Understanding the complexity, patterns, and correlates of alcohol and other substance use among young people seeking help for mental ill-health

Use of alcohol and other substances is complex and multifaceted, with substance use problems related to many individual, social and environmental factors [1]. Substance use may be particularly problematic during adolescence and early adulthood due to factors such as increased risk-taking behaviours and underdevelopment of executive functions [2]. Early substance use is associated with polysubstance use [3], as well as several long-term harms, including substance use disorders, cognitive difficulties, behavioural issues, interrupted education and social functioning, financial and legal problems, and morbidity and early mortality [4].

Problematic substance use also commonly co-occurs with mental ill-health (broad term referring to mental illness and mental health problems). In the general Australian population, the 2019 National Drug Strategy Household Survey (NDSHS) indicated that people with a mental ill-health were twice as likely to smoke cigarettes daily and 1.7 times as likely to have used any illicit substance in the past year compared to their peers without mental ill-health [5]. This discrepancy may be larger among young people. The 2017 Australian Secondary Students’ Alcohol and Drug Survey (ASSAD) found students with a diagnosis of mental disorder were three to six times more likely to report using illicit drugs compared to those without mental ill-health [6]. Similar trends were also observed in other countries such as the UK and the USA [7, 8]. The link between substance use and mental ill-health is related to shared underlying risk factors (e.g. chronic stress, trauma and adverse childhood experiences) and possible bidirectional causal associations [9‐11].

It is critical to understand substance use in youth with mental ill-health given its high prevalence, as well as the negative impact substance use can have on functioning, treatment adherence and the effects of psychotropic medications [12]. Substance use may also contribute to the development and/or worsening of severe mental ill-health such as psychosis and suicide risk [13].

Although substance use is frequently examined in population surveys and cohorts, studies focusing on substance use in young people with emerging mental ill-health are less common [14]. In clinical settings, although substance use might be acknowledged, it may not be seen as serious enough to warrant the attention given limited mental health service resources. However, this represents a significant missed window of opportunity for early identification and intervention for substance use-associated harms in this high-risk population [15]. Hence, there is a need to increase understanding of patterns of substance use and associated risk factors in this population to guide targeted early intervention and treatment approaches.

There is also an opportunity for new methodologies to better capture the nature of substance use in young people. Overall patterns of substance use and associated harms may be difficult to evaluate due to interactions between different substances [16]. Traditionally, latent class analysis has been used to classify individuals into distinct groups [17, 18]. However, this method only evaluates possible subgroups with different substance-using patterns and does not reveal independent associations between the use of different substances.

As an alternative, network analysis allows the evaluation of direct and indirect associations among factors [19, 20]. This technique is increasingly applied, however, to our knowledge, only one study by Rhemtulla and colleagues [21] has adopted this method, albeit in an adult population (to evaluate associations between symptoms of substance use disorder for an individual substance but not between substance types). Network analysis is particularly useful in evaluating complex systems to gain an understanding of influential factors within the system.

Here, we applied network analysis to explore links between use and harms (e.g. health, social, legal or financial problems) associated with different types of substance use in young people seeking help for mental ill-health. The specific aims of the study were to: (i) report prevalence and patterns of substance use in a large cohort of young people seeking primary mental health care for emerging mental ill-health; (ii) examine interrelationships between the use and harms associated with different substances; (iii) explore demographic and clinical correlates associated with the use of substances that plays a more central role among all substances.

The cohort of 1107 young people has been described elsewhere [22]. The median age of participants was 18 years (IQR 16, 20), and 65% were females. Approximately, a third of participants reported LGBTIQA+ (lesbian, gay, bisexual, transgender, intersex, queer/questioning) status. Over three-quarters of participants presented with a primary diagnosis of anxiety and/or depression, and only 2% had a primary diagnosis of substance use disorder.

Substance use and associated harms are complex phenomena and new statistical methods can help to understand their interrelationships. In this analysis, we used a novel statistical approach—network analysis—to examine interconnections between the use and harms of different substances, as well the topology of the substance-using network, to gain an understanding of substance use profiles in young people first presenting to primary youth mental health services. The results reveal the high frequency and significant interconnections of the use of different substances and their associated harms, and show that ATS and cannabis played more central roles with higher overall associations with use and harms of other substances.

The prevalence of substance use in this cohort was considerably higher than the general population. Findings from the 2019 Australian NDSHS show that only 8% of 14–19 years old and 24% of 20–29 years old have ever smoked cigarettes, whereas, in our sample, rates of tobacco use in the past 3 months for 12–17 years old and 18–25 years old were 21% and 43%, respectively [5, 35]. The prevalence of illicit substance use in the past year was reported as 15.9% for 14–19 years old and 31% for 20–29 years old in the national sample [5], whereas 19% of 12–17 years old and 40% of 18–25 years old in our sample reportedly used an illicit substance in the past 3 months. Past 3-month ATS use (8%) was also much higher compared with national figures (0.9% for 14–19 years old and 2.4% for 20–29 years old in the past year) [9].

While most participants were still at the early stage of mental ill-health [22], the interconnections observed between all substance use and harm indicators were substantial. Polysubstance use has been widely reported in youth populations [36, 37]; however, the interplay between different substances has rarely been evaluated. Recent developments in network analysis allow us to better understand the dynamic interactions among symptoms [38].

