|Home | About | Journals | Submit | Contact Us | Français|
To determine the prevalence of recent alcohol, nicotine or cannabis use in young persons presenting for mental healthcare.
A cross-sectional study of young people seeking mental healthcare completed self-report questionnaires regarding their use of alcohol, nicotine or cannabis.
Data were collected from two sites as part of the national headspace services programme.
2122 young people aged 12–30 years provided information as part of a patient register; a subset of N=522 participants also provided more detailed information about their patterns of alcohol use.
Prevalence levels of recent alcohol, nicotine or cannabis use within relevant age bands (12–17, 18–19 and 20–30) or primary diagnostic categories.
The rates for use at least weekly of alcohol for the three age bands were 12%, 39% and 45%, and for cannabis 7%, 14% and 18%, respectively. The rates of daily nicotine use for the three age bands were 23%, 36% and 41%. The pattern of alcohol use was characterised by few abstainers as well as many risky drinkers. Age of onset across all three substances was approximately 15 years. Individuals who used any of the three substances more frequently were likely to be older, male or have psychotic or bipolar disorders.
Frequent use of alcohol, nicotine or cannabis in young people seeking mental healthcare is common. Given the restricted legal access, the patterns of use in those aged 12–17 years are particularly notable. Reductions in substance use needs to be prioritised within services for at-risk young people.
Health prevention priorities in Australia include cessation of nicotine or cannabis use and reduction in alcohol-related harm.1–3 Among those with mental illness, reductions in substance-use-related harm and improved cardiovascular health are key clinical objectives. Epidemiological and longitudinal studies indicate that early substance misuse increases the risk of developing a mental illness and, conversely, early-onset mental disorders are associated with increased risk of alcohol or other substance misuse.4–6
Neurobiological studies increasingly demonstrate the adverse effects of alcohol or cannabis on brain development in teenagers and young adults.7 Young people with mental disorders are at increased risk of later cardiovascular disease.8 One of the most significant modifiable risk factors in this population appears to be cigarette smoking.
Within the new primary-care-based mental health initiative, headspace, the national youth mental health foundation, a strong policy emphasis is placed on the active management of alcohol or other substance-misuse problems. While it is clear that these new centres have the capacity to engage young people, the extent to which concurrent alcohol, nicotine or other substance misuse is being actively managed is yet to be identified.
From a physical health perspective, there is an urgent need to establish an early intervention agenda for these individuals who are at high risk of premature death or physical health morbidity from a range of medical conditions. Of particular note are the high rates of premature death due to cardiovascular disease in persons with depression and the prevalence of metabolic syndrome in those receiving treatments for major mood disorders. Additionally, the high rate of nicotine and cannabis use in young people with major mental health disorders puts them at risk of other smoking-related conditions.
From a mental health perspective, there is a need to better understand the early stages of substance misuse in young people with emerging mental disorders. The identification of substance-use onset and subsequent patterns of use in such young people would help early intervention approaches. In this study, we determined the rates of alcohol, nicotine or cannabis use in young persons entering mental healthcare. Additionally, we sought to assess the patterns of comorbidity with specific mental disorders.
Participants aged 16 years or older provided their own written informed consent and parental/guardian consent was obtained for those under 16 years. The University of Sydney Human Research Ethics Committee approved the study.
