Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Polydrug use in Australian 12-14 year olds from 2006 to 2017: an examination of drug use profiles, emotional control problems, and family relationship characteristics

Polydrug use in Australian 12-14 year olds from 2006 to 2017: an examination of drug use... AUSTRALIAN JOURNAL OF PSYCHOLOGY 2023, VOL. 75, NO. 1, 2174705 https://doi.org/10.1080/00049530.2023.2174705 Polydrug use in Australian 12-14 year olds from 2006 to 2017: an examination of drug use profiles, emotional control problems, and family relationship characteristics a,b,c b b,d e,f a,b,c Adrian B. Kelly , Andrew Munnings , Xiang Zhao , Bosco Rowland , Kristin R. Laurens , a,g h i a,b j Marilyn Campbell , Joanne Williams , Jen A. Bailey , Callula Killingly , Julie Abimanyi-Ochom , k e Peter Kremer and John W. Toumbourou a b Centre for Inclusive Education, Queensland University of Technology, Brisbane, Australia; School of Psychology and Counselling, Queensland University of Technology, Brisbane, Australia; Centre for Child Health and Well-being (Childhood Adversity, Mental Health, and Resilience Theme), Queensland University of Technology, Brisbane, Australia; School of Law, Psychology and Social Work, Örebro University, Örebro, Sweden; School of Psychology and Centre for Social and Early Emotional Development, Deakin University, Melbourne, f g Australia; Faculty of Medicine and the Eastern Health Clinical School, Monash University Melbourne, Australia; School of Early Childhood and Inclusive Education, Queensland University of Technology, Brisbane, Australia; School of Health, Swinburne University of i j Technology, Melbourne, Australia; Social Developmental Research Group, University of Washington, Seattle, WA, USA; School of Health and Social Development, Deakin University, Melbourne, Australia; School of Exercise and Nutrition Sciences and Centre for Sport Research, Deakin University, Melbourne, Australia ABSTRACT ARTICLE HISTORY Received 7 June 2022 Objective: This study examined the nature and prevalence of polydrug use in 12–14 year old Accepted 25 January 2023 Australians. Method: Three Australian school surveys (2006, n=4091; 2009, n=5635; 2017, n=1539; age 12– KEYWORDS 14 years) spanning 11 years were used. Substances included alcohol, tobacco, cannabis, Adolescent; polydrug use; inhalant, and other illicit substances. Risk factors included depressed mood, low emotional family relationships; family control, poor family management and conflict, and academic performance. Latent class ana- conflict; academic lysis was used to discern classes. Regression analyses were used to test the association of risk performance; emotional control factors with classes. Results: Consistent across surveys, there was a class of adolescents who engaged in wide- ranging polydrug use, with prevalences ranging from 0.44% (2006) to 1.78% (2017). Emotional control problems, low academic performance, and poor family management were elevated in the polydrug class. Conclusion: A small proportion of 12–14-year-old adolescents engage in polydrug use. Interventions focusing on family risks and emotional control problems may be beneficial. KEY POINTS What is already known about this topic (1) In Australia, adolescents have generally reduced their use of alcohol and tobacco over recent decades. (2) Most research is based on patterns of use of single substances in mid-to-late adolescence, but we know that a significant proportion of older Australian adolescents engage in polydrug use. (3) Family relationship quality has been associated with drug use amongst older adolescents and young adults but may have an especially significant association with polydrug use amongst younger adolescents given key biopsychosocial transitions occurring around this age. What this research adds: (1) A small but meaningful proportion of Australian 12–14-year-olds engage in polydrug use. (2) The nature of polydrug use amongst young Australian adolescents has shifted since 2006, with profiles showing decreased tobacco use and continuing challenges in addressing alcohol, cannabis and inhalant use amongst young adolescents. This group also reported poor family management, poor emotional control, and academic failure. (3) The results highlight the importance of detection and targeted early intervention for a subgroup of young adolescents who may have developed risky drug use patterns across the transition to high school. CONTACT Adrian B. Kelly a.kelly@qut.edu.au Supplemental data for this article can be accessed at https://doi.org/10.1080/00049530.2023.2174705. © 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 2 A. B. KELLY ET AL. Introduction Amongst Australian adolescents 14–17 years of age, report polydrug use in the past month (Göbel et al., there have been significant downward trends over 2016). In summary, past month prevalence of polydrug recent decades in the use of alcohol and certain use in the past month for middle adolescents ranges other drugs. Since 1998, heavy episodic alcohol use from 4 to 9%. (5+ drinks/day) and weekly alcohol use have steadily Rates of past year/lifetime polydrug use show decreased and there has been a relatively steady large variations, primarily because of differences in increase in the prevalence of alcohol abstinence sampling procedures. Around 20.3% of adolescents (Kelly et al., 2016). There have been significant declines (12–17 years old) were classified as polydrug users in tobacco initiation among 14+ year olds (Australian (A. White et al., 2013). Specifically, 18.3% were lim- Institute of Health and Welfare, 2020). From 2001 to ited range polydrug users (alcohol, tobacco and 2016, there were decreases in the lifetime prevalence cannabis) and 2% were extended polydrug users and recent use (last 12 months) of cannabis amongst (these plus painkillers, ecstasy, amphetamine). 15–19 year olds (from approximately 34% in 2001 to Research from Michigan (US) indicates that 4.6% of 17% in 2016) (Australian Institute of Health and students had a high probability of polydrug use in Welfare, 2020). Amongst 14–19 year olds, there were the past 12 months (Cranford et al., 2013). In substantial drops in recent use of meth/amphetamine a young Brazilian adolescent sample (mean age from 2001 to 2007 (use for non-medical purposes), 12.6 years) recruited for a prevention program and a levelling out of recent use until 2013, and further based on a restricted range of drug types, 1.8% of decreases in recent use by 2016 (Australian Institute of adolescents were classed as polydrug users (Valente Health and Welfare, 2020). While downward trends in et al., 2017). In a Thai sample, 11% of 12- to 15- the recent use of specific substances by middle to later year-olds reported past year alcohol and tobacco aged adolescents are positive from a public health use and high-risk behaviours like fighting and carry- perspective, these trends provide little or no informa- ing a weapon and a further 0.6% reported high-risk tion on adolescents who are polydrug users (recent behaviours and illicit drug use in the past year users of 2+ drug types). (Assanangkornchai et al., 2018). In summary, the rates of adolescent past year/lifetime polydrug use Comparatively few studies are available in Australia vary from 1.8% to 24.6%. and internationally that quantify polydrug use amongst young adolescents (12–14 years of age) (see The overall aim of this study was to examine three Table S1 for details of available studies). Amongst older Australian population surveys to identify substance use subgroups (“classes”) amongst early adolescents, and to American adolescents, polydrug use is reported by examine variations across these subgroups on 1.7% to 2% of students (Banks et al., 2020). Connell et al. (2010) found a class of users that comprised 10% depressed mood, emotion regulation problems, family management practices and academic performance. of the sample and which reported recent alcohol, Establishing the prevalence and patterns of polydrug tobacco, cannabis, and lifetime use of other illicit drugs, inhalants, and nonmedical use of prescribed use amongst early adolescents is important because early onset of substance use is more strongly associated medicines (NUPM). In one of the few nationally repre- with adult drug and alcohol problems than late onset of sentative American studies of polydrug use amongst Grade 8 students, 9% reported a high probability of substance use (Kelly et al., 2019), drug use is associated with great neuropsychological harm for adolescents using any of 5 substances (e-cigarettes, tobacco, can- than for adults (Lubman et al., 2016; Silveri et al., 2016; nabis, NUPM and alcohol) (Miech et al., 2016). In Squeglia & Gray, 2016), and simple prevention messages Australia, polydrug use (primarily alcohol, tobacco, cannabis) is reported by 4.7% of 13.5 year olds in the (e.g., alcohol or tobacco focused universal prevention) may not be useful or relevant to adolescents using prior month (Chan et al., 2016) and these rates are multiple drugs (Kelly et al., 2019; Masterman & Kelly, higher (8.2%) in an older sample of 14.3 year olds (Chan et al., 2017). Amongst Swedish Grade 9 and 11 2003). An examination of the association of family rela- tionship problems (conflict, supervision/management students reporting any illicit drug use, 7% reported problems) may also inform the focus and timing of polydrug use (Miech et al., 2016). Across 25 European countries, 4.7% of adolescents (mean age 13.9 years) family oriented interventions. AUSTRALIAN JOURNAL OF PSYCHOLOGY 3 Materials and methods students (79.1% of students initially approached). Participants for this study were included if they Sample were between 12 and 14 years of age (n = 5,470). 