Context Data on the adverse effects of newly marketed drugs are limited. Voluntary reporting is an important part of postmarketing surveillance but is underused by physicians. Objective To evaluate the effectiveness of educational outreach visits for improving adverse drug reaction (ADR) reporting by physicians. Design, Setting, and Participants A cluster-randomized controlled trial covering all National Health System physicians in the north of Portugal, with intervention in March 2004 through July 2004, and 13 to 16 months of follow-up. A total of 1388 physicians were assigned in 4 spatial clusters to the intervention group, and 5063 were assigned in 11 clusters to the control group. Intervention One-hour educational outreach visits tailored to training needs identified in a previous study. Main Outcome Measures Change in total number of reported ADRs and number of serious, high-causality, unexpected, and new-drug-related ADRs, using generalized linear mixed models adjusted for baseline ADR reporting, age, specialty, and work setting. Results At baseline, ADR reporting rates (per 1000 physician-years) did not differ significantly between the intervention groups and the control groups in reporting ADRs overall (7.6 vs 11.3), nor did they differ significantly by category: serious, 4.3 vs 6.0; high-causality, 5.4 vs 7.6; unexpected, 1.6 vs 3.5; and new-drug-related ADRs, 3.7 vs 3.8. (P>.05 for all comparisons). The control group had no significant increase in ADR reports during follow-up. The adjusted increase in ADR reporting rates attributable to intervention was 90.19 for total ADRs (95% confidence interval [CI], 54.51-125.87; relative risk [RR], 10.23; 95% CI, 3.81-27.51), 30.16 for serious ADRs (95% CI, 18.84-41.47; RR, 6.32; 95% CI, 2.09-19.16), 64.90 for high-causality ADRs (95% CI, 38.38-91.42; RR, 8.75; 95% CI, 3.05-25.07), 28.04 for unexpected ADRs (95% CI, 16.25-39.83; RR, 30.21; 95% CI, 4.54-200.84), and 42.17 for new-drug-related ADRs (95% CI, 21.58-62.76; RR, 8.05; 95% CI, 2.10 -30.83). The greatest difference occurred during the first 4 months after intervention, but differences remained statistically significant for 12 months. Conclusion A targeted outreach program may improve high-quality reporting of ADRs among physicians.