Participants were recruited from two headspace9 sites: (i) the Brain & Mind Research Institute, Camperdown; and (ii) Campbelltown (outer suburban, south-western Sydney). These services specialise in the assessment and early intervention of mental health problems in young people.10 11 Subjects in this study were included if they were: (i) accessing services at one of the above-mentioned headspace sites; (ii) between the ages of 12 and 30 years; and (iii) consented to be enrolled on a patient register. Participants were excluded if they did not have sufficient English-language skills. Thus, this study utilised data obtained from 2122 young people (12–30 years) who consecutively volunteered to enter the patient register between October 2007 and June 2012. A subset of 522 participants also took part in specialised neurobiological research related to clinical outcomes. Hereafter, the entire sample (N=2112) is referred to as the ‘Youth Mental Health’ (YMH) cohort. Referring clinicians were asked to determine primary and secondary diagnoses based on the Diagnostic and Statistical Manual of Mental Disorders-IV criteria. For the purposes of categorisation, clinicians were asked to select one of the following ‘primary diagnosis’ disorders: (i) ‘depression’; (ii) ‘bipolar’; (iii) ‘anxiety’ (including obsessive compulsive, generalised anxiety, agoraphobia/panic and social anxiety disorders); (iv) ‘psychosis’ (including first episode and schizophrenia); (v) ‘behavioural/developmental’ (including attention-deficit hyperactivity, conduct, oppositional defiance and impulse control disorders); (vi) ‘substance use’; (vii) ‘personality’; (viii) ‘eating disorder’; (ix) ‘autism spectrum’; (x) ‘other’; and (xi) ‘unclear, still assessing’.
The questionnaire comprised basic demographic data as well as standardised questionnaires to measure common symptoms, psychosocial functioning, disability and vocational status upon entry to the service (findings reported previously10 11). Participants were asked about their lifetime and current substance use using the first two items of the World Health Organization's ‘alcohol, smoking and substance involvement screening test’ (WHO-ASSIST).12 In item 1, participants were asked whether they ever used: (a) tobacco (nicotine) products; (b) alcoholic beverages; or (c) cannabis. Subsequently, item 2 asked whether their use of each substance in the past 3 months was: (a) ‘never’; (b) ‘once or twice’; (c) ‘monthly’; (d) ‘weekly’; (e) ‘daily or almost daily’.
As a means to compare recent alcohol, nicotine or cannabis use in the YMH cohort with that of the general population, the WHO-ASSIST item 2 data were re-categorised to align with the age bands and use categories of the 2010 National Drug Strategy Household Survey (NDSHS).13 Thus, responses were re-categorised as: (i) ‘daily or almost daily’; (ii) ‘weekly’; (iii) ‘less than weekly’; or (iv) ‘never’. That is, responses of ‘once or twice’ and ‘monthly’ were collapsed to form the ‘less than weekly’ category. Furthermore, the YMH cohort was stratified into three age groups: (i) 12–17 years; (ii) 18–19 years; and (iii) 20–30 years.
Undertaken by psychiatrists, psychologists, mental health nurses or general practitioners with training in mental health. The assessing clinician verified key aspects such as primary and secondary diagnoses/comorbidities, vocational status, medical comorbidity and general functioning via the social and occupational function assessment scale.14
For the sub-sample (N=522) only, the Alcohol Use Disorders Identification Test (AUDIT) was used to assess each participant's level of risky drinking in the past year, as well as their lifetime familiarity.15 16 Total scores range from 0 to 40, with a higher score indicating more problematic drinking. A total score of 1–7 is indicative of low-risk drinking; 8–15 indicates ‘risky drinking’ with a moderate risk of harm; 16–19 indicates a high-risk or ‘harmful’ level of alcohol consumption; and 20–40 indicates a ‘high-risk’ level.
Statistical analyses were performed using SPSS for Windows 20.0. Group differences in demographic and clinical variables were assessed via analysis of variance (ANOVA) or χ2 tests. To determine whether age was a contributing factor to any observed differences in the prevalences of recent substance use, logistic regressions were conducted with a dichotomous dependent variable of substance use (eg, ‘at least weekly’ vs ‘less than weekly’) and with diagnosis and age (and their interaction) entered as predictor variables. The forced entry method was employed and significant regression models were only accepted if the Hosmer-Lemeshow goodness-of-fit statistic was non-significant.
Among male participants, three-quarters (75.9%; 506/667) of respondents (12–30 years) reported a lifetime use of alcohol, 63.3% (401/633) a lifetime use of nicotine and more than half (57.9%; 390/673) a lifetime use of cannabis. Similarly, for female participants, the lifetime rates were 75.6% (573/758), 60.3% (290/440) and 51.3% (391/762), respectively.