2006 survey The mean age of this cohort was 12.5 years (SD = The 2006 study (the Healthy Neighbourhoods Study) 0.64) and 50.0% of this sample were male. involved 7,866 adolescents (52.6% female) from 231 Australian schools (from the States of Victoria, 2017 survey Queensland, and Western Australia). The community The 2017 study recruited participants from 56 second- sampling frame consisted of Statistical Local Areas ary schools across three Australian states (Victoria, (ABS, 2009) with greater than 17,000 inhabitants. Queensland and Western Australia) and utilised the These Statistical Local Areas were stratified into quar- 28 communities that initially participated in the 2006 tiles of socioeconomic disadvantage based on Socio- study. The community sampling frame was identical to Economic Indexes for Areas (SEIFA), which indexes the that in the earlier reported 2006 (Healthy average income and employment status for each resi- Neighbourhoods) Study. The final sample of 12–14- dential postcode in Australia (ABS, 2009). Eligible com- year-old students was 1,539 adolescents with a mean munities were randomly selected from SEIFA quartiles age of 13.0 years (SD = 0.84), and 46.3% of this sample to represent State distributions of advantage/disad- were male. vantage as well as urban and nonurban locations. All procedures followed were in accordance with A total of 164 primary and 82 secondary schools across the ethical standards of the responsible committee all communities were randomly selected. Of the on human experimentation (institutional and national) schools invited to participate, 83% (n = 443) and with the Helsinki Declaration of 1975, as revised in responded, and of these, 52% agreed to participate 2000. Active parental consent was obtained for adoles- (59% and 43% at Grade 6 and 8 levels, respectively). cent participants in the 2006 and 2017 Surveys and Students participated only if written parental consent these surveys were approved by the University of was obtained (67% response rate). The initial sample Melbourne Human Research Ethics Committee and consisted of 7,866 adolescents, with 4,091 adolescents the Deakin University Human Research Ethics aged between 12 and 14 years at the time of survey (M Committee respectively. Passive parental consent was = 12.24; SD = 0.44) included in the present study. 50.2% used for adolescent participants in the 2009 Survey of this sample were male. and this Survey was approved by the University of Melbourne Human Research Ethics Committee and 2009 survey the Victorian Department of Education. The 2009 sample was recruited as part of the HOWRU survey, which used a similar procedure to the 2006 Measures study but was conducted only in Victoria and used passive parental consent. Data collection involved The measures used in the three surveys were based on a two-stage sampling strategy. In the first stage, all the Communities That Care Youth Survey (CTC). This is schools in Victoria, Australia, were stratified into local an epidemiological assessment instrument developed government areas or school regions; schools were in the United States (Arthur et al., 2002) and adapted randomly selected from each strata based on for Australian youth (McMorris et al., 1998-2017), with a probability proportional to each community’s demonstrated reliability and validity (Kelly et al., 2011). grade-level size. Overall, 220 schools from The survey takes approximately 45 minutes for Government, Independent, and Catholic education a participant to complete. sectors took part. In the second stage, one class from each of Grades 7, 9, and 11 at each school was Substance use randomly selected. Passive informed consent from The following items assessed lifetime use of a range of parents was conducted for the majority of schools; substances, each according to a five-point response however, active parent consent was required by scale (1 “never”, 2 “1 or 2 times”, 3 “3 to 5 times”, 4 “6 some Catholic schools. Initially, 13501 students from to 9 times”, 5 “10 or more times”), where higher scores Grades 7, 9, and 11 were approached, among whom indicating increased frequency of use. Alcohol: “In your 37 students declined to participate (0.3%) and 739 lifetime have you had more than just a few sips of an parents declined to provide consent (5.5%). Of the alcoholic beverage (like beer, wine or spirits)?”; remaining participants, 2,047 were absent on the day Tobacco: “In your lifetime have you ever smoked cigar- of survey. The final sample consisted of 10,678 ettes?”; Cannabis: “In your lifetime have you ever used 4 A. B. KELLY ET AL. marijuana (pot, weed, grass)?”; Inhalants: “In your life- Family variables time have you ever sniffed glue, breathed the contents Poor family management was measured using 9 items of an aerosol spray can, or inhaled other gases or on a 4-point scale (1 “definitely yes” to 4 “definitely sprays, in order to get high?”; Other illicit substances: no”). Example items include “My parents ask if I’ve “In your lifetime have you ever used other illegal drugs done my homework”, “Would your parents know if (like cocaine, heroin, ecstasy, or amphetamines/ you did not come home on time”, “The rules in my speed)?”. Participants responses to the lifetime sub- family are clear” (Alphas ranged from .84 to .85 across stance use questions were dichotomised (0 “never surveys). Family conflict was measure using 3 items on used”, 1 “One or more times”). a 4-point scale (1 “definitely yes” to 4 “definitely no”): “We argue about the same things in my family over Depressed mood and over”, “People in my family have serious argu- For the 2006 and 2017 surveys, the Short Mood and ments”, and “People in my family often insult and yell Feelings Questionnaire (SMFQ; Angold et al., 1995) was at each other” (Alphas ranged from .78 to .83 across used. This is a 13-item measure of mood over the past surveys). Parents’ favourable attitude to substance use 2 weeks (alpha = 0.91). Example items include “I felt was measure using 4 items (1 “very wrong”, 2 “wrong”, miserable or unhappy” and “I didn’t enjoy anything at 3 “A little bit wrong”, 4 “Not wrong at all”): “How wrong all”, with responses ranging from 1 (“Not true”) to 3 do your parents feel it would be for you to . . . ” “smoke (“True”). Higher scores indicate a greater level of cigarettes”, “drink beer or wine regularly (at least once depressive symptoms (scores range from 0 to 26). For or twice a month)”, “drink spirits regularly? (at least the 2009 survey, the Kessler Psychological Distress once or twice a month)”, and “use marijuana (pot, Scale (K10) (Kessler et al., 2002) was used. Each of 10 weed, grass)” (Alphas ranged from 0.78 to 0.82 across items (e.g., “In the past 4 weeks, about how often did surveys). you feel hopeless/nervous/worthless?”) is rated using a 5-point Likert scale (1 “none”, 2 “a little”, 3 “some”, 4 Analyses “most”, 5 “all of the time”). K10 scores range from 10– 50 with higher indicating greater levels of psychologi- Statistical analyses were conducted using SPSS cal distress in the past 4 weeks. The internal reliability (Version 25) for the logistic regression and Mplus ver- of the K10 for the present data set was excellent sion 8.1 (Muthen & Muthen, 1998 − 2017) for the latent (Cronbach’s = 0.90). class analysis. To estimate school level effects in sub- stance use, intra-class correlations were calculated for Emotional control the three surveys and for the five substance use cate- Emotional control was measured using four items on gories. Thirteen of fifteen ICCs were below the value a four-point scale (1 “definitely yes”, 2 “yes”, 3 “no”, 4 where multilevel or mixed models are recommended “definitely No”). The four items were “I know how to (.05; Hedges & Hedberg, 2007) and the ICCs for alcohol relax when I feel tense”, “I am always able to keep my use and tobacco use were marginally over 0.05 for the feelings under control”, “I know how to calm down if 2006 survey only (.07 and .06 respectively), so school I am feeling nervous”, and “I control my temper when level effects were not modelled in subsequent people are angry with me” (Cronbach’s Alphas range analyses. 0.74–0.79). Latent class analyses were performed, respectively, on all three surveys using the lifetime substance use Academic problems variables: Tobacco, alcohol, marijuana, inhalant, and Academic failure was measured using 2 items: “Putting other illicit substance use. Fit indices included the them all together, what were your marks like last year?” Akaike Information criteria (AIC; Akaike, 1987), the (5-point scale from 1 “very good” to 5 “very poor”), and Bayesian Information Criteria (BIC; Schwarz, 1978) and “Are your school marks better than the marks of most the Sample Size-adjusted Bayesian Information Criteria students in your class?” (4-point scale from 1 “definitely (Sclove, 1987). Entropy values estimated on the aver- yes” to 4 “definitely no”). Opportunities for prosocial age posterior probabilities were used to evaluate class involvement in school was measured using five items quality, with higher values signalling clear class separa- (e.g., “In my school, students have lots of chances to tion (Muthén, 2004; Nagin, 2005). Model fit statistics help decide things like class activities and rules”, were conducted beginning with a 2-class solution and “There are lots of chances for students in my school sequentially tested up to 4 classes. To evaluate predic- to talk with a teacher one-on-one”) (Alphas 0.57 to 0.65 tors of cluster membership, stepwise logistic regres- across surveys). sion (SPSS Version 25) was used. Independent AUSTRALIAN JOURNAL OF PSYCHOLOGY 5 predictors were grouped by individual predictors a single member. A 2-class model was chosen as (Block 1: depressed mood, emotional control), family the optimal solution as it yielded classification that predictors (Block 2: poor family management, parent’s was clearly distinct and interpretable, had high attitude to substance use), and school predictors average posterior probabilities signalling clear class (Block 3: academic failure, opportunities for prosocial separation, and subgroup sizes which presented as involvement at school). School level clustering of data more interpretable for making meaningful compar- was not modelled because of the extreme zero- isons between the three years (see Table 1). Class 1 inflatedness of polydrug class counts. participants (the ‘at-risk class) were the minority (<1%, n = 18) and were predominantly high fre- quency tobacco and alcohol users and low fre- Results quency cannabis and other illicit drug users (see Figure 1). Class 2 participants (99.6%; n = 4061) Latent class analyses were predominantly low frequency alcohol users Figure 1 depicts the estimated means for lifetime sub- and reported almost nil use of other substances. stance use associated with each class across the three This class was labelled the low-risk group and desig- surveys. nated the reference group in regression analyses. 2006 survey 2009 survey The 2-class and 3-class solutions had the greatest The optimal solution was a two-class solution as it entropy values (both values of >.99), but the 3-class yielded classification that was clearly distinct and had solution was not viable as it included a class with 2017: At-risk group n = 26 2006: At-risk group n = 18 2009: At-risk group n = (1.8%), low-risk n = 1433 101(1.8%), low-risk n = 5369 (0.4%), low-risk n = 4061 (99.6 (98.2%) %) (98.2%) 10 or more times 5 5 6 to 9 times 4 4 3 to 5 times 3 3 2 2 1 or 2 times Never 1 1 1 0 0 0 Substance Type Figure 1. Estimated means for lifetime substance use across the three surveys. Notes. The same five response categories were used for each lifetime substance use question across the three datasets: 1 = “Never”, 2 = “1 or 2 times”, 3 = “3 to 5 times”, 4 = “6 to 9 times”, 5 = “10 or more times”. -♦- = low-risk class; -●- = at-risk class. Other Sub = Other substance use (cocaine, heroin, ecstasy, or amphetamines/speed). Table 1. Fit indices from latent class analyses for the three surveys. 