Objective: A systematic review was undertaken of studies that test the effectiveness of different strategies used to increase general practitioners' rates of screening for and giving advice about hazardous and harmful alcohol consumption.
Method: Resources were MEDLINE, EMBASE, Cinahl and the Cochrane Library (1966-2001). Inclusion criteria were those of the Effective Practice and Organisation of Care Group of the Cochrane Collaboration. A meta-analysis was undertaken, using a random effects model, of 15 programs identified in 12 trials. Effect sizes, calculated using the logged odds ratio, were adjusted by inverse variance weights to control for the sample sizes of the studies.
Results: Analysis of the intervention groups resulted in screening and advice-giving rates of 45% (95% CI: 33%-56%) and analysis of the comparison groups resulted in rates of 32% (95% CI: 20%-43%). The weighted mean effect size (logged odds ratio = 0.73; 95% CI: 0.56-0.90) was heterogeneous. Regression analysis to explain the heterogeneity found a significant effect for alcohol-specific programs compared with general prevention programs in which alcohol was included, and for multicomponent programs compared with single component programs. No significant differences were found between educational-based or office-based interventions.
Conclusions: Although the small numbers of programs studied suggest caution be used in interpreting the results, it seems it is possible to increase the engagement of general practitioners in screening and giving advice for hazardous and harmful alcohol consumption. Although considerably more research of high quality is needed, promising programs are those that have a specific focus on alcohol and those that are multicomponent.