For male participants, the self-reported ‘age of first use’ for each substance was as follows: (i) alcohol=14.6±2.4 years (N=456); (ii) nicotine=14.0±2.6 years (N=385); and (iii) cannabis=14.8±2.9 years (N=127). For female participants, the age of first use for each substance was: (i) alcohol=14.6±2.3 years (N=531); (ii) nicotine=14.2±2.4 years (N=414); and (iii) cannabis=15.6±3.0 years (N=126). The age of first use of cannabis was significantly earlier in male participants (F(1, 252)=5.7, p<0.05).
The prevalence of each category of alcohol use increased with age (see table 1). The prevalence of daily and weekly drinking in the youngest (12–17 years) group was notable (almost 13%). In terms of sex differences, the most substantial differences were the increased rates of daily alcohol use in male participants in the 18–19 years and 20–29 years age groups (9.3% vs 3.6% and 15.4% vs 8.4%, respectively).
Almost one-quarter (23.1%) of the younger group (12–17 years) used nicotine daily (see table 2). For the older groups, the rates were 36% and 41%, respectively. While there were no substantial differences between the younger (12–17-years-old) female and male participants in recent nicotine use, in the older groups, the proportion of daily nicotine use was higher in male participants (40.9% vs 32.0% for 18–19 years; 48.0% vs 34.6% for 20–30 years).
Notably, for the two younger groups (12–17 years and 18–19 years), the prevalence of daily cannabis use (see table 3) was higher than that for daily alcohol use (3.6% vs 1.5% and 8.8% vs 6.0%). Almost 20% of the 12–17 years group used cannabis at least once in the past 3 months compared to more than 30% of the two older groups (33.1% and 36.0%, respectively). As with alcohol and nicotine, the men in the two older age groups were more likely to report using cannabis daily.
In terms of combined substance use, across the entire sample, 17.8% (367/2063) of cases reported ‘at least weekly’ use of both alcohol and tobacco; this combination was more common in male participants (at 20.5%; 196/954) as compared to female participants (15.4%; 171/1109). Similarly, the ‘at least weekly’ use of: (i) alcohol and cannabis; and (ii) nicotine and cannabis was higher in male participants (8.5% and 12.9%, respectively) compared to female participants (5.7% and 8.8%, respectively). Notably, the combination of all three substances used ‘at least weekly’ was at 5.4% in female participants (59/1096) and at 7.9% in male participants (75/945).
The relationships between substance and primary diagnostic categories are shown in tables 44–6. Owing to some limitations (eg, unequal sample sizes) and to facilitate interpretation, ANOVA and regression analyses included only the first five diagnostic categories (ie, depression, bipolar, anxiety, psychosis and behavioural/developmental). Across the five diagnostic groups, there were significant main effects of age for the ‘at least weekly’ alcohol drinkers in both female participants (F(4,229)=7.9, p<0.001) and male participants (F(4,221)=5.9, p<0.001); the behavioural/developmental group was the youngest in both. Despite sex, those with a primary diagnosis of a bipolar spectrum disorder were the most likely to report at least weekly use of alcohol (at 39.4%). Compared to female participants, within the depression and psychosis groups, young male participants reported more frequent use of alcohol.
There were significant main effects of age across the five diagnostic groups for both female (F(4,297)=10.0, p<0.001) and male (F(4,297)=14.7, p<0.001) ‘at least weekly’ users of nicotine. Despite sex, the rate of ‘at least weekly’ use of nicotine was above 30% for four of the five main diagnostic categories, with psychosis having the highest rate at 48.7%. The greatest difference between the sexes in terms of rates of ‘at least weekly’ nicotine use was seen in the psychosis group (37.5% of female participants vs 53.8% of male participants).
There were significant main effects of age across the five diagnostic groups for both female (F(4,76)=3.0, p<0.001) and male (F(4,99)=5.9, p<0.001) ‘at least weekly’ cannabis users. Across all of the diagnostic groups, male patients were more likely to be weekly cannabis users. The group with the highest prevalence of weekly cannabis use was male patients with ‘psychosis’, at 17.5%.