2006 2009 2017 2-class 3-class 4-class 2-class 3-class 4-class 2-class 3-class 4-class SSBIC 13566.88 8410.58 3815.22 35687.46 28820.20 22086.81 6376.148 5001.58 3972.54 AIC 13516.70 8341.58 3727.41 35632.59 28744.76 21990.79 6342.41 4955.19 3913.49 BIC 13617.72 8480.48 3904.19 35738.31 28890.11 22175.80 6426.98 5071.47 4061.48 Entropy 1.00 1.00 0.99 0.99 0.99 0.99 0.99 0.99 0.98 Subgroup size C1 18 17 17 101 279 281 26 34 16 C2 4061 4061 3878 5369 23 22 1433 22 81 C3 1 1 5168 77 1403 144 C4 183 5090 1218 SSBIC = Sample size adjusted Bayesian information criteria. AIC = Akaike information criteria. BIC = Bayesian information criteria. C1 = Class 1; C2 = Class 2; C3 = Class 3; C4 = Class 4. Categorical Response Options 6 A. B. KELLY ET AL. strong interpretability and high average posterior family predictors were entered (Block 2), family conflict probabilities (see Table 1). Class 1 participants (“at- (p = .063), poor family management, and parents’ atti- risk” class) were the minority (1.8%, n = 101) and were tude to substance use predicted the “at-risk” class predominantly high frequency users of tobacco, alco- membership and emotional control was retained as hol, and marijuana users, and lower frequency users of a significant predictor. When school predictors were inhalants and other substance users (see Figure 1). entered (Block 3), the following predictors were signif- Class 2 participants were predominantly low frequency icant: academic failure, poor family management, and alcohol users and the use of other kinds of substances emotional control, and there was a near-significant was rare. This class was labelled the low-risk class and trend for family conflict (p = .082). Psychological dis- was used as the reference group (98.2%; n = 5369). tress, parental attitudes to substance use, and oppor- tunities for prosocial involvement (p = .381) were not significant predictors of class membership. 2017 survey As for the prior surveys, a 2-class model had the high- est entropy, interpretability, and average posterior Discussion probabilities (see Table 1). Class 1 participants (“at- This study investigated profiles of lifetime substance risk” class) were the minority (1.8%, n = 26) and were use across three population surveys, and the associa- predominantly high frequency alcohol, cannabis, inha- tion of individual, familial and school factors with these lant and other substances users (see Figure 1). They latent classes. The great majority of adolescents aged also reported elevated tobacco use. Class 2 (“low risk”) 12–14 years reported almost nil substance use. Each participants were predominantly low frequency alco- survey had a small but notable minority (between .44% hol users and used almost no other kinds of substances and 1.85%) of adolescents reporting the use of multi- (98.2%; n = 1433). ple substances (labelled polydrug users). The topogra- phy of substance use showed meaningful variations Regression analysis across the three surveys. In 2006, polydrug users reported elevated frequencies of tobacco, alcohol, We limited our regression analyses (dependent vari- and cannabis use, and one or two instances of inhalant able class membership) to the 2009 survey, where the and other (illicit) substance use. In 2009, polydrug absolute number of polydrug users was highest . The users reported elevated frequencies of tobacco and results of the logistic regression are presented in alcohol (similar rates to 2006) and more frequent can- Table 2 and summarised here. In Block 1, emotional nabis use than 2006. Inhalant use and other substance control but not psychological distress was significantly use appeared similar to 2006 rates. In 2017, polydrug associated with membership of the at-risk class. When users reported lower frequencies of tobacco use com- pared to 2006, similar levels of alcohol and cannabis Table 2. Logistic regression results showing the prediction of use, and higher frequencies of inhalant and other (illi- at-risk class membership relative to low-risk class membership. cit) substance use. Compared to the low-risk group, Variables B 95% CI R polydrug users reported poorer family management Individual predictors (Block 1) .066 practices and greater problems with emotional con- Emotional Control .29*** [1.19,1.49] Psychological Distress .01 [.97,1.04] trol. The results are consistent with the possibility that Family predictors (Block 2) .184 adolescents are at elevated risk of polydrug use when Emotional Control .15* [1.03,1.30] Psychological Distress −.02 [.95,1.02] parental awareness and supervision of adolescent Poor Family Management .15*** [1.10,1.22] activities are poor and adolescents have difficulties Family Conflict −.11‡ [.80,1.01] managing associated negative emotions. Polydrug Parents’ Attitude to Substance Use .10* [1.01,1.20] School predictors (Block 3) .201 use may be used instrumentally to control negative Emotional Control .13* [1.01,1.28] emotions and engaging with peer networks that use Psychological Distress −.02 [.95,1.02] Poor Family Management .14*** [1.09,1.22] multiple drugs may be a way of escaping distressed Family Conflict −.10† [.80,1.01] and disengaged families. Parents’ Attitude to Substance Use .09* [1.00,1.19] Academic Failure .29** [1.10,1.62] The findings on lifetime prevalence of polydrug use Opportunities for Prosocial Involvement at .06 [.92,1.20] in these Australian 12–14 year old surveys were com- School † ‡ parable to those found in Brazil (1.8%; Valente et al., p = .082, p = .063, *p < .050, **p < .010, ***p < .001. Higher scores on emotional control reflect poor abilities to manage one’s emotions. 2017) and Australian nationally representative studies Higher scores on parents’ attitude to substance use indicates more (2%; White et al., 2013), but below the prevalence rates favourable attitudes related to the offspring’s substance use. Higher scores on the family conflict scale represent lower family conflict. of polydrug use reported in other countries, including AUSTRALIAN JOURNAL OF PSYCHOLOGY 7 the United States (Cranford et al., 2013). The present cannot be established. The cell sizes for polydrug users prevalence rate was higher than prior research in were generally small and necessitated a focus on key Thailand (0.6%; Assanangkornchai et al., 2018) but variables at the exclusion of other potentially important this was unsurprising because their classes included variables. Substance use and problems amongst other a range of high-risk behaviours additional to substance family members were not assessed as part of this study, use (e.g., carrying a weapon), which is rarely reported and it is possible that these factors may be significant in Australian early adolescent populations. The similar- drivers of adolescent polydrug use. We did not include ity of findings from these surveys and prior National sex and age in our analyses because cell sizes for poly- Drug Strategy Household Survey data points to the drug users were small so it was necessary to minimise convergent validity of the present study. It is also the number of independent variables entered into each notable that the findings were relatively consistent model. Also, there were no meaningful differences in sex across informed consent mechanisms, suggesting and age across the three surveys, there was an almost that downward biases associated with active parental even split of genders across the surveys, and the age consent mechanism were minimal: The 2009 survey range for inclusion in the study (12–14 years of age) was (which used passive parental consent) had slightly small. higher polydrug use prevalence rates than both the 2006 (by 1.41%) and the 2017 survey (by 0.07%) (both Conclusion of which used active parental consent). However, the magnitude of these variations was small given that This study showed that a small proportion of 12–14- active parental consent mechanisms were associated year-old adolescents use multiple drugs and that this with the exclusion of 27.5% of the field of potential prevalence has changed very little over the last 14 participants. years, despite great population-level success in the There are significant challenges with detection and reduction of specific drug use, including tobacco and intervention for polydrug users, given the low preva- alcohol. Depressed mood and emotional control pro- lence rate for this class. Interventions for polydrug use blems characterise polydrug users. Tobacco use was might follow a stepped care approach (Toumbourou less prominent in the drug use profiles of the most et al., 2021), where students experiencing problems recent survey. The findings point to the importance of are screened for drug use and targeted individual improving early detection of at-risk adolescents and interventions are delivered by school-based health providing interventions that are tailored to meet com- professionals. Evidence-based interventions for poly- plex presentations. drug using early adolescents might require a comprehensive focus on emotional control pro- blems, strengthening school achievement, and screen- Note ing/intervention for a variety of substances. Intensive 1. We elected not to include the findings of the logis- and sustained multimodal programs addressing aca- tic regression for the 2006 survey and the 2017 demic performance through tutoring, drug use pre- survey because of the very small cell sizes of poly- vention, and family risk factors are effective (Conduct drug users (n = 18 and 26 for the 2006 and 2017 Problems Prevention Research Group, 2004). survey respectively) and the relatively large number Longitudinal research findings and controlled trials of variables in the full model (7 independent vari- point to the value of family-oriented interventions for ables). For completeness we note that for the 2006 dataset the results for the same model were largely addressing adolescent depression and drug problems consistent with the 2009 data set, with significance and this is consistent with the family risks established for emotional control, poor family management and in the present study (Chan et al., 2013; Mason et al., parental attitudes towards