The review addressed a question that was broad in scope. Inclusion and exclusion criteria were defined, but two studies were included that had participants over 21 years old. The authors used appropriate electronic databases and search terms, but did not report search dates. The restriction to published studies in English risked language and publication biases. Attempts were made to reduce reviewer error and bias throughout the review processes review. No quality assessment was reported, so the risk of bias in the included studies was unclear.
Statistical pooling of the effect sizes of different substance-use outcomes (tobacco, alcohol, marijuana and illicit drug) and different study designs may not have been appropriate. The results of non-randomised studies would produce bias and may over-estimate the result compared with randomised studies. The aggregate effect size was lower when tobacco studies were excluded.
Long-term maintenance of the effect of motivational interviewing was unclear as only seven of the 21 studies assessed outcomes over six months and it is unclear how many were assessed at one year or more.
The lack of quality assessment and pooling of data from diverse samples and settings mean that the authors' conclusion should be interpreted with some caution.