 Moreover, to our knowledge, a feature of all the trials aimed at preventing or alleviating the effects of substance misuse in adolescents is their failure to collect detailed economic information and, therefore, to assess the cost-effectiveness of the interventions. It is imperative that economic evaluations of these interventions are conducted and that resources in this area are allocated in a manner that is both clinically and cost effective. The effects of broader macroeconomic measures affecting prices of substances, and tighter controls around illicit markets, sales practices and enforcement, also remain the basis of future enquiry. Our study also generated subtle differences in sociodemographic predictors of individual and multiple substance use with those observed in England.13 In particular, the English data suggest that girls are at increased risk of multiple substance use whereas our study suggests that boys are at increased risk. In addition, the English data suggest that the prevalence of individual and multiple substance use across years is higher amongst white adolescents whereas our study suggests that they are higher amongst non-white adolescents. This highlights the need for policy responses that are informed by an understanding of localized behavioural patterns. Limitations of this study There are a number of study caveats that should be borne in mind by readers. First, although the overall study population included 134 387 adolescents, the samples for some of the sociodemographic subgroups within some study years were relatively small. Caution is therefore required when drawing conclusions about the sociodemographic risk profiles of adolescent substance misuers in Scotland. Second, the type and degree of substance use was self-reported by adolescents, a method that has previously been shown to only have fair validity when corroborated against biochemical test results.33 Moreover, the definitions of self-reported heavy substance use that we applied have not been widely used in international surveys.10,12,34,35 Third, the categorization of key covariates within our multivariate models, namely ethnicity and socioeconomic status, was driven by the design of the SALSUS questionnaires, and does not reflect the more granulated and personalized approaches to ethnicity and socioeconomic profiling applied in some national36,37 and international38,39 surveys. Fourth, the summary statistics generated by our statistical approaches for estimating time trends in substance use do not fully convey peaks and troughs in prevalence within intermediate years. Fifth, as noted above, a number of socialization, cultural and environmental factors were not collected within SALSUS and were therefore omitted from our analyses. Sixth, the study excluded other health risk behaviours during adolescence, such as early or risky sexual behaviours, which often co-occur with substance use and compound the risk of long-term adverse sequelae.9 Finally, our study does not prove causality between recent policy initiatives introduced by the Scottish Government, or changes in behavioural, inter-personal and social factors, and trends in the prevalence, and patterns, of individual and multiple substance use. Conclusions This study reveals that, in keeping with other nations of the UK, the prevalence of regular individual and multiple substance use amongst adolescents in Scotland has declined since the turn of the 21st century. Of particular concern, however, is the upward trend in heavy alcohol and heavy illicit drug use, which should be the focus of future prevention efforts. Targeted strategies should be informed by the risk profiles of substance misusers and evidence around the clinical and cost-effectiveness of preventive interventions. Supplementary data Supplementary data are available at the Journal of Public Health online. Acknowledgements We would like to thanks participants of the successive ‘Scottish Schools Adolescent Lifestyle and Substance Use Survey (SALSUS)’ annual surveys, conducted by the Information Services Division, NHS National Services Scotland, on behalf of Scottish Government. Funding Stavros Petrou receives support from the UK National Institute for Health Research (NIHR) as a NIHR Senior Investigator. The Warwick Clinical Trials Unit, University of Warwick, benefited from facilities funded through the Birmingham Science City Translational Medicine Clinical Research and Infrastructure Trials Platform, with support from Advantage West Midlands. The views herein expressed are those of the authors and not necessarily those of the funding bodies. Conflicts of interest None to declare. References 1 World Health Organization (WHO). Global Report: Mortality Attributable to Tobacco . Geneva: WHO, 2012. 2 Doll R, Peto R, Boreham J et al.  . Mortality in relation to smoking: 50 years’ observations on male British doctors. Br Med J  2004; 328: 1519. Google Scholar CrossRef Search ADS   3 Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med  2006; 3: e442. Google Scholar CrossRef Search ADS PubMed  4 Lim SS, Vos T, Flaxman AD et al.  . A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet  2012; 380: 2224– 60. Google Scholar CrossRef Search ADS PubMed  5 Rehm J, Shield KD. Global alcohol-attributable deaths from cancer, liver cirrhosis, and injury in 2010. Alcohol Res Curr Rev  2013; 35: 174– 83. 6 Fischbach P. The role of illicit drug use in sudden death in the young. Cardiol Young  2017; 27: S75– S9. Google Scholar CrossRef Search ADS PubMed  7 World Health Organization (WHO). Global Health Risks: Mortality and Burden of Disease Attributable to Selected Major Risks . Geneva: WHO, 2009. 8 Booth-Butterfield M. Embedded health behaviors from adolescence to adulthood: the impact of tobacco. Health Commun  2003; 15: 171– 84. Google Scholar CrossRef Search ADS PubMed  9 Hale DR, Viner RM. The correlates and course of multiple health risk behaviour in adolescence. BMC Public Health  2016; 16: 458. Google Scholar CrossRef Search ADS PubMed  10 Warren CW, Jones NR, Peruga A et al.  . Global youth tobacco surveillance, 2000–2007. MMWR Surveill Summ  2008; 57: 1– 28. Google Scholar PubMed  11 Eaton DK, Kann L, Kinchen S et al.  . Youth risk behavior surveillance—United States, 2011. MMWR Surveill Summ  2012; 61: 1– 162. Google Scholar PubMed  12 Merikangas KR, McClair VL. Epidemiology of substance use disorders. Hum Genet  2012; 131: 779– 89. Google Scholar CrossRef Search ADS PubMed  13 Hale D, Viner R. Trends in the prevalence of multiple substance use in adolescents in England, 1998–2009. J Public Health  2013; 35: 367– 74. Google Scholar CrossRef Search ADS   14 Levin KA, Dundas R, Miller M et al.  . Socioeconomic and geographic inequalities in adolescent smoking: a multilevel cross-sectional study of 15 year olds in Scotland. Soc Sci Med  2014; 107: 162– 70. Google Scholar CrossRef Search ADS PubMed  15 Currie C, Corbett J. Scottish Schools Adolescent Lifestyle and Substance Use Survey (SALSUS) 2004 National Report . Edinburgh: University of Edinburgh, 2005. 16 Black C, Sewel K, Murray L. Scottish Schools Adolescent Lifestyle and Substance Use Survey (SALSUS) 2010 National Report . Edinburgh: University of Edinburgh, 2011. 17 Currie C, Levin K, Todd J. Health Behaviour in School-Aged Children: World Health Organization Collaborative Cross-National Study (HBSC): Findings From the 2010 HBSC Survey in Scotland . Edinburgh: HBSC, 2008. 18 Ralston K, Dundas R, Leyland AH. A comparison of the Scottish Index of Multiple Deprivation (SIMD) 2004 with the 2009 + 1 SIMD: does choice of measure affect the interpretation of inequality in mortality? Int J Health Geogr  2014; 13: 27. Google Scholar CrossRef Search ADS PubMed  19 Sterne JA, White IR, Carlin JB et al.  . Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. Br Med J  2009; 338: b2393. Google Scholar CrossRef Search ADS   20 Ji CY, Hu PJ, Song Y. The epidemiology of alcohol consumption and misuse among Chinese college students. Alcohol Alcohol  2012; 47: 464– 72. Google Scholar CrossRef Search ADS PubMed  21 Bridges S, Gill V, Omole T et al.  . Smoking, Drinking and Drug Use Among Young People in England in 2010 . London: National Centre for Social Research and the National Foundation for Educational Research, 2011. 22 Hibell B, Guttormsson U, Ahlström S et al.  . The 2011 ESPAD Report: Substance Use Among Students in 36 European Countries . Stockholm: ESPAD, 2012. 23 Resnick MD, Bearman PS, Blum RW et al.  . Protecting adolescents from harm. Findings from the National Longitudinal Study on Adolescent Health. J Am Med Assoc  1997; 278: 823– 32. Google Scholar CrossRef Search ADS   24 Sweeting H, Jackson C, Haw S. Changes in the socio-demographic patterning of late adolescent health risk behaviours during the 1990s: analysis of two West of Scotland cohort studies. BMC Public Health  2011; 11: 829. Google Scholar CrossRef Search ADS PubMed  25 Jackson C, Sweeting H, Haw S. Clustering of substance use and sexual risk behaviour in adolescence: analysis of two cohort studies. BMJ Open  2012; 2: e000661. Google Scholar CrossRef Search ADS PubMed  26 Oetting ER, Donnermeyer JF, Trimble JE et al.  . Primary socialization theory: culture, ethnicity, and cultural identification. The links between culture and substance use. IV. Subst Use Misuse  1998; 33: 2075– 107. Google Scholar CrossRef Search ADS PubMed  27 Veitch C. Impact of rurality on environmental determinants and hazards. Aust J Rural Health  2009; 17: 16– 20. Google Scholar CrossRef Search ADS PubMed  28 Kloep M, Hendry LB, Ingebrigtsen JE et al.  . Young people in ‘drinking’ societies? Norwegian, Scottish and Swedish adolescents’ perceptions of alcohol use. Health Educ Res  2001; 16: 279– 91. Google Scholar CrossRef Search ADS PubMed  29 Pearce J, Hiscock R, Moon G et al.  . The neighbourhood effects of geographical access to tobacco retailers on individual smoking behaviour. J Epidemiol Community Health  2009; 63: 69– 77. Google Scholar CrossRef Search ADS PubMed  30 Thomas RE, Baker PR, Thomas BC. Family-based interventions in preventing children and adolescents from using tobacco: a systematic review and meta-analysis. Acad Pediatr  2016; 16: 419– 29. Google Scholar CrossRef Search ADS PubMed  31 Foxcroft DR, Tsertsvadze A. Universal alcohol misuse prevention programmes for children and adolescents: Cochrane systematic reviews. Perspect Public Health  2012; 132: 128– 34. Google Scholar CrossRef Search ADS PubMed  32 Vermeulen-Smit E, Verdurmen JE, Engels RC. The effectiveness of family interventions in preventing adolescent illicit drug use: a systematic review and meta-analysis of randomized controlled trials. Clin Child Fam Psychol Rev  2015; 18: 218– 39. Google Scholar CrossRef Search ADS PubMed  33 Hale DR, Viner RM. Policy responses to multiple risk behaviours in adolescents. J Public Health  2012; 34( Suppl 1): i11– 9. Google Scholar CrossRef Search ADS   34 Williams RJ, Nowatzki N. Validity of adolescent self-report of substance use. Subst Use Misuse  2005; 40: 1– 13. Google Scholar CrossRef Search ADS PubMed  35 Bitancourt T, Tissot MC, Fidalgo TM et al.  . Factors associated with illicit drugs’ lifetime and frequent/heavy use among students results from a population survey. Psychiatry Res  2016; 237: 290– 5. Google Scholar CrossRef Search ADS PubMed  36 Gattamorta KA, Varela A, McCabe BE et al.  . Psychiatric symptoms, parental attachment, and reasons for use as correlates of heavy substance use among treatment-seeking Hispanic adolescents. Subst Use Misuse  2017; 52: 392– 400. Google Scholar CrossRef Search ADS PubMed  37 Burton J, Nandi A, Platt L. Measuring ethnicity: challenges and opportunities for survey research. Ethnic Racial Stud  2010; 33: 1332– 49. Google Scholar CrossRef Search ADS   38 Shavers VL. Measurement of socioeconomic status in health disparities research. J Natl Med Assoc  2007; 99: 1013– 23. Google Scholar PubMed  39 Cheng TL, Goodman E. Committee on Pediatric Research. Race, ethnicity, and socioeconomic status in research on child health. Pediatrics  2015; 135: e225– 37. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Public Health Oxford University Press

Epidemiological trends and risk factors for tobacco, alcohol and drug use among adolescents in Scotland, 2002–13

Loading next page...
 
/lp/ou_press/epidemiological-trends-and-risk-factors-for-tobacco-alcohol-and-drug-CdssmVLKIb
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1741-3842
eISSN
1741-3850
D.O.I.
10.1093/pubmed/fdy006
Publisher site
See Article on Publisher Site

Abstract

Abstract Background This study estimates trends in prevalence, and patterns, of individual and multiple substance use between 2002 and 2013 amongst adolescents in Scotland. Methods The study uses data from 134 387 participants of the biennial national ‘Scottish Schools Adolescent Lifestyle and Substance Use Survey’ on smoking, alcohol and illicit drug use. Current regular use and current heavy use of smoking, alcohol, illicit drugs and multiple substances was measured. Time trends in the prevalence of each outcome were estimated using univariate and multivariate logistic regression. Results Regular smoking, alcohol, illicit drug and multiple substance use declined significantly amongst adolescents in Scotland. However, multivariate analyses that focussed upon high-risk levels of these behaviours revealed an upward linear trend in heavy alcohol (OR = 1.06; 95% CI: 1.04, 1.07) and heavy illicit drug (OR = 1.04; 95% CI: 1.00, 1.08) use (P < 0.05). Non-white pupils were more likely to be involved in individual and multiple substance use than ethnically white British pupils. In comparison to pupils from the least deprived socioeconomic quintile, pupils from the most deprived quintile had increased odds of 1.41 (95% CI: 1.02, 1.97; P < 0.05) and 1.62 (95% CI: 1.14, 2.29; P < 0.05) of being regular and heavy multiple substance users, respectively. Conclusions Further effort is required to tackle heavy alcohol and heavy illicit drug use amongst adolescents in Scotland. Prevention strategies should be informed by the risk profiles of substance misusers and evidence around the clinical and cost-effectiveness of preventive interventions. adolescents, alcohol, illicit drugs, Scotland, tobacco Introduction The adverse sequelae of tobacco, alcohol and illicit drug use are well documented.1–6 The World Health Organization (WHO) has highlighted that understanding the prevalence and role of these behavioural risk factors should play a crucial part in developing clear and effective strategies for improving global health.7 In industrialized nations, initiation of tobacco, alcohol and illicit drug use tends to occur during adolescence, a critical period of life in which risky behaviours often result in embeddedness during the remaining life course.8 Moreover, risky adolescent behaviours such as tobacco, alcohol and illicit drug use often co-occur, which in turn compounds the risk of a host of adverse health, social and economic consequences.9 The prevalence of risky adolescent behaviours during adolescence varies by behaviour and across jurisdictions. Data for the years 2000–07 from 140 WHO member states collected as part of the Global Youth Tobacco Survey revealed that ~9.5% of 13–15 year olds smoked cigarettes, with prevalence ranging from 4.9% in the Eastern Mediterranean Region to 19.2% in the European Region.10 In the United States, surveillance data collected during 2010 and 2011 revealed that 18.1% of high school students in grades 9–12 had smoked cigarettes during the 30 days before the survey, with evidence of higher prevalence amongst male (19.9%) than female (16.1%) students.11 Furthermore, data compiled across national surveys in the United States reveal that more than one half of adolescents in the United States report alcohol use, and nearly one-fourth report exposure to illicit drugs.12 In the United Kingdom (UK), recent evidence based on national surveys suggests a reduction in the prevalence of cigarette smoking, alcohol use and illicit drug use amongst adolescents in England.13 However, close scrutiny of these epidemiological