substance use, and non- 2003, 2012). Longitudinal research would help to significance for depressed mood. For the 2017 sur- establish dominant paths of influence, which in turn vey, none of the variables were significant. The most likely account of the pattern of nonsignifi - would guide the relative weighting program compo- cance for the 2017 survey compared to the 2009 nents and delivery modes. survey is lack of statistical power. A strength of this study is that it utilised three large- scale Australian samples, it included a mix of pathways to participation permitting redress of potential sample Disclosure statement bias (Kelly & Halford, 2007), and it focused on young adolescents (12–14 year olds) where data is scarce. The No potential conflict of interest was reported by the author(s). study is cross-sectional and so causal relationships 8 A. B. KELLY ET AL. Data availability statement parenting and school factors moderate this association? Addictive Behaviors, 64, 78–81. https://doi.org/10.1016/j. The data that support the findings of this study are available addbeh.2016.08.004 from the last author, Prof. John Toumbourou, Deakin Chan, G. C. K., Kelly, A. B., Hides, L., Quinn, C., & Williams, J. W. University, john.toumbourou@deakin.edu.au, upon reason- (2016). Does gender moderate the relationship between able request. polydrug use and sexual risk-taking among Australian secondary school students under 16 years of age? Drug and Alcohol Review, 35(6), 750–754. https://doi.org/10. Funding 1111/dar.12394 Chan, G. C. K., Kelly, A. B., & Toumbourou, J. W. (2013). Kristin R. Laurens was supported by the Australian Research Accounting for the association of family conflict and very Council [FT170100294]; National Health and Medical young adolescent female alcohol use: The role of Research Council [APP1087781]; Centre for Inclusive depressed mood. Journal of Studies on Alcohol and Drugs, Education, Queensland University of Technology, Brisbane, 74(3), 396–505. https://doi.org/10.15288/jsad.2013.74.396 Australia. Conduct Problems Prevention Research Group. (2004). The effects of the fast track program on serious problem out- comes at the end of elementary school. Journal of Clinical ORCID Child & Adolescent Psychology, 33(4), 650–661. https://doi. org/10.1207/s15374424jccp3304_1 Adrian B. Kelly http://orcid.org/0000-0001-5546-4994 Connell, C., Gilreath, T., Aklin, W., & Brex, R. (2010). Social- Xiang Zhao http://orcid.org/0000-0003-1054-9462 ecological influences on patterns of substance use among Bosco Rowland http://orcid.org/0000-0003-0192-809X non-metropolitan high school students. American Journal Kristin R. Laurens http://orcid.org/0000-0002-3987-6486 of Community Psychology, 45(1), 36–48. https://doi.org/10. Marilyn Campbell http://orcid.org/0000-0002-4477-2366 1007/s10464-009-9289-x Julie Abimanyi-Ochom http://orcid.org/0000-0002-4760- Cranford, J. A., McCabe, S. E., & Boyd, C. J. (2013). Adolescents’ nonmedical use and excessive medical use of prescription John W. Toumbourou http://orcid.org/0000-0002-8431- medications and the identification of substance use sub- groups. Addictive Behaviors, 38(11), 2768–2771. https://doi. org/10.1016/j.addbeh.2013.06.015 Göbel, K., Scheithauer, H., Bräker, A. -B., Jonkman, H., & References Soellner, R. (2016). Substance use patterns among adoles- cents in Europe: A latent class analysis. Substance Use & ABS. (2009). Socio-economic indices for areas. Misuse, 51(9), 1130–1138. https://doi.org/10.3109/ Akaike, H. (1987). Factor analysis and AIC. Psychometrika, 52, 10826084.2016.1160120 317–332. https://doi.org/10.1007/BF02294359 Hedges, L. V., & Hedberg, E. C. (2007). Intraclass correlation Angold, A., Costello, E. J., Messer, S. C., & Pickles, A. (1995). values for planning group-randomized trials in education. Development of a short questionnaire for use in epide- Educational Evaluation and Policy Analysis, 29(1), 60–87. miological studies of depression in children and https://doi.org/10.3102/0162373707299706 adolescents. International Journal of Methods in Kelly, A. B., Chan, G. C. K., Weier, M., Quinn, C., Gullo, M., Psychiatric Research, 5, 237–249. http://devepi.duhs.duke. Connor, J. P., & Hall, W. D. (2016). Parental supply of alcohol edu/AngoldMFQarticle.pdf to Australian minors: An analysis of six nationally represen- Arthur, M. W., Hawkins, J. D., Pollard, J. A., Catalano, R. F., & tative surveys spanning 15 years. BMC Public Health, 15(1), Baglioni, A. J. (2002). Measuring risk and protective factors 325. https://doi.org/10.1186/s12889-016-3004-2 for substance use, delinquency, and other adolescent pro- Kelly, A. B., & Halford, W. K. (2007). Responses to ethical blem behaviors: The Communities That Care Youth challenges in conducting research with Australian Survey. Evaluation Review, 26, 575–601. https://doi.org/ adolescents. Australian Journal of Psychology, 59(1), 10.1177/0193841X0202600601 24–33. https://doi.org/10.1080/00049530600944358 Assanangkornchai, S., Li, J., McNeil, E., & Saingam, D. (2018). Kelly, A. B., O’flaherty, M., Connor, J. P., Homel, R., Clusters of alcohol and drug use and other health-risk Toumbourou, J. W., Patton, G. C., & Williams, J. (2011). behaviors among Thai secondary school students: The influence of parents, siblings and peers on pre- and A latent class analysis. BMC Public Health, 18(1), 1272. early-teen smoking: A multilevel model. Drug and Alcohol https://doi.org/10.1186/s12889-018-6205-z Review, 30(4), 381–387. https://doi.org/10.1111/j.1465- Australian Institute of Health and Welfare. (2020). Alcohol, 3362.2010.00231.x tobacco & other drugs in Australia. https://www.aihw.gov. Kelly, A. B., Weier, M., & Hall, W. D. (2019). First use of illicit au/reports/alcohol/alcohol-tobacco-other-drugs-australia Banks, D. E., Bello, M. S., Crichlow, Q., Leventhal, A. M., Barnes- drugs: The state of current knowledge. In G. Girolamo, Najor, J. V., & Zapolski, T. C. B. (2020). Differential typolo- P. McGorry, & N. Sartorius (Eds.), Age of onset of mental gies of current substance use among black and white disorders: Etiopatogenetic and treatment implications high-school adolescents: A latent class analysis. Addictive (pp.149–168). Springer International. Behaviors, 106. https://doi.org/10.1016/j.addbeh.2020. Kessler, R. C., Andrews, G., Colpe, L. J., Hiripi, E., Mroczek, D., 106356 Normand, S. L., Walters, E., & Zaslavsky, A. (2002). Short Chan, G. C. K., Kelly, A. B., Carroll, A., & Williams, J. W. (2017). screening scales to monitor population prevalences and Peer drug use and adolescent polysubstance use: Do trends in nonspecific psychological distress. Psychological AUSTRALIAN JOURNAL OF PSYCHOLOGY 9 Medicine, 32(6), 959–976. https://doi.org/10.1017/ Sclove, S. L. (1987). Application of model-selection criteria to S0033291702006074 some problems in multivariate analysis. Psychometrika, 52 Lubman, D. I., Yücel, M., & Hall, W. D. (2016). Substance use (3), 333–343. https://doi.org/10.1007/BF02294360 and the adolescent brain: A toxic combination? Journal of Silveri, M. M., Dager, A. D., Cohen-Gilbert, J. E., & Sneider, J. T. Psychopharmacology (Oxford), 21(8), 792–794. https://doi. (2016). Neurobiological signatures associated with alcohol org/10.1177/0269881107078309 and drug use in the human adolescent brain. Neuroscience Mason, W. A., Haggerty, K. P., Fleming, A. P., & Casey- and Biobehavioral Reviews, 70, 244–259. https://doi.org/10. Goldstein, M. (2012). Family intervention to prevent 1016/j.neubiorev.2016.06.042 depression and substance use among adolescents of Squeglia, L. M., & Gray, K. M. (2016). Alcohol and drug use and depressed parents. Journal of Child and Family Studies, 21 the developing brain. Current Psychiatry Reports, 18(5), (6), 891–905. https://doi.org/10.1007/s10826-011-9549-x 1–10. https://doi.org/10.1007/s11920-016-0689-y Mason, W. A., Kosterman, R., Hawkins, J. D., Haggerty, K. P., & Toumbourou, J. W., Heerde, J., Kelly, A. B., & Bailey, J. (2021). Spoth, R. (2003). Reducing adolescents’ growth in substance Marijuana use. In W. O’Donohue & Z. Zimmerman (Eds.), use and deliquency: Randomized trial effects of a parent- Handbook of evidence-based prevention of behavioral dis- training prevention intervention. Prevention Science, 4(3), orders in integrated care: A stepped care approach (pp 251– 203–212. https://doi.org/10.1023/A:1024653923780 265, Chapter 11). Springer Nature http://doi.org/10.1007/ Masterman, P. W., & Kelly, A. B. (2003). Reaching adolescents 978-3-030-83469-2_11 . who drink harmfully: Fitting intervention to developmen- Valente, J. Y., Cogo-Moreira, H., & Sanchez, Z. M. (2017). tal reality. Journal of Substance Abuse Treatment, 24(4), Gradient of association between parenting styles and pat- 347–355. https://doi.org/10.1016/S0740-5472(03)00047-3 terns of drug use in adolescence: A latent class analysis. Miech, R. A., O’malley, P. M., Johnston, L. D., & Patrick, M. E. Drug and alcohol dependence, 180, 272–278. https://doi. (2016). E-cigarettes and the drug use patterns of org/10.1016/j.drugalcdep.2017.08.015 adolescents. Nicotine & Tobacco Research, 18(5), 654–659. White, A., Chan, G. C. K., Quek, L. H., Connor, J. P., Baker, P., https://doi.org/10.1093/ntr/ntv217 Saunders, J. B., & Kelly, A. B. (2013). The topography of Muthén, B. (2004). Latent Variable Analysis: Growth Mixture multiple drug use among adolescent Australians: Findings Modeling and Related Techniques for Longitudinal Data. from the national drug strategy household survey. In D. Kaplan (Ed.), Handbook of Quantitative Methodology Addictive Behaviors, 38(4), 2038–2073. http://doi.org/10. for the Social Sciences (pp. 345–368). Sage. https://doi.org/ 1016/j.addbeh.2013.01.001 10.4135/9781412986311.n19 White, A., Chan, G. C., Quek, L. -H., Connor, J. P., Muthén, L. K., & Muthén, B. O. (1998-2017). Mplus User’s Guide (8th ed.). Muthén & Muthén. Saunders, J. B., Baker, P., Brackenridge, C., Kelly, A. B. (2013). The topography of multiple drug use among Nagin, D. S. (2005). Group-based modeling of development. adolescent Australians: Findings from the national Harvard University Press. drug strategy household survey. Addictive Behaviors, Schwarz, G. (1978). Estimating the Dimension of a Model. Annals of Statistics, 6, 461–464. https://doi.org/10.1214/ 38(4), 2068–2073.https://doi.org/10.1016/j.addbeh.2013. aos/1176344136 01.001 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Australian Journal of Psychology Taylor & Francis

Polydrug use in Australian 12-14 year olds from 2006 to 2017: an examination of drug use profiles, emotional control problems, and family relationship characteristics

Download PDF
9 pages

Loading next page...