Among the remaining six diagnostic groups (see tables 44–6), the ‘substance use’ group, as expected, showed the highest levels of ‘at least weekly’ use of alcohol, nicotine and cannabis. Of note, over 63% of male patients with a personality disorder reported at least weekly nicotine use, compared to 46% of female patients. Interestingly, those with ‘unclear’ diagnoses showed moderately high prevalences of ‘at least weekly’ use across all three substances.
In female participants, for nicotine, the significant model (χ2 (9)=40.9, p<0.001) included only age as a significant predictor (p<0.01). Similarly, for cannabis, the model (χ2 (9)=21.7, p=.01) only included age (p<0.05). In both models, as indicated by the β coefficients for age (0.334 and 0.308, respectively), ‘at least weekly’ users were older. The regression model for current alcohol use in female patients was not acceptable (at p<0.01). In male patients, for nicotine, the highly significant model (χ2 (9)=83.8, p<0.001) included age (p<0.001), diagnosis (p<0.01) and age-by-diagnosis (p<0.01) as predictors. For cannabis, the model (χ2 (13)=59.5, p<0.001) only included age (p<0.01). As seen in female patients, the β coefficients for age were positive in both models (0.265 and 0.162, respectively), indicating that ‘at least weekly’ users were older and, in the case of tobacco, more likely to be of a particular diagnosis (ie, psychosis; see table 5). The regression model for current alcohol use in male patients was not acceptable (at p<0.01).
Table 7 displays the demographics and social-occupational functioning of the subset (N=522) sample. As predicted, abstainers (N=159) were the youngest group; ANOVA comparing age across the five drinking categories was highly significant (F(4,521)=27.1, p<0.001). Scheffe's post hoc tests confirmed that only the abstainers’ group significantly (p<0.001) differed (in age) from any other group. After omitting the abstainers, there were no significant differences across the four drinking categories in terms of distribution of gender or age; however, there was a difference in age of psychiatric onset (F(3,289)=3.2, p<0.05).
Table 8 displays the proportions of individuals across the alcohol use categories by each primary diagnosis group. As above, ANOVA and regression analyses only included the first five diagnostic groups (see table 8). As observed in the larger sample, the five groups were found to differ in mean age (F(4,493)=25.4, p<0.001). In terms of ‘risky’ drinking (ie, hazardous or harmful or high-risk), the five diagnostic groups differed with the bipolar group having the largest proportion (46.7%) and the behavioural/developmental group having the lowest proportions of ‘risky’ drinkers (25.1%). A logistic regression with the dependent variable being ‘no or low-risk’ drinking (ie, abstainers and low-risk) versus ‘risky’ (the remaining three categories) was found to show acceptable goodness-of-fit and a significant model (χ2 (9)=27.0, p<0.001); however, there were no significant predictors (age, diagnosis, age by diagnosis).
Despite adolescence being the peak period for the onset of both mental and substance misuse disorders, primary-care-based studies indicate that these problems are not being effectively managed in young people. National community-based surveys have indicated the extent to which young people do not access care for either mental disorders or alcohol/substance misuse, and this lack of care is most evident for young men and for those with alcohol/substance misuse.17
This is the first study to examine substance-use patterns in a large sample (2000+) of young people accessing headspace services (albeit across two sites within the metropolitan city of Sydney). The data presented here demonstrate that among young people who present for mental healthcare within the headspace network, alcohol, nicotine or cannabis uses are common. Given the comorbidity with significant mental health problems, these patterns of substance use are likely to contribute to increased risk of poor physical and/or mental health outcomes. Future studies will help to determine the representativeness of these data by evaluating other samples (currently, there are more than 40 headspace sites nationally).
