Introduction and aims: Polysubstance use is associated with adverse health and social outcomes, but few studies have investigated whether these associations differ between individuals engaged in different patterns of illicit drug and non-prescription medication use.
Design and methods: Latent class analysis (LCA) was used to identify patterns of drug use in the Global Drug Survey, a purposive sample collected in late 2012 and surveyed using an online questionnaire including past-year drug use, sociodemographics, mental illness, involvement in violence and sexual behaviour. The sample analysed (n = 14, 869; median age 27 years; 68.5% male) included those residing in the UK (n = 5869), Australia (n = 6313) and the USA (n = 2687).
Results: LCA of cannabis, ecstasy, cocaine, stimulants, nitrous, ketamine, benzodiazepines and opioid painkiller use identified six classes: no polysubstance use (Class 1, 49.1%); cannabis and ecstasy (Class 2, 23.6%); all illicit drugs (Class 3, 9.4%); ecstasy and cocaine (Class 4, 8.3%); cannabis and medication (Class 5, 5.9%); and all drugs (Class 6, 3.8%). Participants diagnosed with anxiety were most likely to belong to Class 5 [odds ratio (OR) 2.66, 95% confidence interval (CI) 2.10-3.38]. Violent behaviour was most strongly associated with Class 6 membership (OR 1.9, 95% CI 1.36-2.64). Sexual risk-taking also predicted membership of this class (OR 5.79, 95% CI 4.66-7.18) and Class 4 (OR 4.41, 95% CI 3.57-5.43).
Discussion and conclusions: Five heterogeneous groups of polysubstance users were identified in this international sample covering the UK, Australia and USA. Anxiety disorders were associated with medication and cannabis use, while high-risk behaviours predicted use of cocaine and ecstasy, or wide-ranging polysubstance use including ketamine and medications.
Keywords: comorbidity; epidemiology; latent class analysis; polysubstance; psychopathology.
© 2015 Australasian Professional Society on Alcohol and other Drugs.