 
/lp/taylor-francis/polydrug-use-in-australian-12-14-year-olds-from-2006-to-2017-an-MTjAQFXjX0

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Taylor & Francis
Copyright
© 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
ISSN
1742-9536
eISSN
0004-9530
DOI
10.1080/00049530.2023.2174705
Publisher site
See Article on Publisher Site

Abstract

AUSTRALIAN JOURNAL OF PSYCHOLOGY 2023, VOL. 75, NO. 1, 2174705 https://doi.org/10.1080/00049530.2023.2174705 Polydrug use in Australian 12-14 year olds from 2006 to 2017: an examination of drug use profiles, emotional control problems, and family relationship characteristics a,b,c b b,d e,f a,b,c Adrian B. Kelly , Andrew Munnings , Xiang Zhao , Bosco Rowland , Kristin R. Laurens , a,g h i a,b j Marilyn Campbell , Joanne Williams , Jen A. Bailey , Callula Killingly , Julie Abimanyi-Ochom , k e Peter Kremer and John W. Toumbourou a b Centre for Inclusive Education, Queensland University of Technology, Brisbane, Australia; School of Psychology and Counselling, Queensland University of Technology, Brisbane, Australia; Centre for Child Health and Well-being (Childhood Adversity, Mental Health, and Resilience Theme), Queensland University of Technology, Brisbane, Australia; School of Law, Psychology and Social Work, Örebro University, Örebro, Sweden; School of Psychology and Centre for Social and Early Emotional Development, Deakin University, Melbourne, f g Australia; Faculty of Medicine and the Eastern Health Clinical School, Monash University Melbourne, Australia; School of Early Childhood and Inclusive Education, Queensland University of Technology, Brisbane, Australia; School of Health, Swinburne University of i j Technology, Melbourne, Australia; Social Developmental Research Group, University of Washington, Seattle, WA, USA; School of Health and Social Development, Deakin University, Melbourne, Australia; School of Exercise and Nutrition Sciences and Centre for Sport Research, Deakin University, Melbourne, Australia ABSTRACT ARTICLE HISTORY Received 7 June 2022 Objective: This study examined the nature and prevalence of polydrug use in 12–14 year old Accepted 25 January 2023 Australians. Method: Three Australian school surveys (2006, n=4091; 2009, n=5635; 2017, n=1539; age 12– KEYWORDS 14 years) spanning 11 years were used. Substances included alcohol, tobacco, cannabis, Adolescent; polydrug use; inhalant, and other illicit substances. Risk factors included depressed mood, low emotional family relationships; family control, poor family management and conflict, and academic performance. Latent class ana- conflict; academic lysis was used to discern classes. Regression analyses were used to test the association of risk performance; emotional control factors with classes. Results: Consistent across surveys, there was a class of adolescents who engaged in wide- ranging polydrug use, with prevalences ranging from 0.44% (2006) to 1.78% (2017). Emotional control problems, low academic performance, and poor family management were elevated in the polydrug class. Conclusion: A small proportion of 12–14-year-old adolescents engage in polydrug use. Interventions focusing on family risks and emotional control problems may be beneficial. KEY POINTS What is already known about this topic (1) In Australia, adolescents have generally reduced their use of alcohol and tobacco over recent decades. (2) Most research is based on patterns of use of single substances in mid-to-late adolescence, but we know that a significant proportion of older Australian adolescents engage in polydrug use. (3) Family relationship quality has been associated with drug use amongst older adolescents and young adults but may have an especially significant association with polydrug use amongst younger adolescents given key biopsychosocial transitions occurring around this age. What this research adds: (1) A small but meaningful proportion of Australian 12–14-year-olds engage in polydrug use. (2) The nature of polydrug use amongst young Australian adolescents has shifted since 2006, with profiles showing decreased tobacco use and continuing challenges in addressing alcohol, cannabis and inhalant use amongst young adolescents. This group also reported poor family management, poor emotional control, and academic failure. (3) The results highlight the importance of detection and targeted early intervention for a subgroup of young adolescents who may have developed risky drug use patterns across the transition to high school. CONTACT Adrian B. Kelly a.kelly@qut.edu.au Supplemental data for this article can be accessed at https://doi.org/10.1080/00049530.2023.2174705. © 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 2 A. B. KELLY ET AL. Introduction Amongst Australian adolescents 14–17 years of age, report polydrug use in the past month (Göbel et al., there have been significant downward trends over 2016). In summary, past month prevalence of polydrug recent decades in the use of alcohol and certain use in the past month for middle adolescents ranges other drugs. Since 1998, heavy episodic alcohol use from 4 to 9%. (5+ drinks/day) and weekly alcohol use have steadily Rates of past year/lifetime polydrug use show decreased and there has been a relatively steady large variations, primarily because of differences in increase in the prevalence of alcohol abstinence sampling procedures. Around 20.3% of adolescents (Kelly et al., 2016). There have been significant declines (12–17 years old) were classified as polydrug users in tobacco initiation among 14+ year olds (Australian (A. White et al., 2013). Specifically, 18.3% were lim- Institute of Health and Welfare, 2020). From 2001 to ited range polydrug users (alcohol, tobacco and 2016, there were decreases in the lifetime prevalence cannabis) and 2% were extended polydrug users and recent use (last 12 months) of cannabis amongst (these plus painkillers, ecstasy, amphetamine). 15–19 year olds (from approximately 34% in 2001 to Research from Michigan (US) indicates that 4.6% of 17% in 2016) (Australian Institute of Health and students had a high probability of polydrug use in Welfare, 2020). Amongst 14–19 year olds, there were the past 12 months (Cranford et al., 2013). In substantial drops in recent use of meth/amphetamine a young Brazilian adolescent sample (mean age from 2001 to 2007 (use for non-medical purposes), 12.6 years) recruited for a prevention program and a levelling out of recent use until 2013, and further based on a restricted range of drug types, 1.8% of decreases in recent use by 2016 (Australian Institute of adolescents were classed as polydrug users (Valente Health and Welfare, 2020). While downward trends in et al., 2017). In a Thai sample, 11% of 12- to 15- the recent use of specific substances by middle to later year-olds reported past year alcohol and tobacco aged adolescents are positive from a public health use and high-risk behaviours like fighting and carry- perspective, these trends provide little or no informa- ing a weapon and a further 0.6% reported high-risk tion on adolescents who are polydrug users (recent behaviours and illicit drug use in the past year users of 2+ drug types). (Assanangkornchai et al., 2018). In summary, the rates of adolescent past year/lifetime polydrug use Comparatively few studies are available in Australia vary from 1.8% to 24.6%. and internationally that quantify polydrug use amongst young adolescents (12–14 years of age) (see The overall aim of this study was to examine three Table S1 for details of available studies). Amongst older Australian population surveys to identify substance use subgroups (“classes”) amongst early adolescents, and to American adolescents, polydrug use is reported by examine variations across these subgroups on 1.7% to 2% of students (Banks et al., 2020). Connell et al. (2010) found a class of users that comprised 10% depressed mood, emotion regulation problems, family management practices and academic performance. of the sample and which reported recent alcohol, Establishing the prevalence and patterns of polydrug tobacco, cannabis, and lifetime use of other illicit drugs, inhalants, and nonmedical use of prescribed use amongst early adolescents is important because early onset of substance use is more strongly associated medicines (NUPM). In one of the few nationally repre- with adult drug and alcohol problems than late onset of sentative American studies of polydrug use amongst Grade 8 students, 9% reported a high probability of substance use (Kelly et al., 2019), drug use is associated with great neuropsychological harm for adolescents using any of 5 substances (e-cigarettes, tobacco, can- than for adults (Lubman et al., 2016; Silveri et al., 2016; nabis, NUPM and alcohol) (Miech et al., 2016). In Squeglia & Gray, 2016), and simple prevention messages Australia, polydrug use (primarily alcohol, tobacco, cannabis) is reported by 4.7% of 13.5 year olds in the (e.g., alcohol or tobacco focused universal prevention) may not be useful or relevant to adolescents using prior month (Chan et al., 2016) and these rates are multiple drugs (Kelly et al., 2019; Masterman & Kelly, higher (8.2%) in an older sample of 14.3 year olds (Chan et al., 2017). Amongst Swedish Grade 9 and 11 2003). An examination of the association of family rela- tionship problems (conflict, supervision/management students reporting any illicit drug use, 7% reported problems) may also inform the focus and timing of polydrug use (Miech et al., 2016). Across 25 European countries, 4.7% of adolescents (mean age 13.9 years) family oriented interventions. AUSTRALIAN JOURNAL OF PSYCHOLOGY 3 Materials and methods students (79.1% of students initially approached). Participants for this study were included if they Sample were between 12 and 14 years of age (n = 5,470). 