BACKGROUND: Educational outreach visits are commonly used to promote changes in prescribing in family practice. However, the effectiveness of outreach visits has not been evaluated across a range of settings. AIM: To estimate the effectiveness of educational outreach visits on United Kingdom (UK) general practice prescribing and to examine the extent to which practice characteristics influenced outcome. DESIGN OF STUDY: Randomised controlled trial. SETTING: General practices in 12 health authorities in England. METHOD: Educational outreach visits were made to practices that received two of four guidelines. Each practice provided data on treatment of patients for all four guidelines for both pre and post-intervention periods. The primary outcome is average effect across all four guidelines. Secondary analyses examined the predictive effect of practice and guideline characteristics. RESULTS: Seventy per cent of practices approached agreed to take part in the intervention. Overall, educational outreach was associated with a significant improvement in prescribing practice (odds ratio [OR] = 1.24 [95% CI = 1.07 to 1.42]), a 5.2% (95% CI = 1.7% to 8.7%) increase in the number of patients treated within the guideline recommendations. Smaller practices (two or fewer full-time equivalent practitioners) responded much more favourably to educational outreach than larger practices. Smaller practices improved their performance in line with the guidelines by 13.5% (95% CI = 6% to 20.9%) attributable to outreach, while larger practices improved by only 1.4% (95% CI = -2.4% to 5.3%, P-value for interaction <0.001). CONCLUSION: In large practices, educational outreach alone is unlikely to achieve worthwhile change. There is good evidence to support the use of educational outreach visits in small practices.