Compared to their age-matched peers in the general population (ie, based on the findings of the 2010 NDSHS13), the patients in this study showed some differences in the rates of current substance use. Of note, our youngest group (12–17 years) was twice as likely to report weekly alcohol use compared to 12–17-year- olds in the general population (ie, 10.3% vs 5.1%; see table 4.3 in ref. 13). These comparisons should be treated somewhat cautiously as the NDSHS13 determined the frequency of use over the past 12 months, whereas our study assessed the past 3 months; furthermore, the NDSHS evaluated ‘daily’ use, whereas we asked about ‘daily or almost daily’ use. Noting such limitations, ‘daily’ alcohol use in our YMH samples was at least three times greater than that observed in the general population (ie, across the age groups13). Similarly, ‘daily’ nicotine use in our YMH samples is at least twice as high as the general population estimates (again across the three age groups; see table 3.3 in ref. 13). With regard to ‘any’ recent cannabis use, our older YMH samples (ie, 18–19 and 20–30 years) were 1.5 times more likely to report recent cannabis use compared to their peers in the general population (see table 6.4 in ref. 13).
In this study, the relationships between recent substance use and diagnosis were complex and were mainly affected by age and gender. The logistic regression models for nicotine or cannabis use were acceptable, and in three of these models, only age was a significant predictor of weekly use, the exception being nicotine use in male patients where age and diagnosis both contributed. The difference in the prevalences of weekly nicotine use among male patients with an anxiety disorder as compared to male patients with a psychotic disorder is notable. In general, weekly substance use appeared to be more likely if an individual was an older male and diagnosed with psychosis or bipolar disorder.
Information regarding the prevalence of frequent substance use among younger people seeking mental healthcare is limited. There is a general consensus that substance misuse in individuals with a mental disorder is common,6 18 19 particularly in treatment-seeking populations, with evidence to suggest that ‘problematic substance use is the most common comorbid condition among people with a major mental illness and is associated with poorer patient outcomes’.20 In a large cross-sectional study of over 45 000 Australians attending primary care, 12% of respondents were identified as having any mental disorder with concurrent substance misuse.18
Previous research has tended to evaluate the comorbidity of mental and substance-use disorder in broad adult samples (ie, ‘18 and above’). Despite this, the key associations we have observed in the current YMH cohort are consistent with those observed in the literature. For example, a study of over 40 000 adults in the USA found that comorbid substance use and mood/anxiety disorders are among the most prevalent of psychiatric disorders;21 and it was notable that bipolar disorders were more strongly related to the substance-use disorders than any other mood or anxiety disorder. Similarly, other population-based studies have demonstrated stronger associations between psychosis and nicotine22 or cannabis use.23
Of considerable relevance to preventing later poor physical health, at least weekly nicotine use was highly prevalent in those with psychosis (almost 50%). In an Australian study24 of 1812 individuals with severe psychotic disorders, for those aged 18–24 years, the rate of nicotine use was 70.6%. When adopting the same parameters, albeit over the past 3 months, the current study yields a rate of 62.6% among young people who are early in the course of their disorder.
The presence of an anxiety or mood disorder has been shown to be the largest determinant of treatment seeking in cannabis users, regardless of the level of use.22 A recent study,25 utilising a cohort of Australian secondary school students, reported that by the time participants reached 29 years, their daily cannabis use was significantly associated with an anxiety disorder. In the current study, frequent cannabis use was particularly common in female participants with an affective disorder, whereas for male participants other diagnoses (including psychosis and bipolar disorder) tended to be associated with increased rates of frequent use.
Among the subset of YMH patients (N=522), the prevalence of low-risk drinkers (34%) is the same as that predicted in the population.13 However, the rates of risky (ie, hazardous: 20% plus harmful: 7% plus high risk: 9%) YMH drinkers are substantially higher than in the general population (23%).13 Notably, other researchers26 27 have found that hazardous levels of alcohol use tend to peak in the 20–25-year age range, suggesting that, if left untreated, there may be an escalation in problematic use in the risky drinking groups. With regard to risky drinking categories, key diagnostic groups (ie, bipolar and psychosis) tended to have higher rates of risky drinkers.