2006 survey The mean age of this cohort was 12.5 years (SD = The 2006 study (the Healthy Neighbourhoods Study) 0.64) and 50.0% of this sample were male. involved 7,866 adolescents (52.6% female) from 231 Australian schools (from the States of Victoria, 2017 survey Queensland, and Western Australia). The community The 2017 study recruited participants from 56 second- sampling frame consisted of Statistical Local Areas ary schools across three Australian states (Victoria, (ABS, 2009) with greater than 17,000 inhabitants. Queensland and Western Australia) and utilised the These Statistical Local Areas were stratified into quar- 28 communities that initially participated in the 2006 tiles of socioeconomic disadvantage based on Socio- study. The community sampling frame was identical to Economic Indexes for Areas (SEIFA), which indexes the that in the earlier reported 2006 (Healthy average income and employment status for each resi- Neighbourhoods) Study. The final sample of 12–14- dential postcode in Australia (ABS, 2009). Eligible com- year-old students was 1,539 adolescents with a mean munities were randomly selected from SEIFA quartiles age of 13.0 years (SD = 0.84), and 46.3% of this sample to represent State distributions of advantage/disad- were male. vantage as well as urban and nonurban locations. All procedures followed were in accordance with A total of 164 primary and 82 secondary schools across the ethical standards of the responsible committee all communities were randomly selected. Of the on human experimentation (institutional and national) schools invited to participate, 83% (n = 443) and with the Helsinki Declaration of 1975, as revised in responded, and of these, 52% agreed to participate 2000. Active parental consent was obtained for adoles- (59% and 43% at Grade 6 and 8 levels, respectively). cent participants in the 2006 and 2017 Surveys and Students participated only if written parental consent these surveys were approved by the University of was obtained (67% response rate). The initial sample Melbourne Human Research Ethics Committee and consisted of 7,866 adolescents, with 4,091 adolescents the Deakin University Human Research Ethics aged between 12 and 14 years at the time of survey (M Committee respectively. Passive parental consent was = 12.24; SD = 0.44) included in the present study. 50.2% used for adolescent participants in the 2009 Survey of this sample were male. and this Survey was approved by the University of Melbourne Human Research Ethics Committee and 2009 survey the Victorian Department of Education. The 2009 sample was recruited as part of the HOWRU survey, which used a similar procedure to the 2006 Measures study but was conducted only in Victoria and used passive parental consent. Data collection involved The measures used in the three surveys were based on a two-stage sampling strategy. In the first stage, all the Communities That Care Youth Survey (CTC). This is schools in Victoria, Australia, were stratified into local an epidemiological assessment instrument developed government areas or school regions; schools were in the United States (Arthur et al., 2002) and adapted randomly selected from each strata based on for Australian youth (McMorris et al., 1998-2017), with a probability proportional to each community’s demonstrated reliability and validity (Kelly et al., 2011). grade-level size. Overall, 220 schools from The survey takes approximately 45 minutes for Government, Independent, and Catholic education a participant to complete. sectors took part. In the second stage, one class from each of Grades 7, 9, and 11 at each school was Substance use randomly selected. Passive informed consent from The following items assessed lifetime use of a range of parents was conducted for the majority of schools; substances, each according to a five-point response however, active parent consent was required by scale (1 “never”, 2 “1 or 2 times”, 3 “3 to 5 times”, 4 “6 some Catholic schools. Initially, 13501 students from to 9 times”, 5 “10 or more times”), where higher scores Grades 7, 9, and 11 were approached, among whom indicating increased frequency of use. Alcohol: “In your 37 students declined to participate (0.3%) and 739 lifetime have you had more than just a few sips of an parents declined to provide consent (5.5%). Of the alcoholic beverage (like beer, wine or spirits)?”; remaining participants, 2,047 were absent on the day Tobacco: “In your lifetime have you ever smoked cigar- of survey. The final sample consisted of 10,678 ettes?”; Cannabis: “In your lifetime have you ever used 4 A. B. KELLY ET AL. marijuana (pot, weed, grass)?”; Inhalants: “In your life- Family variables time have you ever sniffed glue, breathed the contents Poor family management was measured using 9 items of an aerosol spray can, or inhaled other gases or on a 4-point scale (1 “definitely yes” to 4 “definitely sprays, in order to get high?”; Other illicit substances: no”). Example items include “My parents ask if I’ve “In your lifetime have you ever used other illegal drugs done my homework”, “Would your parents know if (like cocaine, heroin, ecstasy, or amphetamines/ you did not come home on time”, “The rules in my speed)?”. Participants responses to the lifetime sub- family are clear” (Alphas ranged from .84 to .85 across stance use questions were dichotomised (0 “never surveys). Family conflict was measure using 3 items on used”, 1 “One or more times”). a 4-point scale (1 “definitely yes” to 4 “definitely no”): “We argue about the same things in my family over Depressed mood and over”, “People in my family have serious argu- For the 2006 and 2017 surveys, the Short Mood and ments”, and “People in my family often insult and yell Feelings Questionnaire (SMFQ; Angold et al., 1995) was at each other” (Alphas ranged from .78 to .83 across used. This is a 13-item measure of mood over the past surveys). Parents’ favourable attitude to substance use 2 weeks (alpha = 0.91). Example items include “I felt was measure using 4 items (1 “very wrong”, 2 “wrong”, miserable or unhappy” and “I didn’t enjoy anything at 3 “A little bit wrong”, 4 “Not wrong at all”): “How wrong all”, with responses ranging from 1 (“Not true”) to 3 do your parents feel it would be for you to . . . ” “smoke (“True”). Higher scores indicate a greater level of cigarettes”, “drink beer or wine regularly (at least once depressive symptoms (scores range from 0 to 26). For or twice a month)”, “drink spirits regularly? (at least the 2009 survey, the Kessler Psychological Distress once or twice a month)”, and “use marijuana (pot, Scale (K10) (Kessler et al., 2002) was used. Each of 10 weed, grass)” (Alphas ranged from 0.78 to 0.82 across items (e.g., “In the past 4 weeks, about how often did surveys). you feel hopeless/nervous/worthless?”) is rated using a 5-point Likert scale (1 “none”, 2 “a little”, 3 “some”, 4 Analyses “most”, 5 “all of the time”). K10 scores range from 10– 50 with higher indicating greater levels of psychologi- Statistical analyses were conducted using SPSS cal distress in the past 4 weeks. The internal reliability (Version 25) for the logistic regression and Mplus ver- of the K10 for the present data set was excellent sion 8.1 (Muthen & Muthen, 1998 − 2017) for the latent (Cronbach’s = 0.90). class analysis. To estimate school level effects in sub- stance use, intra-class correlations were calculated for Emotional control the three surveys and for the five substance use cate- Emotional control was measured using four items on gories. Thirteen of fifteen ICCs were below the value a four-point scale (1 “definitely yes”, 2 “yes”, 3 “no”, 4 where multilevel or mixed models are recommended “definitely No”). The four items were “I know how to (.05; Hedges & Hedberg, 2007) and the ICCs for alcohol relax when I feel tense”, “I am always able to keep my use and tobacco use were marginally over 0.05 for the feelings under control”, “I know how to calm down if 2006 survey only (.07 and .06 respectively), so school I am feeling nervous”, and “I control my temper when level effects were not modelled in subsequent people are angry with me” (Cronbach’s Alphas range analyses. 0.74–0.79). Latent class analyses were performed, respectively, on all three surveys using the lifetime substance use Academic problems variables: Tobacco, alcohol, marijuana, inhalant, and Academic failure was measured using 2 items: “Putting other illicit substance use. Fit indices included the them all together, what were your marks like last year?” Akaike Information criteria (AIC; Akaike, 1987), the (5-point scale from 1 “very good” to 5 “very poor”), and Bayesian Information Criteria (BIC; Schwarz, 1978) and “Are your school marks better than the marks of most the Sample Size-adjusted Bayesian Information Criteria students in your class?” (4-point scale from 1 “definitely (Sclove, 1987). Entropy values estimated on the aver- yes” to 4 “definitely no”). Opportunities for prosocial age posterior probabilities were used to evaluate class involvement in school was measured using five items quality, with higher values signalling clear class separa- (e.g., “In my school, students have lots of chances to tion (Muthén, 2004; Nagin, 2005). Model fit statistics help decide things like class activities and rules”, were conducted beginning with a 2-class solution and “There are lots of chances for students in my school sequentially tested up to 4 classes. To evaluate predic- to talk with a teacher one-on-one”) (Alphas 0.57 to 0.65 tors of cluster membership, stepwise logistic regres- across surveys). sion (SPSS Version 25) was used. Independent AUSTRALIAN JOURNAL OF PSYCHOLOGY 5 predictors were grouped by individual predictors a single member. A 2-class model was chosen as (Block 1: depressed mood, emotional control), family the optimal solution as it yielded classification that predictors (Block 2: poor family management, parent’s was clearly distinct and interpretable, had high attitude to substance use), and school predictors average posterior probabilities signalling clear class (Block 3: academic failure, opportunities for prosocial separation, and subgroup sizes which presented as involvement at school). School level clustering of data more interpretable for making meaningful compar- was not modelled because of the extreme zero- isons between the three years (see Table 1). Class 1 inflatedness of polydrug class counts. participants (the ‘at-risk class) were the minority (<1%, n = 18) and were predominantly high fre- quency tobacco and alcohol users and low fre- Results quency cannabis and other illicit drug users (see Figure 1). Class 2 participants (99.6%; n = 4061) Latent class analyses were predominantly low frequency alcohol users Figure 1 depicts the estimated means for lifetime sub- and reported almost nil use of other substances. stance use associated with each class across the three This class was labelled the low-risk group and desig- surveys. nated the reference group in regression analyses. 