data suggests mixed patterns of multiple substance use amongst adolescents that differ by sociodemographic profile, highlighting areas where future preventive efforts should be targeted.13 In contrast to other industrialized nations, there is a relative paucity of published national epidemiological evidence on substance use amongst adolescents in Scotland. Levin et al.14 analysed national data on 2692 15 year olds included in The Scottish Health Behaviour in School-aged Children Study, conducted between March and June 2010. They found that 13.6% of boys and 18.9% of girls were current smokers, broadly in accordance with data from unpublished reports based on The Scottish Schools Adolescent Lifestyle and Substance Use Survey (SALSUS).15,16 They also found that prevalence of smoking was the highest amongst those living in the second most deprived socioeconomic quintile. Recent Scotland-wide epidemiological data on alcohol and illicit drug use amongst adolescents has, to our knowledge, been restricted to unpublished reports.15–17 Since the Scottish Parliament was established in 1999, the Scottish Government has introduced a number of policy initiatives aimed at tackling substance misuse amongst adolescents. This has included, but not limited to, a ban on tobacco advertising in 2002, an increase in the age for tobacco sales from 16 to 18 years in 2007, a ban on the display of cigarettes for sale in shops and self-service sales from automatic vending machines in 2010, and national frameworks aimed at mitigating the damaging impacts that alcohol and drug misuse have on families and communities, including young people. The objective of this study was to estimate trends in the prevalence, and patterns, of individual and multiple substance use amongst adolescents in Scotland against this policy background. Methods Data sources Data from SALSUS formed the basis of this empirical investigation. SALSUS is a continuation of a series of biennial national surveys on smoking, alcohol and illicit drug use among young people that were carried out jointly in Scotland and England between 1982 and 2000.15 From 2002, Scotland has carried out its own national survey, namely SALSUS, which generates epidemiological data on substance use among adolescents, and provides a vehicle for monitoring progress towards Scottish Government targets on smoking, alcohol and illicit drug use. The SALSUS surveys were carried out biennially between 2002 and 2010 (2002: n = 23 090; 2004: n = 7 062; 2006: n = 23 180; 2008: n = 10 063; 2010: n = 37 307) and subsequently after a three year period in 2013 (n = 33 685). SALSUS takes the form of a confidential, self-completed questionnaire completed by secondary school second year (S2) (average age of 13 years) and secondary school fourth year S4 (average age of 15 years) pupils in school settings. In each survey year, the Scottish Government schools database was used as the sampling frame and included all state funded, grant-maintained and independent secondary schools across the country, but excluded schools dedicated to children with additional support needs. Each survey adopted a multistage sample design that determined the probability of being a selected school, and the probability of being a selected class within that school. With the exception of the 2004 survey, weighting for school type and age group non-response was also applied within local authority strata to ensure that the samples were representative both at a national and a local authority level; the weighting system applied in 2004 ensured that the sample was nationally representative. Consent to participate was provided both by schools and the pupils and their parents. The overall response rate, calculated from the school, class and pupil response rates, varied between 57% in 2006 and 65% in 2002. Detailed methodology for each of the SALSUS surveys, including survey design, sampling strategy, questionnaire design and consent procedures is described in the appendices of the SALSUS annual reports (http://www.scotpho.org.uk/publications/overview-of-key-data-sources/surveys-cross-sectional/scottish-schools-adolescent-lifestyle-a-substance-use-survey). SALSUS data are publicly available and were downloaded from the UK Data Archive on 16 May 2016 (http://www.data-archive.ac.uk/). Outcome measures All substance use measures were self-reported by the adolescents as part of questionnaires completed under exam conditions. Each questionnaire was returned to the class teacher in a sealed envelope without the reporting of names to ensure confidentiality. The outcome variables were defined separately for two levels of substance misuse: current regular use and current heavy use. Current regular smoking was defined as usually smoking one cigarette a week or more; current regular alcohol use was defined as drinking once a week or more on average; and current illicit drug use was defined as having taken any illicit drugs in the last month. These definitions were broadly consistent with those applied in previous studies of adolescent substance use in the UK.13 Current regular multiple substance use was defined as engaging simultaneously in all these behaviours. Using these definitions, data were available for all study years with the exception of 2004. With regards to current heavy substance use, the report of smoking at least 60 cigarettes in last week, drinking at least 21 units of alcohol in the last week and taking illicit drugs most days were considered measures of heavy smoking, alcohol and illicit drug use, respectively. In keeping with the operational definition adopted by SALSUS, heavy multiple substance use was defined as engaging in at least two out of three of these behaviours.15,16 Using these definitions, data were available for all study years with the exception of 2002 for heavy smoking, 2002, 2004 and 2006 for heavy alcohol use, and 2004 for heavy illicit drug use. Sociodemographic data incorporated in the SALSUS surveys included gender (male, female), school year (S2, S4; indicative of age) and ethnicity (Scottish/white British, white other, other ethnicity, do not know/refused to answer). It also included socioeconomic quintile derived from Scottish Index of Multiple Deprivation (SIMD)18 ranks that were themselves derived from postcodes for home addresses reported by the pupils. Socioeconomic data were only available from 2006 onwards. Statistical analysis The prevalence of individual and multiple regular and heavy substance use behaviours was calculated for the total sample in each survey year and separately by gender, school year, ethnicity and socioeconomic quintile within each survey year. Statistical analysis provided a description of the time trends for each outcome measure in two alternative ways: (i) percentage change between the first and last survey year available and its 95% confidence interval (CI), calculated using univariate logistic regression with survey year as the only independent variable considered on a nominal scale; and (ii) annual change adjusted for the independent variables (gender, school year, ethnicity and socioeconomic quintile) using multivariate logistic regression with survey year treated as a continuous independent variable. The rationale for these two different methods of time trend calculation was to evaluate the effects of restricting the trend to be linear on a logarithmic scale with survey year as a continuous predictor variable and of adjusting for the independent variables other than survey year. The univariate logistic regressions were used to calculate the marginal distributions of the survey year estimates (via the post-estimation command ‘margins’ in STATA). Subsequently, the differences between the first versus last survey year estimates were divided by their respective time spans to generate average annual changes. The latter were then multiplied by 100 to generate average annual percentage point changes. In addition, estimates of marginal odds and associated standard errors within the univariate regressions were converted into