This study is limited by several factors. First, the key substance-use measures were self-report only and not confirmed by interview. Furthermore, participants were not asked about the amount of nicotine and cannabis use; only the recent (past 3 months) frequency of use was asked for. Follow-up longitudinal studies of these patients would be important to determine the long-term patterns of such substance use. Despite these limitations, this study shows that the frequent use of alcohol, nicotine or cannabis in young people seeking mental healthcare is common. Given the restricted legal access, and in comparison to their peers in the general population, the patterns of use in the YMH patients aged 12–17 years are particularly notable. Reductions in the use of these substances need to be prioritised within services provided to these at-risk young people. Traditionally, mental health services have been separate to interventions that target substance use; however, there are growing suggestions that complex young people (with comorbid mental health and substance-use problems) would be most receptive to integrated rather than sequential or parallel approaches.28 29
Contributors: DFH, EMS, SLN and IBH designed the study and wrote the protocol. DFH and ML reviewed the literature; DFH, DW and ML conducted the statistical analyses. DFH, ML, DW and IBH drafted the manuscript. DFH, DW, BGW and SLN were involved in study coordination, administration of neuropsychological and data analyses. JL and BGW contributed to the study interpretation and drafts of the manuscript. All authors contributed to and have approved the final manuscript.
Funding: Funding for the headspace project comes from the Australian Government. The evaluation of these services is supported by an NHMRC Australian Fellowship to IBH (No. 464914). DFH was supported by a grant from the NSW Ministry of Health, Mental Health and Drug & Alcohol Office. SLN was funded by an NHMRC Clinical Research Fellowship (No. 402864). This research was further supported by NHMRC Program Grant (No. 566529) and Centres of Clinical Research Excellence Grant (No. 264611). These funding agencies had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.
Competing interests: DFH is currently supported by a grant from the NSW Ministry of Health, Mental Health and Drug & Alcohol Office as well as an NHMRC Australia Fellowship (awarded to Professor Hickie). In 2007, he received honoraria for educational seminars from Janssen-Cilag. EMS is the (unpaid) Clinical Director of Headspace Services at the BMRI, the (unpaid) Co-ordinator of the Youth Mental Health Research Program at the BMRI and Deputy Director of St Vincent's Private Hospital Young Adult Mental Health Unit. She has received honoraria for educational seminars related to the clinical management of depressive disorders supported by Servier and Eli-Lilly pharmaceuticals. She has participated in a national advisory board for the antidepressant compound Pristiq, manufactured by Pfizer. IBH is a member of the Medical Advisory Panel for BUPA Health Insurance (Australia) and also a Board Member of Psychosis Australia Trust. From 2012, he is a Commissioner in Australia's new National Mental Health Commission. He was until January 2012 a director of headspace: the national youth mental health foundation. IBH was previously the chief executive officer (till 2003) and clinical adviser (till 2006) of beyondblue, an Australian National Depression Initiative. He is supported principally for clinical research in depression and health services and population health initiatives related to anxiety and depression by an NHMRC Australian Medical Research Fellowship (2007–2012). He has led projects for health professionals and the community supported by governmental, community agency and pharmaceutical industry partners (Wyeth, Eli Lily, Servier, Pfizer, AstraZeneca) for the identification and management of depression and anxiety. He has received honoraria for presentations of his own work at educational seminars supported by a number of non-government organisations and the pharmaceutical industry (including Pfizer, Servier and Astra Zeneca). He has served on advisory boards convened by the pharmaceutical industry in relation to specific antidepressants, including nefazodone, duloxetine and desvenlafaxine. He leads a new investigator-initiated study of the effects of agomelatine on circadian parameters (supported in part by Servier but also by other NHMRC funding) and has participated in a multicentre clinical trial of the effects of agomelatine on sleep architecture in depression and a Servier-supported study of major depression and sleep disturbance in primary care settings. In addition to national and international government-based grant bodies, investigator-initiated mental health research at the BMRI, he has been supported by various pharmaceutical manufacturers (including Servier and Pfizer) and not-for-profit entities (including the Heart Foundation, beyondblue and the BUPA Foundation).
Ethics approval: The University of Sydney Human Research Ethics Committee.
Provenance and peer review: Not commissioned; externally peer reviewed.
Data sharing statement: No additional data are available.