2006 survey 2009 survey The 2-class and 3-class solutions had the greatest The optimal solution was a two-class solution as it entropy values (both values of >.99), but the 3-class yielded classification that was clearly distinct and had solution was not viable as it included a class with 2017: At-risk group n = 26 2006: At-risk group n = 18 2009: At-risk group n = (1.8%), low-risk n = 1433 101(1.8%), low-risk n = 5369 (0.4%), low-risk n = 4061 (99.6 (98.2%) %) (98.2%) 10 or more times 5 5 6 to 9 times 4 4 3 to 5 times 3 3 2 2 1 or 2 times Never 1 1 1 0 0 0 Substance Type Figure 1. Estimated means for lifetime substance use across the three surveys. Notes. The same five response categories were used for each lifetime substance use question across the three datasets: 1 = “Never”, 2 = “1 or 2 times”, 3 = “3 to 5 times”, 4 = “6 to 9 times”, 5 = “10 or more times”. -♦- = low-risk class; -●- = at-risk class. Other Sub = Other substance use (cocaine, heroin, ecstasy, or amphetamines/speed). Table 1. Fit indices from latent class analyses for the three surveys. 2006 2009 2017 2-class 3-class 4-class 2-class 3-class 4-class 2-class 3-class 4-class SSBIC 13566.88 8410.58 3815.22 35687.46 28820.20 22086.81 6376.148 5001.58 3972.54 AIC 13516.70 8341.58 3727.41 35632.59 28744.76 21990.79 6342.41 4955.19 3913.49 BIC 13617.72 8480.48 3904.19 35738.31 28890.11 22175.80 6426.98 5071.47 4061.48 Entropy 1.00 1.00 0.99 0.99 0.99 0.99 0.99 0.99 0.98 Subgroup size C1 18 17 17 101 279 281 26 34 16 C2 4061 4061 3878 5369 23 22 1433 22 81 C3 1 1 5168 77 1403 144 C4 183 5090 1218 SSBIC = Sample size adjusted Bayesian information criteria. AIC = Akaike information criteria. BIC = Bayesian information criteria. C1 = Class 1; C2 = Class 2; C3 = Class 3; C4 = Class 4. Categorical Response Options 6 A. B. KELLY ET AL. strong interpretability and high average posterior family predictors were entered (Block 2), family conflict probabilities (see Table 1). Class 1 participants (“at- (p = .063), poor family management, and parents’ atti- risk” class) were the minority (1.8%, n = 101) and were tude to substance use predicted the “at-risk” class predominantly high frequency users of tobacco, alco- membership and emotional control was retained as hol, and marijuana users, and lower frequency users of a significant predictor. When school predictors were inhalants and other substance users (see Figure 1). entered (Block 3), the following predictors were signif- Class 2 participants were predominantly low frequency icant: academic failure, poor family management, and alcohol users and the use of other kinds of substances emotional control, and there was a near-significant was rare. This class was labelled the low-risk class and trend for family conflict (p = .082). Psychological dis- was used as the reference group (98.2%; n = 5369). tress, parental attitudes to substance use, and oppor- tunities for prosocial involvement (p = .381) were not significant predictors of class membership. 2017 survey As for the prior surveys, a 2-class model had the high- est entropy, interpretability, and average posterior Discussion probabilities (see Table 1). Class 1 participants (“at- This study investigated profiles of lifetime substance risk” class) were the minority (1.8%, n = 26) and were use across three population surveys, and the associa- predominantly high frequency alcohol, cannabis, inha- tion of individual, familial and school factors with these lant and other substances users (see Figure 1). They latent classes. The great majority of adolescents aged also reported elevated tobacco use. Class 2 (“low risk”) 12–14 years reported almost nil substance use. Each participants were predominantly low frequency alco- survey had a small but notable minority (between .44% hol users and used almost no other kinds of substances and 1.85%) of adolescents reporting the use of multi- (98.2%; n = 1433). ple substances (labelled polydrug users). The topogra- phy of substance use showed meaningful variations Regression analysis across the three surveys. In 2006, polydrug users reported elevated frequencies of tobacco, alcohol, We limited our regression analyses (dependent vari- and cannabis use, and one or two instances of inhalant able class membership) to the 2009 survey, where the and other (illicit) substance use. In 2009, polydrug absolute number of polydrug users was highest . The users reported elevated frequencies of tobacco and results of the logistic regression are presented in alcohol (similar rates to 2006) and more frequent can- Table 2 and summarised here. In Block 1, emotional nabis use than 2006. Inhalant use and other substance control but not psychological distress was significantly use appeared similar to 2006 rates. In 2017, polydrug associated with membership of the at-risk class. When users reported lower frequencies of tobacco use com- pared to 2006, similar levels of alcohol and cannabis Table 2. Logistic regression results showing the prediction of use, and higher frequencies of inhalant and other (illi- at-risk class membership relative to low-risk class membership. cit) substance use. Compared to the low-risk group, Variables B 95% CI R polydrug users reported poorer family management Individual predictors (Block 1) .066 practices and greater problems with emotional con- Emotional Control .29*** [1.19,1.49] Psychological Distress .01 [.97,1.04] trol. The results are consistent with the possibility that Family predictors (Block 2) .184 adolescents are at elevated risk of polydrug use when Emotional Control .15* [1.03,1.30] Psychological Distress −.02 [.95,1.02] parental awareness and supervision of adolescent Poor Family Management .15*** [1.10,1.22] activities are poor and adolescents have difficulties Family Conflict −.11‡ [.80,1.01] managing associated negative emotions. Polydrug Parents’ Attitude to Substance Use .10* [1.01,1.20] School predictors (Block 3) .201 use may be used instrumentally to control negative Emotional Control .13* [1.01,1.28] emotions and engaging with peer networks that use Psychological Distress −.02 [.95,1.02] Poor Family Management .14*** [1.09,1.22] multiple drugs may be a way of escaping distressed Family Conflict −.10† [.80,1.01] and disengaged families. Parents’ Attitude to Substance Use .09* [1.00,1.19] Academic Failure .29** [1.10,1.62] The findings on lifetime prevalence of polydrug use Opportunities for Prosocial Involvement at .06 [.92,1.20] in these Australian 12–14 year old surveys were com- School † ‡ parable to those found in Brazil (1.8%; Valente et al., p = .082, p = .063, *p < .050, **p < .010, ***p < .001. Higher scores on emotional control reflect poor abilities to manage one’s emotions. 2017) and Australian nationally representative studies Higher scores on parents’ attitude to substance use indicates more (2%; White et al., 2013), but below the prevalence rates favourable attitudes related to the offspring’s substance use. Higher scores on the family conflict scale represent lower family conflict. of polydrug use reported in other countries, including AUSTRALIAN JOURNAL OF PSYCHOLOGY 7 the United States (Cranford et al., 2013). The present cannot be established. The cell sizes for polydrug users prevalence rate was higher than prior research in were generally small and necessitated a focus on key Thailand (0.6%; Assanangkornchai et al., 2018) but variables at the exclusion of other potentially important this was unsurprising because their classes included variables. Substance use and problems amongst other a range of high-risk behaviours additional to substance family members were not assessed as part of this study, use (e.g., carrying a weapon), which is rarely reported and it is possible that these factors may be significant in Australian early adolescent populations. The similar- drivers of adolescent polydrug use. We did not include ity of findings from these surveys and prior National sex and age in our analyses because cell sizes for poly- Drug Strategy Household Survey data points to the drug users were small so it was necessary to minimise convergent validity of the present study. It is also the number of independent variables entered into each notable that the findings were relatively consistent model. Also, there were no meaningful differences in sex across informed consent mechanisms, suggesting and age across the three surveys, there was an almost that downward biases associated with active parental even split of genders across the surveys, and the age consent mechanism were minimal: The 2009 survey range for inclusion in the study (12–14 years of age) was (which used passive parental consent) had slightly small. higher polydrug use prevalence rates than both the 2006 (by 1.41%) and the 2017 survey (by 0.07%) (both Conclusion of which used active parental consent). However, the magnitude of these variations was small given that This study showed that a small proportion of 12–14- active parental consent mechanisms were associated year-old adolescents use multiple drugs and that this with the exclusion of 27.5% of the field of potential prevalence has changed very little over the last 14 participants. years, despite great population-level success in the There are significant challenges with detection and reduction of specific drug use, including tobacco and intervention for polydrug users, given the low preva- alcohol. Depressed mood and emotional control pro- lence rate for this class. Interventions for polydrug use blems characterise polydrug users. Tobacco use was might follow a stepped care approach (Toumbourou less prominent in the drug use profiles of the most et al., 2021), where students experiencing problems recent survey. The findings point to the importance of are screened for drug use and targeted individual improving early detection of at-risk adolescents and interventions are delivered by school-based health providing interventions that are tailored to meet com- professionals. Evidence-based interventions for poly- plex presentations. drug using early adolescents might require a comprehensive focus on emotional control pro- blems, strengthening school achievement, and screen- Note ing/intervention for a variety of substances. Intensive 1. We elected not to include the findings of the logis- and sustained multimodal programs addressing aca- tic regression for the 2006 survey and the 2017 demic performance through tutoring, drug use pre- survey because of the very small cell sizes of poly- vention, and family risk factors are effective (Conduct drug users (n = 18 and 26 for the 2006 and 2017 Problems Prevention Research Group, 2004). survey respectively) and the relatively large number Longitudinal research findings and controlled trials of variables in the full model (7 independent vari- point to the value of family-oriented interventions for ables). For completeness we note that for the 2006 dataset the results for the same model were largely addressing adolescent depression and drug problems consistent with the 2009 data set, with significance and this is consistent with the family risks established for emotional control, poor family management and in the present