Background: Guidelines are commonly used to promote quality and equity of services, but do doctors follow them, and why do or don’t they? Interventions aimed at behaviour change are complex. Controlled trials are well-suited to establish associations between interventions and outcomes, but tend to offer little information about causal processes. Process is often the primary focus of qualitative research. Objectives: To suggest how combining a review of trials with a review of qualitative studies can provide a more complete understanding of intervention effectiveness. Methods: Systematic review of controlled trials on the effect of guideline dissemination and implementation strategies. Systematic review of qualitative studies on GPs’ attitudes to and experiences with guidelines. Results: The review of trials concluded that the majority of interventions using guidelines produced modest to moderate effects. However, there was considerable variation in these effects. The studies were reanalysed on the basis of methodological quality, control group, and number of intervention components, but these attempts to explain heterogeneity of results gave no clear answers. In the review of qualitative research, GPs offered several reasons for not following guidelines: questioning the guidelines, GPs’ experience, preserving the doctor-patient relationship, professional responsibility, practical issues, and guideline format. Synthesis of these results revealed that GPs’ reasons differed according to whether the guideline in question was prescriptive (encouraged a certain type of behaviour or treatment) or proscriptive (discouraged certain treatments or behaviours). If the prescriptive-proscriptive distinction truly guides behaviour, this is an important input in future guideline development and implementation. One way of testing this hypothesis could be through a reanalysis of the trials included in Grimshaw et al’s review. Conclusions: Our examination of the two reviews suggests that a combination of their results may further our understanding of guideline compliance. Other guideline implementation research suggests that the nature of the behaviour promoted by the guideline, for instance whether it recommends increasing or reducing the behaviour, may influence its adoption. The prescriptive-proscriptive distinction may be a fundamental dimension of a typology of behaviours and could be used to explore heterogeneity in the context of quantitative synthesis.