odds ratios (ORs) and corresponding 95% CIs, and presented graphically as time trends. Interactions between the independent variables were also investigated. In keeping with previous similar studies with large samples,13 missing data were handled by using listwise deletion, except for known confounding effects of non-response such as refusing to answer or answering ‘don’t know’ to the question on respondent’s ethnicity. The latter may be an indicator of social vulnerability (e.g. for immigrants) and correlated with higher risk of substance use. All regression analyses accounted for clustering of respondents within primary sampling units and the varying probability of selecting an individual from the target population. The distribution of multivariate regression residuals was examined to verify whether the assumptions for model estimation had been met. In addition, two types of sensitivity analysis were performed: (i) by adding independent variables related to family circumstances (free school meal entitlement, household composition, maternal and paternal knowledge of children’s activities, amount of pocket money) in order to verify the stability of time trends for substance use; and (ii) by applying more conservative definitions of current regular use of tobacco and alcohol covering the past month rather than the past week in order to enhance comparability with some international surveys.10,19 All analyses were performed using STATA software (StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP). Results Prevalence of substance use Table 1 summarizes the weighted prevalence of adolescent substance use by type of substance and heaviness of use for the overall samples in each survey year. The weighted prevalence of regular smoking, alcohol, illicit drug and multiple substance use declined steadily between 2002 and 2013 from 13.06 to 5.22%, 27.09 to 6.78%, 14.27 to 5.61% and 5.48 to 1.44%, respectively. Uninterrupted declines in regular substance use were also observed in each sociodemographic subgroup with the exception of all ethnicity subgroups for which the prevalence of regular smoking, alcohol and multiple substance use peaked in 2006, and all socioeconomic subgroups for which the prevalence of regular smoking, illicit drug and multiple substance use peaked in 2008 (data not reported). Table 1 Number (%)* of adolescents using substances by type of substance, heaviness of use and year of survey   2002  2004  2006  2008  2010  2013  Total surveyed (N)  23 090  7062  23 180  10 063  37 307  33 685  Regular substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  2 949 (13.06)  –  2 118 (9.67)  891 (9.54)  2 857 (7.80)  1 670 (5.22)   Drinking  6 166 (27.09)  –  4 472 (20.12)  1 579 (16.76)  4 645 (12.90)  2 236 (6.78)   Drug use  3 165 (14.27)  –  1 787 (8.31)  769 (8.13)  2 502 (6.95)  1 849 (5.61)   Multiple substances  1 222 (5.48)  –  719 (3.40)  332 (3.53)  961 (2.59)  476 (1.44)  Heavy substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  –  171 (2.42)  389 (1.89)  339 (3.76)  658 (1.81)  327 (1.08)   Drinking  –  –  –  511 (5.32)  1 587 (4.29)  840 (2.55)   Drug use  416 (1.93)  –  203 (0.95)  85 (0.83)  323 (0.85)  294 (0.88)   Multiple substances  –  –  44 (0.25)  124 (1.32)  410 (1.08)  200 (0.60)    2002  2004  2006  2008  2010  2013  Total surveyed (N)  23 090  7062  23 180  10 063  37 307  33 685  Regular substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  2 949 (13.06)  –  2 118 (9.67)  891 (9.54)  2 857 (7.80)  1 670 (5.22)   Drinking  6 166 (27.09)  –  4 472 (20.12)  1 579 (16.76)  4 645 (12.90)  2 236 (6.78)   Drug use  3 165 (14.27)  –  1 787 (8.31)  769 (8.13)  2 502 (6.95)  1 849 (5.61)   Multiple substances  1 222 (5.48)  –  719 (3.40)  332 (3.53)  961 (2.59)  476 (1.44)  Heavy substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  –  171 (2.42)  389 (1.89)  339 (3.76)  658 (1.81)  327 (1.08)   Drinking  –  –  –  511 (5.32)  1 587 (4.29)  840 (2.55)   Drug use  416 (1.93)  –  203 (0.95)  85 (0.83)  323 (0.85)  294 (0.88)   Multiple substances  –  –  44 (0.25)  124 (1.32)  410 (1.08)  200 (0.60)  *Note: Survey weights were not applied for the number of respondents using a substance (n) and the total surveyed (N). This differed from the respective percentages in parentheses, which were multiplied by survey weights, i.e. (n/N) × (survey weight) ×100. View Large Table 1 Number (%)* of adolescents using substances by type of substance, heaviness of use and year of survey   2002  2004  2006  2008  2010  2013  Total surveyed (N)  23 090  7062  23 180  10 063  37 307  33 685  Regular substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  2 949 (13.06)  –  2 118 (9.67)  891 (9.54)  2 857 (7.80)  1 670 (5.22)   Drinking  6 166 (27.09)  –  4 472 (20.12)  1 579 (16.76)  4 645 (12.90)  2 236 (6.78)   Drug use  3 165 (14.27)  –  1 787 (8.31)  769 (8.13)  2 502 (6.95)  1 849 (5.61)   Multiple substances  1 222 (5.48)  –  719 (3.40)  332 (3.53)  961 (2.59)  476 (1.44)  Heavy substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  –  171 (2.42)  389 (1.89)  339 (3.76)  658 (1.81)  327 (1.08)   Drinking  –  –  –  511 (5.32)  1 587 (4.29)  840 (2.55)   Drug use  416 (1.93)  –  203 (0.95)  85 (0.83)  323 (0.85)  294 (0.88)   Multiple substances  –  –  44 (0.25)  124 (1.32)  410 (1.08)  200 (0.60)    2002  2004  2006  2008  2010  2013  Total surveyed (N)  23 090  7062  23 180  10 063  37 307  33 685  Regular substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  2 949 (13.06)  –  2 118 (9.67)  891 (9.54)  2 857 (7.80)  1 670 (5.22)   Drinking  6 166 (27.09)  –  4 472 (20.12)  1 579 (16.76)  4 645 (12.90)  2 236 (6.78)   Drug use  3 165 (14.27)  –  1 787 (8.31)  769 (8.13)  2 502 (6.95)  1 849 (5.61)   Multiple substances  1 222 (5.48)  –  719 (3.40)  332 (3.53)  961 (2.59)  476 (1.44)  Heavy substance use  n (%)  n (%)  n (%)  n (%)  n (%)  n (%)   Smoking  –  171 (2.42)  389 (1.89)  339 (3.76)  658 (1.81)  327 (1.08)   Drinking  –  –  –  511 (5.32)  1 587 (4.29)  840 (2.55)   Drug use  416 (1.93)  –  203 (0.95)  85 (0.83)  323 (0.85)  294 (0.88)   Multiple substances  –  –  44 (0.25)  124 (1.32)  410 (1.08)  200 (0.60)  *Note: Survey weights were not applied for the number of respondents using a substance (n) and the total surveyed (N). This differed from the respective percentages in parentheses, which were multiplied by survey weights, i.e. (n/N) × (survey weight) ×100. View Large In contrast to the pattern for regular substance use, the weighted prevalence of heavy smoking, alcohol and multiple substance use peaked in 2008 and declined thereafter. Moreover, this pattern was observed for all gender, school year (age), ethnicity and socioeconomic subgroups. Trends in substance use: univariate regressions Table 2 summarizes the results of the univariate logistic regressions estimating time trends in substance use by type of substance and heaviness of use with survey year as the only independent variable. Average annual percentage point declines of 0.65, 1.69, 0.72 and 0.34% were estimated for regular smoking, alcohol, illicit drug and multiple substance use, respectively, between 2002 and 2013 (P < 0.0001). Lower average annual percentage point declines of 0.14, 0.46 and 0.09% were estimated for heavy smoking, alcohol and illicit drug use, respectively, between the first and last years surveyed for each of these substances (P < 0.0001). Notably, however, an average annual percentage point increase of 0.04% was estimated for heavy multiple substance use between the first and last years surveyed (P < 0.0001). Graphical representations of the time trends in probabilities of regular and heavy substance use, by type of substance, are presented in Fig. 1. Table 2 Average annual average percentage point change in substance use between first and last year surveyed; all adolescents Type of substance use  Time span (complete calendar years)a  Annual average % point changeb  SE  95% CI  P-value  Regular substance use   Smoking  12  −0.653  0.007  (−0.535, −0.771)  <0.0001   Drinking  12  −1.692  0.008  (−1.559, −1.825)  <0.0001   Drug use  12  −0.722  0.006  (−0.621, −0.823)  <0.0001   Multiple substances  12  −0.336  0.004  (−0.276, −0.396)  <0.0001  Heavy substance use   Smoking  10  −0.135  0.002  (−0.092, −0.177)  <0.0001   Drinking  6  −0.462  0.003  (−0.367, −0.558)  <0.0001   Drug use  12  −0.087  0.001  (−0.064, −0.111)  <0.0001   Multiple substances  8  0.043  0.001  (0.023, 0.064)  <0.0001  Type of substance use  Time span (complete calendar years)a  Annual average % point changeb  SE  95% CI  P-value  Regular substance use   Smoking  12  −0.653  0.007  (−0.535, −0.771)  <0.0001   Drinking  12  −1.692  0.008  (−1.559, −1.825)  <0.0001   Drug use  12  −0.722  0.006  (−0.621, −0.823)  <0.0001   Multiple substances  12  −0.336  0.004  (−0.276, −0.396)  <0.0001  Heavy substance use   Smoking  10  −0.135  0.002  (−0.092, −0.177)  <0.0001   Drinking  6  −0.462  0.003  (−0.367, −0.558)  <0.0001   Drug use  12  −0.087  0.001  (−0.064, −0.111)  <0.0001   Multiple substances  8  0.043  0.001  (0.023, 0.064)  <0.0001  SE, standard error; CI, confidence interval. aTime span (years) between first and last year surveyed. bAnnual percentage point change in substance use between first and last year surveyed. View Large Table 2 Average annual average percentage point change in substance use between first and last year surveyed; all adolescents Type of substance use  Time span (complete calendar years)a  Annual average % point changeb  SE  95% CI  P-value  Regular substance use   Smoking  12  −0.653  0.007  (−0.535, −0.771)  <0.0001   Drinking  12  −1.692  0.008  (−1.559, −1.825)  <0.0001   Drug use  12  −0.722  0.006  (−0.621, −0.823)  <0.0001   Multiple substances  12  −0.336  0.004  (−0.276, −0.396)  <0.0001  Heavy substance use   Smoking  10  −0.135  0.002  (−0.092, −0.177)  <0.0001   Drinking  6  −0.462  0.003  (−0.367, −0.558)  <0.0001   Drug use  12  −0.087  0.001  (−0.064, −0.111)  <0.0001   Multiple substances  8  0.043  0.001  (0.023, 0.064)  <0.0001  Type of substance use  Time span (complete calendar years)a  Annual average % point changeb  SE  95% CI  P-value  Regular substance use   Smoking  12  −0.653  0.007  (−0.535, −0.771)  <0.0001   Drinking  12  −1.692  0.008  (−1.559, −1.825)  <0.0001   Drug use  12  −0.722  0.006  (−0.621, −0.823)  <0.0001   Multiple substances  12  −0.336  0.004  (−0.276, −0.396)  <0.0001  Heavy substance use   Smoking  10  −0.135  0.002  (−0.092, −0.177)  <0.0001   Drinking  6  −0.462  0.003  (−0.367, −0.558)  <0.0001   Drug use  12  −0.087  0.001  (−0.064, −0.111)  <0.0001   Multiple substances  8  0.043  0.001  (0.023, 0.064)  <0.0001  SE, standard error; CI, confidence interval. aTime span (years) between first and last year surveyed. bAnnual percentage point change in substance use between first and last year surveyed. View Large Fig. 1 View largeDownload slide Time trends in probability of regular and heavy substance use; all adolescents. Solid lines represent regular substance use; dashed lines represent heavy substance use. Fig. 1 View largeDownload slide Time trends in probability of regular and heavy substance use; all adolescents. Solid lines represent regular substance use; dashed lines represent heavy substance use. Appendices 1–4 summarize the results of the univariate logistic regressions estimating time trends in substance use for each type of substance and heaviness of use, by sociodemographic subgroup. For each gender (Appendix 1) and school year (Appendix 2) subgroup, there were statistically significant average annual percentage point decreases in regular smoking, alcohol, illicit drug and multiple substance use and in heavy smoking, alcohol and illicit drug use, but also a statistically significant increase in heavy multiple substance use. A similar pattern was observed for the ethnicity subgroups with the exception of regular smoking, for which a non-statistically significant decline was estimated in all the ethnicity subgroups, and heavy illicit drug use, for which an increase in use was estimated in all the ethnicity subgroups (Appendix 3). Similarly, when the analyses were replicated by socioeconomic quintile, a temporal increase in heavy drug use was estimated within each quintile, although the estimated average annual percentage point increase was only statistically significant (P = 0.039) in the most deprived socioeconomic quintile (Appendix 4). Trends and patterns in substance use: multivariate regressions The multivariate regressions revealed significant downward linear trends over time for all forms of regular substance use: ORs (95% CIs) of 0.92 (0.91, 0.94), 0.84 (0.83, 0.85), 0.96 (0.95, 0.97) and 0.91 (0.87, 0.93) for regular smoking, alcohol, illicit drug and multiple substance use, respectively (P < 0.001) (Table 3). In contrast, the multivariate regressions revealed significant upward linear trends over time for heavy alcohol (OR = 1.06; 95% CI: 1.04, 1.07) and illicit drug (OR = 1.04; 95% CI: 1.00, 1.08) use (P < 0.05). Table 3 Factors predicting individual risk behaviours amongst all adolescents (n = 96721)a 2002–13   Regular substance use  Heavy substance use  Smoking  Drinking  Drug use  Multiple  Smoking  Drinking  Drug use  Multiple  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Year  0.92* (0.91, 0.94)  0.84* (0.83, 0.85)  0.96* (0.95, 0.97)  0.91* (0.87, 0.93)  0.91* (0.89, 0.93)  1.06** (1.04, 1.07)  1.04** (1.00, 1.08)  1.02 (0.99, 1.05)  Gender   Maleb  1.00                 Female  1.28* (1.19, 1.37)  0.96 (0.92, 1.01)  0.75* (0.69, 0.80)  0.91* (0.89, 0.93)  0.99 (0.88, 1.10)  0.68* (0.63, 0.74)  0.32* (0.25, 0.40)  0.57* (0.48, 0.68)  School year   S2 (13 years old)b  1.00                 S4 (15 years old)  4.95* (4.36, 5.63)  4.49* (4.09, 4.92)  4.62* (4.22, 5.07)  1.01 (0.90, 1.13)  6.15* (5.06, 7.49)  4.02* (3.56, 4.54)  3.61* (2.97, 4.37)  5.98* (4.62, 7.74)  Ethnicity   Scottish/White Britishb  1.00     White Other  1.12 (0.90, 1.39)  0.99 (0.83, 1.18)  1.38* (1.12, 1.69)  1.11 (0.70, 1.78)  0.90 (0.65, 1.26)  1.21 (0.96, 1.53)  1.80** (1.08, 3.00)  1.28 (0.70, 2.34)   Other Ethnicity  1.20** (1.00, 1.43)  1.04 (0.85, 1.27)  1.70* (1.41, 2.06)  1.39* (1.09, 1.77)  1.40** (1.00, 1.96)  1.49* (1.17, 1.90)  5.00* (3.96, 6.30)  2.72* (1.92, 3.85)   Don’t know/refused  2.86* (2.45, 3.35)  2.19* (1.93, 2.48)  1.79* (1.41, 2.27)  2.12* (1.67, 2.69)  4.21* (2.96, 5.98)  2.44* (1.93, 3.10)  6.84* (4.78, 9.79)  3.54* (2.42, 5.17)  Socioeconomic quintile   Fifth (least deprived)b  1.00  1.00  1.00  1.00  1.00  1.00  1.00  1.00   Fourth  1.21* (1.07, 1.37)  1.18* (1.10, 1.28)  1.03 (0.94, 1.13)  0.98 (0.83, 1.16)  0.98 (0.73, 1.32)  1.11 (0.94, 1.30)  0.73 (0.53, 1.01)  0.67** (0.50, 0.89)   Third  1.41* (1.25, 1.59)  1.32* (1.21, 1.45)  1.21* (1.06, 1.39)  1.15 (0.94, 1.39)  1.38** (1.06, 1.78)  1.41* (1.17, 1.70)  0.91 (0.67, 1.25)  1.13 (0.79, 1.61)   Second  1.70* (1.49, 1.94)  1.32* (1.17, 1.49)  1.33* (1.15, 1.53)  1.25* (1.01, 1.54)  1.86* (1.48, 2.32)  1.53* (1.31, 1.79)  1.03 (0.62, 1.71)  1.32 (0.91, 1.89)   First (most deprived)  1.96* (1.58, 2.44)  1.40* (1.18, 1.66)  1.54* (1.30, 1.82)  1.41* (1.02, 1.97)  2.27* (1.68, 3.09)  1.56* (1.22, 1.99)  1.11 (0.89, 1.38)  1.62** (1.14, 2.29)    Regular substance use  Heavy substance use  Smoking  Drinking  Drug use  Multiple  Smoking  Drinking  Drug use  Multiple  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Year  0.92* (0.91, 0.94)  0.84* (0.83, 0.85)  0.96* (0.95, 0.97)  0.91* (0.87, 0.93)  0.91* (0.89, 0.93)  1.06** (1.04, 1.07)  1.04** (1.00, 1.08)  1.02 (0.99, 1.05)  Gender   Maleb  1.00                 Female  1.28* (1.19, 1.37)  0.96 (0.92, 1.01)  0.75* (0.69, 0.80)  0.91* (0.89, 0.93)  0.99 (0.88, 1.10)  0.68* (0.63, 0.74)  0.32* (0.25, 0.40)  0.57* (0.48, 0.68)  School year   S2 (13 years old)b  1.00                 S4 (15 years old)  4.95* (4.36, 5.63)  4.49* (4.09, 4.92)  4.62* (4.22, 5.07)  1.01 (0.90, 1.13)  6.15* (5.06, 7.49)  4.02* (3.56, 4.54)  3.61* (2.97, 4.37)  5.98* (4.62, 7.74)  Ethnicity   Scottish/White Britishb  1.00     White Other  1.12 (0.90, 1.39)  0.99 (0.83, 1.18)  1.38* (1.12, 1.69)  