The review question was clear in terms of the study design, intervention and participants' age. The authors did not define smoking initiation, and there was variation in the definitions used in the included studies (such as ever smoked and 30-day recall). Several relevant sources were searched but, since the search strategy was restricted to trials published in English, some relevant studies might have been omitted. A limited effort was made to locate unpublished studies. There was no assessment of validity, other than to assess the included trials for blinding of the intervention. In addition, the included studies were limited by the use of self-reported outcomes and short duration of follow-up. It does not appear that any attempts were made to minimise bias, e.g. in the data extraction and study selection processes.
Insufficient description of the participants and interventions used and the lack of information on the control interventions mean that it was difficult to interpret the results. There was also insufficient information on the number of practices involved in each of the included trials. The authors noted that only one of the included trials accounted for practice-level clustering in the analysis.
The review did not provide sufficient information on the context of the included studies (e.g. whether a median income of $40 to $50,000 relates to a low or high socioeconomic position), which limits the understanding of the generalisability of the results.
The authors' conclusions about there being a lack of evidence to support smoking prevention interventions at health care practices follow on from the results of the review.