Youth drug use is a severe problem worldwide. This review focuses on Cognitive-Behavioural Therapy (CBT) as a treatment for young people who misuse non-opioid drugs, such as cannabis, amphetamines, ecstasy and cocaine, which are strongly associated with a range of health and social problems. CBT is an individualized and multicomponent intervention that combines behavioural and cognitive therapy. While behavioural therapy mainly focuses on external settings and observable behaviour, cognitive therapy is concerned with internal cognitive processes. The primary focus of CBT is to reduce users’ positive expectations about drug use, to enhance their self-confidence to resist drugs, and to improve their skills for problem-solving and for coping with daily life stressors.

The objective of this review is to assess the effectiveness of CBT for young people (aged 13-21) in outpatient treatment for non-opioid drug use and to explore any factors that may moderate outcomes.

The literature search yielded a total of 18,514 references, of which 394 were deemed potentially relevant and retrieved for eligibility determination. Of these, 360 did not fulfil the screening criteria and were excluded. Four records were unobtainable. A total of seven unique studies, reported in 17 papers, were included in the review.

Meta-analysis was used to examine the effects of CBT on drug use reduction, social and family functioning, school problems, treatment retention and criminal activity compared to a group of other interventions (Adolescent Community Reinforcement Approach (ACRA), Chestnut Bloomington Outpatient (CBOP) (+Assertive Continuing Care (ACC)), Drugs Harm Psychoeducation curriculum (DHPE), Functional Family Therapy (FFT), Interactional Therapy (IT), Multidimensional Family Therapy (MDFT), and Psychoeducational Therapy (PET)).

Our main objective was to evaluate the current evidence on the effect of CBT on abstinence and drug use reduction for young people in outpatient treatment for non-opioid drug use. Seven randomised trials, involving 953 participants, were included in this review. Each of the seven included studies compared CBT to another intervention. We analysed the effects in the short term (from the start of treatment to up to 6 months thereafter), medium term (from 6 months to less than 12 months after the start of treatment), and long term (12 months or more after the start of treatment).

We analysed CBT that was delivered with an add-on component such as motivational interviewing (four studies) separately from CBT that was delivered without an add-on component (three studies). Based on meta-analysis of data from the four included studies analysing CBT with an add-on component, there was no evidence of a relative effect of CBT for the reduction of youth drug use frequency compared to other interventions (ACRA, CBOP (+ACC), DHPE, FFT and MDFT). The random effects standardized mean difference was -0.14 (95% CI -0.64, 0.36) for the short term based on four studies, -0.06 (95% CI -0.44, 0.32) for the medium term based on four studies and -0.15 (95% CI -0.36, 0.06) for the long term based on two studies.

Based on meta-analysis of data from the four included studies analysing CBT without an add-on component, there was no evidence of a relative effect of CBT for the reduction of youth drug use frequency compared to other interventions (IT, MDFT, and PET ). The random effects standardized mean difference was -0.13 (95% CI -0.68, 0.42) for the short term based on two studies, -0.08 (95% CI -0.48, 0.31) for the medium term based on three studies and 0.02 (95% CI -0.48, 0.52) for the long term based on two studies.

Thus, the available data does not support the hypothesis that there is a drug use reduction effect from using CBT with young drug users compared to other interventions (ACRA, CBOP (+ACC), DHPE, FFT, IT, MDFT, and PET ).

Statistically significant heterogeneity was present in the short term. In the medium term statistically significant heterogeneity was present between studies analysing CBT with an add-on component. In the analysis of studies without an add-on component there was no statistically significant heterogeneity in the medium term. Due to the low power of detecting heterogeneity with only two studies included in the analysis, this result should be interpreted with caution. There was no heterogeneity between studies in the long term; however, with only two studies included in the analyses the power to detect heterogeneity was low.

The primary outcome measured as recovery could only be analysed in the long term. The meta-analysis of CBT with an add-on component was inconclusive as the eight different comparison combinations analysed showed different results. Only one study analysing CBT without an add-on component provided data on recovery status. The reported effect was not statistically significant.