1.11 (0.70, 1.78)  0.90 (0.65, 1.26)  1.21 (0.96, 1.53)  1.80** (1.08, 3.00)  1.28 (0.70, 2.34)   Other Ethnicity  1.20** (1.00, 1.43)  1.04 (0.85, 1.27)  1.70* (1.41, 2.06)  1.39* (1.09, 1.77)  1.40** (1.00, 1.96)  1.49* (1.17, 1.90)  5.00* (3.96, 6.30)  2.72* (1.92, 3.85)   Don’t know/refused  2.86* (2.45, 3.35)  2.19* (1.93, 2.48)  1.79* (1.41, 2.27)  2.12* (1.67, 2.69)  4.21* (2.96, 5.98)  2.44* (1.93, 3.10)  6.84* (4.78, 9.79)  3.54* (2.42, 5.17)  Socioeconomic quintile   Fifth (least deprived)b  1.00  1.00  1.00  1.00  1.00  1.00  1.00  1.00   Fourth  1.21* (1.07, 1.37)  1.18* (1.10, 1.28)  1.03 (0.94, 1.13)  0.98 (0.83, 1.16)  0.98 (0.73, 1.32)  1.11 (0.94, 1.30)  0.73 (0.53, 1.01)  0.67** (0.50, 0.89)   Third  1.41* (1.25, 1.59)  1.32* (1.21, 1.45)  1.21* (1.06, 1.39)  1.15 (0.94, 1.39)  1.38** (1.06, 1.78)  1.41* (1.17, 1.70)  0.91 (0.67, 1.25)  1.13 (0.79, 1.61)   Second  1.70* (1.49, 1.94)  1.32* (1.17, 1.49)  1.33* (1.15, 1.53)  1.25* (1.01, 1.54)  1.86* (1.48, 2.32)  1.53* (1.31, 1.79)  1.03 (0.62, 1.71)  1.32 (0.91, 1.89)   First (most deprived)  1.96* (1.58, 2.44)  1.40* (1.18, 1.66)  1.54* (1.30, 1.82)  1.41* (1.02, 1.97)  2.27* (1.68, 3.09)  1.56* (1.22, 1.99)  1.11 (0.89, 1.38)  1.62** (1.14, 2.29)  aMultivariate analysis based on sample with complete data for outcomes and all covariates. bReference category. OR denotes odds ratio; CI denotes confidence interval. *P < 0.001; **P < 0.05. View Large Table 3 Factors predicting individual risk behaviours amongst all adolescents (n = 96721)a 2002–13   Regular substance use  Heavy substance use  Smoking  Drinking  Drug use  Multiple  Smoking  Drinking  Drug use  Multiple  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Year  0.92* (0.91, 0.94)  0.84* (0.83, 0.85)  0.96* (0.95, 0.97)  0.91* (0.87, 0.93)  0.91* (0.89, 0.93)  1.06** (1.04, 1.07)  1.04** (1.00, 1.08)  1.02 (0.99, 1.05)  Gender   Maleb  1.00                 Female  1.28* (1.19, 1.37)  0.96 (0.92, 1.01)  0.75* (0.69, 0.80)  0.91* (0.89, 0.93)  0.99 (0.88, 1.10)  0.68* (0.63, 0.74)  0.32* (0.25, 0.40)  0.57* (0.48, 0.68)  School year   S2 (13 years old)b  1.00                 S4 (15 years old)  4.95* (4.36, 5.63)  4.49* (4.09, 4.92)  4.62* (4.22, 5.07)  1.01 (0.90, 1.13)  6.15* (5.06, 7.49)  4.02* (3.56, 4.54)  3.61* (2.97, 4.37)  5.98* (4.62, 7.74)  Ethnicity   Scottish/White Britishb  1.00     White Other  1.12 (0.90, 1.39)  0.99 (0.83, 1.18)  1.38* (1.12, 1.69)  1.11 (0.70, 1.78)  0.90 (0.65, 1.26)  1.21 (0.96, 1.53)  1.80** (1.08, 3.00)  1.28 (0.70, 2.34)   Other Ethnicity  1.20** (1.00, 1.43)  1.04 (0.85, 1.27)  1.70* (1.41, 2.06)  1.39* (1.09, 1.77)  1.40** (1.00, 1.96)  1.49* (1.17, 1.90)  5.00* (3.96, 6.30)  2.72* (1.92, 3.85)   Don’t know/refused  2.86* (2.45, 3.35)  2.19* (1.93, 2.48)  1.79* (1.41, 2.27)  2.12* (1.67, 2.69)  4.21* (2.96, 5.98)  2.44* (1.93, 3.10)  6.84* (4.78, 9.79)  3.54* (2.42, 5.17)  Socioeconomic quintile   Fifth (least deprived)b  1.00  1.00  1.00  1.00  1.00  1.00  1.00  1.00   Fourth  1.21* (1.07, 1.37)  1.18* (1.10, 1.28)  1.03 (0.94, 1.13)  0.98 (0.83, 1.16)  0.98 (0.73, 1.32)  1.11 (0.94, 1.30)  0.73 (0.53, 1.01)  0.67** (0.50, 0.89)   Third  1.41* (1.25, 1.59)  1.32* (1.21, 1.45)  1.21* (1.06, 1.39)  1.15 (0.94, 1.39)  1.38** (1.06, 1.78)  1.41* (1.17, 1.70)  0.91 (0.67, 1.25)  1.13 (0.79, 1.61)   Second  1.70* (1.49, 1.94)  1.32* (1.17, 1.49)  1.33* (1.15, 1.53)  1.25* (1.01, 1.54)  1.86* (1.48, 2.32)  1.53* (1.31, 1.79)  1.03 (0.62, 1.71)  1.32 (0.91, 1.89)   First (most deprived)  1.96* (1.58, 2.44)  1.40* (1.18, 1.66)  1.54* (1.30, 1.82)  1.41* (1.02, 1.97)  2.27* (1.68, 3.09)  1.56* (1.22, 1.99)  1.11 (0.89, 1.38)  1.62** (1.14, 2.29)    Regular substance use  Heavy substance use  Smoking  Drinking  Drug use  Multiple  Smoking  Drinking  Drug use  Multiple  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  OR (95% CI)  Year  0.92* (0.91, 0.94)  0.84* (0.83, 0.85)  0.96* (0.95, 0.97)  0.91* (0.87, 0.93)  0.91* (0.89, 0.93)  1.06** (1.04, 1.07)  1.04** (1.00, 1.08)  1.02 (0.99, 1.05)  Gender   Maleb  1.00                 Female  1.28* (1.19, 1.37)  0.96 (0.92, 1.01)  0.75* (0.69, 0.80)  0.91* (0.89, 0.93)  0.99 (0.88, 1.10)  0.68* (0.63, 0.74)  0.32* (0.25, 0.40)  0.57* (0.48, 0.68)  School year   S2 (13 years old)b  1.00                 S4 (15 years old)  4.95* (4.36, 5.63)  4.49* (4.09, 4.92)  4.62* (4.22, 5.07)  1.01 (0.90, 1.13)  6.15* (5.06, 7.49)  4.02* (3.56, 4.54)  3.61* (2.97, 4.37)  5.98* (4.62, 7.74)  Ethnicity   Scottish/White Britishb  1.00     White Other  1.12 (0.90, 1.39)  0.99 (0.83, 1.18)  1.38* (1.12, 1.69)  1.11 (0.70, 1.78)  0.90 (0.65, 1.26)  1.21 (0.96, 1.53)  1.80** (1.08, 3.00)  1.28 (0.70, 2.34)   Other Ethnicity  1.20** (1.00, 1.43)  1.04 (0.85, 1.27)  1.70* (1.41, 2.06)  1.39* (1.09, 1.77)  1.40** (1.00, 1.96)  1.49* (1.17, 1.90)  5.00* (3.96, 6.30)  2.72* (1.92, 3.85)   Don’t know/refused  2.86* (2.45, 3.35)  2.19* (1.93, 2.48)  1.79* (1.41, 2.27)  2.12* (1.67, 2.69)  4.21* (2.96, 5.98)  2.44* (1.93, 3.10)  6.84* (4.78, 9.79)  3.54* (2.42, 5.17)  Socioeconomic quintile   Fifth (least deprived)b  1.00  1.00  1.00  1.00  1.00  1.00  1.00  1.00   Fourth  1.21* (1.07, 1.37)  1.18* (1.10, 1.28)  1.03 (0.94, 1.13)  0.98 (0.83, 1.16)  0.98 (0.73, 1.32)  1.11 (0.94, 1.30)  0.73 (0.53, 1.01)  0.67** (0.50, 0.89)   Third  1.41* (1.25, 1.59)  1.32* (1.21, 1.45)  1.21* (1.06, 1.39)  1.15 (0.94, 1.39)  1.38** (1.06, 1.78)  1.41* (1.17, 1.70)  0.91 (0.67, 1.25)  1.13 (0.79, 1.61)   Second  1.70* (1.49, 1.94)  1.32* (1.17, 1.49)  1.33* (1.15, 1.53)  1.25* (1.01, 1.54)  1.86* (1.48, 2.32)  1.53* (1.31, 1.79)  1.03 (0.62, 1.71)  1.32 (0.91, 1.89)   First (most deprived)  1.96* (1.58, 2.44)  1.40* (1.18, 1.66)  1.54* (1.30, 1.82)  1.41* (1.02, 1.97)  2.27* (1.68, 3.09)  1.56* (1.22, 1.99)  1.11 (0.89, 1.38)  1.62** (1.14, 2.29)  aMultivariate analysis based on sample with complete data for outcomes and all covariates. bReference category. OR denotes odds ratio; CI denotes confidence interval. *P < 0.001; **P < 0.05. View Large When sociodemographic risk factors were considered, girls had a significantly increased odds of being a regular smoker than boys (OR = 1.28; 95% CI: 1.19, 1.37), but also had a significantly decreased odds of regular use of illicit drugs (OR = 0.75; 95% CI: 0.69, 0.80) and multiple substances (OR = 0.91; 95% CI: 0.89, 0.93) and heavy use of alcohol (OR = 0.68; 95% CI: 0.63, 0.74), illicit drugs (OR = 0.32; 95% CI: 0.25, 0.40) and multiple substances (OR = 0.57; 95% CI: 0.48, 0.68) (P < 0.001). With the exception of regular multiple substance use, S4 (average age of 15 years) pupils had a significantly increased odds of all forms of regular and heavy substance use than S2 (average age of 13 years) pupils. Ethnically white non-British pupils had a significantly increased odds of being a regular (OR = 1.38; 95% CI: 1.12, 1.69; P < 0.001) and heavy (OR = 1.80; 95% CI: 1.08, 3.00; P < 0.05) illicit drug user than ethnically white British pupils. With the exception of regular alcohol use, non-white pupils and pupils who were unaware or refused to identify their ethnicity had a significantly increased odds of taking all forms of regular and heavy substance use compared to ethnically white British pupils. With regard to socioeconomic status, significantly increased odds of regular smoking and alcohol use were estimated with increasing levels of socioeconomic deprivation. Finally, in comparison to pupils from the least deprived socioeconomic quintile, pupils from the most deprived quintile had increased odds of 1.41 (95% CI: 1.02, 1.97; P < 0.05) and 1.62 (95% CI: 1.14, 2.29; P < 0.05) of being regular and heavy multiple substance users, respectively. Additional analyses The distribution of multivariate logistic regression residuals was centred around zero and approximately normal. Also, the residuals’ correlation with the main exposure variable (survey years) was close to zero (details not shown), suggesting that key assumptions for the parameter estimation had been met. The sensitivity analyses revealed that estimates of temporal trends in substance