Some substance-using sub-clusters were identified in the zero-order substance use and harm correlation network, such as the connection between cannabis with tobacco and hallucinogens as well as the connection between ATS with cocaine, hallucinogens and opiates. These major connections remain strong in the partial correlation network, which indicates that these links were not being confounded by other substance use [39]. Although these links could be associated with common underlying factors, it is also possible that early interventions and treatment for one substance may directly impact other substance use by young people. However, longitudinal data are required to confirm the potential causal associations.

Among all the substance groups evaluated, ATS had the highest centrality followed by cannabis. The high centrality, in this content, indicates stronger and wider independent correlations with other substances. In this cohort, ATS use was a marker of the severe end of substance use with young people in this group frequently using other substances (such as cannabis, cocaine, hallucinogens and opiates) and experiencing social and functional impairment associated with their substance use. Similar findings are also observed from cohorts of young people from the general community [40, 41]. Although there was no strong evidence suggesting clinical symptoms were associated with ATS use when other factors were considered, this could be a result of a lack of statistical power due to the small sample size in this group.

Cannabis use also had an important role in the overall substance use network with strong links, particularly with tobacco and hallucinogens. In our cohort, 69% (114 out of 166) of daily tobacco smokers also used cannabis, which is substantially higher compared to non-daily tobacco smokers (18%). These unequal rates are equivalent to an unadjusted odds ratio of 9.8, which is much higher compared with estimates from the general population [42, 43]. Although the link between cannabis and tobacco use has been well established (e.g. gateway effect of tobacco, common co-administration pattern with forms such as blunts, spliffs or vaping, and shared genetic predispositions and environmental factors) [44, 45], this link seems to be stronger among young people with mental ill-health, warranting further investigation. Co-use, particularly co-administration, of cannabis and tobacco can lead to a greater level of harms [45, 46]. The co-use pattern may also become more significant in recent years with vaping devices becoming more popular and accessible [47, 48]. Targeted interventions, particularly delivered in mental health care settings, are needed.

Interestingly, although the primarily alcohol/tobacco group presented with the lowest level of substance use and harms among all three substance use subgroups, harms such as “lead to problems” and “failed normal expectations” were commonly reported. This indicates the high vulnerability of substance use-related harms in the mental health help-seeking youth.

Considering the central role of ATS and cannabis, we further divided the population into four groups and assessed how demographics and clinical factors were associated with substance-using patterns. Results suggest that substance use groups were most strongly related to non-clinical factors. Living in regional areas was found to be one of the most important predictors of using substances, particularly ATS. At the population level, rates of cannabis and ATS use were only slightly higher in regional areas compared with metropolitan areas [5, 49]. In our cohort, less than 6% of participants from metropolitan areas used ATS compared with more than 15% from regional areas. This substantive difference suggests the need for improved integration of care for substance and mental health, particularly in regional areas of Australia.

Another important issue to note is that substance use was associated with engagement in education and/or employment. Those who were working or both studying and working were found more likely to use substances; this was found consistently in models adjusted for other factors such as age and sex. The relative risk ratio of using ATS relative to no substance use was more than 7 times higher in the working only group compared with the studying only group after controlling for confounding factors such as age. Although this high risk may relate to access to money, it is also possible that substance use, particularly ATS use, was related to exposure through workplaces or early school disengagement in this population [40].

There are some limitations of the study. The prevalence of substance use and harms in this cohort may be biased due to participation bias or response bias [22]. Our sample shared some similarities and differences with headspace samples reported previously, but overall it was largely representative of the patient characteristics [22]. However, care must be taken in interpreting the results. The findings may not generalise to other populations in Australia and other countries. For example, a high proportion (30%) of young people reported an LGBTIQA + status, which, although comparable to 26% in the national headspace data [50], may be higher compared with other health service settings. The sample also does not represent young people seeking help, but do not have access to care. However, the analytical framework provided here can be replicated in other datasets. Only cross-sectional data were used to evaluate connections between substance use variables, which limits our ability to understand the directions of the network and longitudinal causal associations between substance use. Cross-sectional data also cannot differentiate between-individual relationships and within-individual relationships [32], which suggests that the observed associations may reflect either substance using preferences between people or temporal variations in drug of choice.

We found that at first presentation to youth primary mental health services, rates of comorbid substance use were disproportionally high compared to the general population. Moreover, a substantive subset of young people presenting for care reported frequent use of a range of substances, as well as clinically relevant harms associated with their use. Despite this, only about 2–3% of headspace clients present primarily for alcohol and other substance use [51]. This reveals the need for screening and integrated care for both substance use and mental health to accommodate the diverse needs of young people presented in the primary mental healthcare setting. As this high discrepancy may also be a result of a lack of willingness to disclose to clinicians, methods for substance use screening and assessment in clinical services should be co-designed with young people’s involvement. More integrated care could also leverage the ready presentation of young people with substance use for mental health care to facilitate early intervention for substance use problems. Clinicians should also be aware that cannabis and ATS use could be a marker for more complex and multifaceted substance-using patterns, warranting a focus on targeted treatments for these substances.