study (Chan et al., 2013; Mason et al., parental attitudes towards substance use, and non- 2003, 2012). Longitudinal research would help to significance for depressed mood. For the 2017 sur- establish dominant paths of influence, which in turn vey, none of the variables were significant. The most likely account of the pattern of nonsignifi - would guide the relative weighting program compo- cance for the 2017 survey compared to the 2009 nents and delivery modes. survey is lack of statistical power. A strength of this study is that it utilised three large- scale Australian samples, it included a mix of pathways to participation permitting redress of potential sample Disclosure statement bias (Kelly & Halford, 2007), and it focused on young adolescents (12–14 year olds) where data is scarce. The No potential conflict of interest was reported by the author(s). study is cross-sectional and so causal relationships 8 A. B. KELLY ET AL. Data availability statement parenting and school factors moderate this association? Addictive Behaviors, 64, 78–81. https://doi.org/10.1016/j. The data that support the findings of this study are available addbeh.2016.08.004 from the last author, Prof. John Toumbourou, Deakin Chan, G. C. K., Kelly, A. B., Hides, L., Quinn, C., & Williams, J. W. University, john.toumbourou@deakin.edu.au, upon reason- (2016). Does gender moderate the relationship between able request. polydrug use and sexual risk-taking among Australian secondary school students under 16 years of age? Drug and Alcohol Review, 35(6), 750–754. https://doi.org/10. Funding 1111/dar.12394 Chan, G. C. K., Kelly, A. B., & Toumbourou, J. W. (2013). Kristin R. Laurens was supported by the Australian Research Accounting for the association of family conflict and very Council [FT170100294]; National Health and Medical young adolescent female alcohol use: The role of Research Council [APP1087781]; Centre for Inclusive depressed mood. Journal of Studies on Alcohol and Drugs, Education, Queensland University of Technology, Brisbane, 74(3), 396–505. https://doi.org/10.15288/jsad.2013.74.396 Australia. Conduct Problems Prevention Research Group. (2004). The effects of the fast track program on serious problem out- comes at the end of elementary school. Journal of Clinical ORCID Child & Adolescent Psychology, 33(4), 650–661. https://doi. org/10.1207/s15374424jccp3304_1 Adrian B. Kelly http://orcid.org/0000-0001-5546-4994 Connell, C., Gilreath, T., Aklin, W., & Brex, R. (2010). Social- Xiang Zhao http://orcid.org/0000-0003-1054-9462 ecological influences on patterns of substance use among Bosco Rowland http://orcid.org/0000-0003-0192-809X non-metropolitan high school students. American Journal Kristin R. Laurens http://orcid.org/0000-0002-3987-6486 of Community Psychology, 45(1), 36–48. https://doi.org/10. Marilyn Campbell http://orcid.org/0000-0002-4477-2366 1007/s10464-009-9289-x Julie Abimanyi-Ochom http://orcid.org/0000-0002-4760- Cranford, J. A., McCabe, S. E., & Boyd, C. J. (2013). Adolescents’ nonmedical use and excessive medical use of prescription John W. Toumbourou http://orcid.org/0000-0002-8431- medications and the identification of substance use sub- groups. Addictive Behaviors, 38(11), 2768–2771. https://doi. org/10.1016/j.addbeh.2013.06.015 Göbel, K., Scheithauer, H., Bräker, A. -B., Jonkman, H., & References Soellner, R. (2016). Substance use patterns among adoles- cents in Europe: A latent class analysis. Substance Use & ABS. (2009). Socio-economic indices for areas. Misuse, 51(9), 1130–1138. https://doi.org/10.3109/ Akaike, H. (1987). Factor analysis and AIC. Psychometrika, 52, 10826084.2016.1160120 317–332. https://doi.org/10.1007/BF02294359 Hedges, L. V., & Hedberg, E. C. (2007). Intraclass correlation Angold, A., Costello, E. J., Messer, S. C., & Pickles, A. (1995). values for planning group-randomized trials in education. Development of a short questionnaire for use in epide- Educational Evaluation and Policy Analysis, 29(1), 60–87. miological studies of depression in children and https://doi.org/10.3102/0162373707299706 adolescents. International Journal of Methods in Kelly, A. B., Chan, G. C. K., Weier, M., Quinn, C., Gullo, M., Psychiatric Research, 5, 237–249. http://devepi.duhs.duke. Connor, J. P., & Hall, W. D. (2016). Parental supply of alcohol edu/AngoldMFQarticle.pdf to Australian minors: An analysis of six nationally represen- Arthur, M. W., Hawkins, J. D., Pollard, J. A., Catalano, R. F., & tative surveys spanning 15 years. BMC Public Health, 15(1), Baglioni, A. J. (2002). Measuring risk and protective factors 325. https://doi.org/10.1186/s12889-016-3004-2 for substance use, delinquency, and other adolescent pro- Kelly, A. B., & Halford, W. K. (2007). Responses to ethical blem behaviors: The Communities That Care Youth challenges in conducting research with Australian Survey. Evaluation Review, 26, 575–601. https://doi.org/ adolescents. Australian Journal of Psychology, 59(1), 10.1177/0193841X0202600601 24–33. https://doi.org/10.1080/00049530600944358 Assanangkornchai, S., Li, J., McNeil, E., & Saingam, D. (2018). Kelly, A. B., O’flaherty, M., Connor, J. P., Homel, R., Clusters of alcohol and drug use and other health-risk Toumbourou, J. W., Patton, G. C., & Williams, J. (2011). behaviors among Thai secondary school students: The influence of parents, siblings and peers on pre- and A latent class analysis. BMC Public Health, 18(1), 1272. early-teen smoking: A multilevel model. Drug and Alcohol https://doi.org/10.1186/s12889-018-6205-z Review, 30(4), 381–387. https://doi.org/10.1111/j.1465- Australian Institute of Health and Welfare. (2020). Alcohol, 3362.2010.00231.x tobacco & other drugs in Australia. https://www.aihw.gov. Kelly, A. B., Weier, M., & Hall, W. D. (2019). First use of illicit au/reports/alcohol/alcohol-tobacco-other-drugs-australia Banks, D. E., Bello, M. S., Crichlow, Q., Leventhal, A. M., Barnes- drugs: The state of current knowledge. In G. Girolamo, Najor, J. V., & Zapolski, T. C. B. (2020). Differential typolo- P. McGorry, & N. Sartorius (Eds.), Age of onset of mental gies of current substance use among black and white disorders: Etiopatogenetic and treatment implications high-school adolescents: A latent class analysis. Addictive (pp.149–168). Springer International. Behaviors, 106. https://doi.org/10.1016/j.addbeh.2020. Kessler, R. C., Andrews, G., Colpe, L. J., Hiripi, E., Mroczek, D., 106356 Normand, S. L., Walters, E., & Zaslavsky, A. (2002). Short Chan, G. C. K., Kelly, A. B., Carroll, A., & Williams, J. W. (2017). screening scales to monitor population prevalences and Peer drug use and adolescent polysubstance use: Do trends in nonspecific psychological distress. Psychological AUSTRALIAN JOURNAL OF PSYCHOLOGY 9 Medicine, 32(6), 959–976. https://doi.org/10.1017/ Sclove, S. L. (1987). Application of model-selection criteria to S0033291702006074 some problems in multivariate analysis. Psychometrika, 52 Lubman, D. I., Yücel, M., & Hall, W. D. (2016). Substance use (3), 333–343. https://doi.org/10.1007/BF02294360 and the adolescent brain: A toxic combination? Journal of Silveri, M. M., Dager, A. D., Cohen-Gilbert, J. E., & Sneider, J. T. Psychopharmacology (Oxford), 21(8), 792–794. https://doi. (2016). Neurobiological signatures associated with alcohol org/10.1177/0269881107078309 and drug use in the human adolescent brain. Neuroscience Mason, W. A., Haggerty, K. P., Fleming, A. P., & Casey- and Biobehavioral Reviews, 70, 244–259. https://doi.org/10. Goldstein, M. (2012). Family intervention to prevent 1016/j.neubiorev.2016.06.042 depression and substance use among adolescents of Squeglia, L. M., & Gray, K. M. (2016). Alcohol and drug use and depressed parents. Journal of Child and Family Studies, 21 the developing brain. Current Psychiatry Reports, 18(5), (6), 891–905. https://doi.org/10.1007/s10826-011-9549-x 1–10. https://doi.org/10.1007/s11920-016-0689-y Mason, W. A., Kosterman, R., Hawkins, J. D., Haggerty, K. P., & Toumbourou, J. W., Heerde, J., Kelly, A. B., & Bailey, J. (2021). Spoth, R. (2003). Reducing adolescents’ growth in substance Marijuana use. In W. O’Donohue & Z. Zimmerman (Eds.), use and deliquency: Randomized trial effects of a parent- Handbook of evidence-based prevention of behavioral dis- training prevention intervention. Prevention Science, 4(3), orders in integrated care: A stepped care approach (pp 251– 203–212. https://doi.org/10.1023/A:1024653923780 265, Chapter 11). Springer Nature http://doi.org/10.1007/ Masterman, P. W., & Kelly, A. B. (2003). Reaching adolescents 978-3-030-83469-2_11 . who drink harmfully: Fitting intervention to developmen- Valente, J. Y., Cogo-Moreira, H., & Sanchez, Z. M. (2017). tal reality. Journal of Substance Abuse Treatment, 24(4), Gradient of association between parenting styles and pat- 347–355. https://doi.org/10.1016/S0740-5472(03)00047-3 terns of drug use in adolescence: A latent class analysis. Miech, R. A., O’malley, P. M., Johnston, L. D., & Patrick, M. E. Drug and alcohol dependence, 180, 272–278. https://doi. (2016). E-cigarettes and the drug use patterns of org/10.1016/j.drugalcdep.2017.08.015 adolescents. Nicotine & Tobacco Research, 18(5), 654–659. White, A., Chan, G. C. K., Quek, L. H., Connor, J. P., Baker, P., https://doi.org/10.1093/ntr/ntv217 Saunders, J. B., & Kelly, A. B. (2013). The topography of Muthén, B. (2004). Latent Variable Analysis: Growth Mixture multiple drug use among adolescent Australians: Findings Modeling and Related Techniques for Longitudinal Data. from the national drug strategy household survey. In D. Kaplan (Ed.), Handbook of Quantitative Methodology Addictive Behaviors, 38(4), 2038–2073. http://doi.org/10. for the Social Sciences (pp. 345–368). Sage. https://doi.org/ 1016/j.addbeh.2013.01.001 10.4135/9781412986311.n19 White, A., Chan, G. C., Quek, L. -H., Connor, J. P., Muthén, L. K., & Muthén, B. O. (1998-2017). Mplus User’s Guide (8th ed.). Muthén & Muthén. Saunders, J. B., Baker, P., Brackenridge, C., Kelly, A. B. (2013). The topography of multiple drug use among Nagin, D. S. (2005). Group-based modeling of development. adolescent Australians: Findings from the national Harvard University Press. drug strategy household survey. Addictive Behaviors, Schwarz, G. (1978). Estimating the Dimension of a Model. Annals of Statistics, 6, 461–464. https://doi.org/10.1214/ 38(4), 2068–2073.https://doi.org/10.1016/j.addbeh.2013. aos/1176344136 01.001

Journal

Australian Journal of Psychology – Taylor & Francis

Published: Dec 31, 2023

Keywords: Adolescent; polydrug use; family relationships; family conflict; academic performance; emotional control

There are no references for this article.