use remained robust to the incorporation of independent variables related to family circumstances (Appendix 5) and application of more conservative definitions of current regular use (Appendix 6). Discussion Main findings of this study This study revealed that regular smoking, alcohol, illicit drug and multiple substance use declined significantly amongst adolescents in Scotland over the period 2002–13. However, multivariate analyses that focussed upon high-risk levels of these behaviours revealed an upward trend over this time horizon in heavy alcohol and illicit drug use. Sociodemographic patterns within the study data suggest complex gender profiles with girls more likely to be regular smokers, but boys more likely to be use alcohol, illicit drugs and multiple substances in risky ways. Older adolescents were significantly more likely to use individual substances either regularly or in risky ways than younger adolescents. Our results also suggest that non-white pupils and those who were unaware or refused to identify their ethnicity were more likely to be involved in individual and multiple substance use. Furthermore, we observed an association between socioeconomic deprivation and an increased likelihood of being involved in all types of individual and multiple substance use. What is already known on this topic? A number of large cross-sectional surveys have revealed high levels of risky behaviours in adolescents that vary by behaviour and jurisdiction.10–12,14 The findings of this study affirm analyses of national representative data from several industrialized nations, which previously suggested that adolescent substance use has been declining since the turn of the 21st century.13,20,21 Furthermore, the sociodemographic predictors of individual and multiple substance use, by heaviness of use, revealed by this study are broadly consistent with the previous literature.9,13,14,22–24 A number of theoretical and small observational studies have identified socialization, cultural and environmental mechanisms for the initiation and sustenance of adolescent substance use,25–28 but data on these factors are largely absent from national cross-sectional and longitudinal surveys. What this study adds Our study findings provide a transparent, nuanced account of recent declines in the prevalence of regular substance misuse amongst adolescents in Scotland against concerted policy initiatives aimed at its prevention. However, in contrast to recent epidemiological evidence from England, which showed a significant downward linear trend for a combination of risky alcohol use (either heavy regular drinking or binge drinking), regular smoking and regular illicit drug use amongst 11–15 year olds between 1998 and 2009 (OR = 0.90; 95% CI: 0.88, 0.93; P < 0.001),13 we did not observe a decline in the prevalence of multiple heavy substance use. Differences between our results and those from England may be explained by a number of factors including differences in the time horizons of the underpinning data, categorization of individual and multiple exposures, and covariates incorporated into each set of models. Nevertheless, we cannot discount the possibility that patterns of behavioural risk factors differ between adolescents in the two nations. Our findings highlight the need for implementation of effective prevention strategies that particularly target heavy alcohol and heavy illicit drug use amongst adolescents in Scotland. Randomized controlled trials of family-based or school-based interventions aimed at preventing adolescents misusing tobacco,29 alcohol30 or illicit drugs31 have been carried out to good effect. However, less is known about the effectiveness of prevention programmes targeting high-risk behaviours in adolescence, nor about the common antecedents to multiple risk factors that should be the focus of future prevention efforts.32 Moreover, to our knowledge, a feature of all the trials aimed at preventing or alleviating the effects of substance misuse in adolescents is their failure to collect detailed economic information and, therefore, to assess the cost-effectiveness of the interventions. It is imperative that economic evaluations of these interventions are conducted and that resources in this area are allocated in a manner that is both clinically and cost effective. The effects of broader macroeconomic measures affecting prices of substances, and tighter controls around illicit markets, sales practices and enforcement, also remain the basis of future enquiry. Our study also generated subtle differences in sociodemographic predictors of individual and multiple substance use with those observed in England.13 In particular, the English data suggest that girls are at increased risk of multiple substance use whereas our study suggests that boys are at increased risk. In addition, the English data suggest that the prevalence of individual and multiple substance use across years is higher amongst white adolescents whereas our study suggests that they are higher amongst non-white adolescents. This highlights the need for policy responses that are informed by an understanding of localized behavioural patterns. Limitations of this study There are a number of study caveats that should be borne in mind by readers. First, although the overall study population included 134 387 adolescents, the samples for some of the sociodemographic subgroups within some study years were relatively small. Caution is therefore required when drawing conclusions about the sociodemographic risk profiles of adolescent substance misuers in Scotland. Second, the type and degree of substance use was self-reported by adolescents, a method that has previously been shown to only have fair validity when corroborated against biochemical test results.33 Moreover, the definitions of self-reported heavy substance use that we applied have not been widely used in international surveys.10,12,34,35 Third, the categorization of key covariates within our multivariate models, namely ethnicity and socioeconomic status, was driven by the design of the SALSUS questionnaires, and does not reflect the more granulated and personalized approaches to ethnicity and socioeconomic profiling applied in some national36,37 and international38,39 surveys. Fourth, the summary statistics generated by our statistical approaches for estimating time trends in substance use do not fully convey peaks and troughs in prevalence within intermediate years. Fifth, as noted above, a number of socialization, cultural and environmental factors were not collected within SALSUS and were therefore omitted from our analyses. Sixth, the study excluded other health risk behaviours during adolescence, such as early or risky sexual behaviours, which often co-occur with substance use and compound the risk of long-term adverse sequelae.9 Finally, our study does not prove causality between recent policy initiatives introduced by the Scottish Government, or changes in behavioural, inter-personal and social factors, and trends in the prevalence, and patterns, of individual and multiple substance use. Conclusions This study reveals that, in keeping with other nations of the UK, the prevalence of regular individual and multiple substance use amongst adolescents in Scotland has declined since the turn of the 21st century. Of particular concern, however, is the upward trend in heavy alcohol and heavy illicit drug use, which should be the focus of future prevention efforts. Targeted strategies should be informed by the risk profiles of substance misusers and evidence around the clinical and cost-effectiveness of preventive interventions. Supplementary data Supplementary data are available at the Journal of Public Health online. Acknowledgements We would like to thanks participants of the successive ‘Scottish Schools Adolescent Lifestyle and Substance Use Survey (SALSUS)’ annual surveys, conducted by the Information Services Division, NHS National Services Scotland, on behalf of Scottish Government. Funding Stavros Petrou receives support from the UK National Institute for Health Research (NIHR) as a NIHR Senior Investigator. The Warwick Clinical Trials Unit, University of Warwick, benefited from facilities funded through the Birmingham Science City Translational Medicine Clinical Research and Infrastructure Trials Platform, with support from Advantage West Midlands. The views herein expressed are those of the authors and not necessarily those of the funding bodies. Conflicts of interest None to declare. References 1 World Health Organization (WHO). Global Report: Mortality Attributable to Tobacco . Geneva: WHO, 2012. 2 Doll R, Peto R, Boreham J et al.  . Mortality in relation to smoking: 50 years’ observations on male British doctors. Br Med J  2004; 328: 1519. Google Scholar CrossRef Search ADS   3 Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med  2006; 3: e442. Google Scholar CrossRef Search ADS PubMed  4 Lim SS, Vos T, Flaxman AD et al.  . A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet  2012; 380: 2224– 60. Google Scholar CrossRef Search ADS PubMed  5 Rehm J, Shield KD. Global alcohol-attributable deaths from cancer, liver cirrhosis, and injury in 2010. Alcohol Res Curr Rev  2013; 35: 174– 83. 6 Fischbach P. The role of illicit drug use in sudden death in the young. Cardiol Young  2017; 27: S75– S9. Google Scholar CrossRef Search ADS PubMed  7 World Health Organization (WHO). Global Health Risks: Mortality and Burden of Disease Attributable to Selected Major Risks . Geneva: WHO, 2009. 8 Booth-Butterfield M. Embedded health behaviors from adolescence to adulthood: the impact of tobacco. Health Commun  2003; 15: 171– 84. Google Scholar CrossRef Search ADS PubMed  9 Hale DR, Viner RM. The correlates and course of multiple health risk behaviour in adolescence. BMC Public Health  2016; 16: 458. Google Scholar CrossRef Search ADS PubMed  10 Warren CW, Jones NR, Peruga A et al.  . Global youth tobacco surveillance, 2000–2007. MMWR Surveill Summ  2008; 57: 1– 28. Google Scholar PubMed  11 Eaton DK, Kann L, Kinchen S et al.  . Youth risk behavior surveillance—United States, 2011. MMWR Surveill Summ  2012; 61: 1– 162. Google Scholar PubMed  12 Merikangas KR, McClair VL. Epidemiology of substance use disorders. Hum Genet  2012; 131: 779– 89. Google Scholar CrossRef Search ADS PubMed  13 Hale D, Viner R. Trends in the prevalence of multiple substance use in adolescents in England, 1998–2009. J Public Health  2013; 35: 367– 74. Google Scholar CrossRef Search ADS   14 Levin KA, Dundas R, Miller M et al.  . Socioeconomic and geographic inequalities in adolescent smoking: a multilevel cross-sectional study of 15 year olds in Scotland. Soc Sci Med  2014; 107: 162– 70. Google Scholar CrossRef Search ADS PubMed  15 Currie C, Corbett J. Scottish Schools Adolescent Lifestyle and Substance Use Survey (SALSUS) 2004 National Report . Edinburgh: University of Edinburgh, 2005. 16 Black C, Sewel K, Murray L. Scottish Schools Adolescent Lifestyle and Substance Use Survey (SALSUS) 2010 National Report . Edinburgh: University of Edinburgh, 2011. 17 Currie C, Levin K, Todd J. Health Behaviour in School-Aged Children: World Health Organization Collaborative Cross-National Study (HBSC): Findings From the 2010 HBSC Survey in Scotland . Edinburgh: HBSC, 2008. 18 Ralston K, Dundas R, Leyland AH. A comparison of the Scottish Index of Multiple Deprivation (SIMD) 2004 with the 2009 + 1 SIMD: does choice of measure affect the interpretation of inequality in mortality? Int J Health Geogr  2014; 13: 27. Google Scholar CrossRef Search ADS PubMed  19 Sterne JA, White IR, Carlin JB et al.  . Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. Br Med J  2009; 338: b2393. Google Scholar CrossRef Search ADS   20 Ji CY, Hu PJ, Song Y. The epidemiology of alcohol consumption and misuse among Chinese college students. Alcohol Alcohol  2012; 47: 464– 72. Google Scholar CrossRef Search ADS PubMed  21 Bridges S, Gill V, Omole T et al.  . Smoking, Drinking and Drug Use Among Young People in England in 2010 . London: National Centre for Social Research and the National Foundation for Educational Research, 2011. 22 Hibell B, Guttormsson U, Ahlström S et al.  . The 2011 ESPAD Report: Substance Use Among Students in 36 European Countries . Stockholm: ESPAD, 2012. 23 Resnick MD, Bearman PS, Blum RW et al.  . Protecting adolescents from harm. Findings from the National Longitudinal Study on Adolescent Health. J Am Med Assoc  1997; 278: 823– 32. Google Scholar CrossRef Search ADS   24 Sweeting H, Jackson C, Haw S. Changes in the socio-demographic patterning of late adolescent health risk behaviours during the 1990s: analysis of two West of Scotland cohort studies. BMC Public Health  2011; 11: 829. Google Scholar CrossRef Search ADS PubMed  25 Jackson C, Sweeting H, Haw S. Clustering of substance use and sexual risk behaviour in adolescence: analysis of two cohort studies. BMJ Open  2012; 2: e000661. Google Scholar CrossRef Search ADS PubMed  26 Oetting ER, Donnermeyer JF, Trimble JE et al.  . Primary socialization theory: culture, ethnicity, and cultural identification. The links between culture and substance use. IV. Subst Use Misuse  1998; 33: 2075– 107. Google Scholar CrossRef Search ADS PubMed  27 Veitch C. Impact of rurality on environmental determinants and hazards. Aust J Rural Health  2009; 17: 16– 20. Google Scholar CrossRef Search ADS PubMed  28 Kloep M, Hendry LB, Ingebrigtsen JE et al.  . Young people in ‘drinking’ societies? Norwegian, Scottish and Swedish adolescents’ perceptions of alcohol use. Health Educ Res  2001; 16: 279– 91. Google Scholar CrossRef Search ADS PubMed  29 Pearce J, Hiscock R, Moon G et al.  . The neighbourhood effects of geographical access to tobacco retailers on individual smoking behaviour. J Epidemiol Community Health  2009; 63: 69– 77. Google Scholar CrossRef Search ADS PubMed  30 Thomas RE, Baker PR, Thomas BC. Family-based interventions in preventing children and adolescents from using tobacco: a systematic review and meta-analysis. Acad Pediatr  2016; 16: 419– 29. Google Scholar CrossRef Search ADS PubMed  31 Foxcroft DR, Tsertsvadze A. Universal alcohol misuse prevention programmes for children and adolescents: Cochrane systematic reviews. Perspect Public Health  2012; 132: 128– 34. Google Scholar CrossRef Search ADS PubMed  32 Vermeulen-Smit E, Verdurmen JE, Engels RC. The effectiveness of family interventions in preventing adolescent illicit drug use: a systematic review and meta-analysis of randomized controlled trials. Clin Child Fam Psychol Rev  2015; 18: 218– 39. Google Scholar CrossRef Search ADS PubMed  33 Hale DR, Viner RM. Policy responses to multiple risk behaviours in adolescents. J Public Health  2012; 34( Suppl 1): i11– 9. Google Scholar CrossRef Search ADS   34 Williams RJ, Nowatzki N. Validity of adolescent self-report of substance use. Subst Use Misuse  2005; 40: 1– 13. Google Scholar CrossRef Search ADS PubMed  35 Bitancourt T, Tissot MC, Fidalgo TM et al.  . Factors associated with illicit drugs’ lifetime and frequent/heavy use among students results from a population survey. Psychiatry Res  2016; 237: 290– 5. Google Scholar CrossRef Search ADS PubMed  36 Gattamorta KA, Varela A, McCabe BE et al.  . Psychiatric symptoms, parental attachment, and reasons for use as correlates of heavy substance use among treatment-seeking Hispanic adolescents. Subst Use Misuse  2017; 52: 392– 400. Google Scholar CrossRef Search ADS PubMed  37 Burton J, Nandi A, Platt L. Measuring ethnicity: challenges and opportunities for survey research. Ethnic Racial Stud  2010; 33: 1332– 49. Google Scholar CrossRef Search ADS   38 Shavers VL. Measurement of socioeconomic status in health disparities research. J Natl Med Assoc  2007; 99: 1013– 23. Google Scholar PubMed  39 Cheng TL, Goodman E. Committee on Pediatric Research. Race, ethnicity, and socioeconomic status in research on child health. Pediatrics  2015; 135: e225– 37. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Journal of Public HealthOxford University Press

Published: Feb 2, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off