Spontaneous reporting of adverse drug reactions (ADRs) is the method most widely used by pharmacovigilance systems, with the principal limitation being the physician's underreporting.This study sought to evaluate the results of workshop and telephone-interview interventions designed to improve the quantity and relevance of ADR reporting by physicians.A cluster-randomized controlled trial was conducted on 6579 physicians in northern Portugal in 2008. Following randomization, we allocated 1034 physicians to a telephone-interview intervention, 438 to a workshop intervention and the remaining 5107 to the control group. At the workshop, a real clinical case was presented and participants were then asked to report on it by completing the relevant form. In the telephone intervention, participants were asked (i) whether they had ever had any suspicion of ADRs; (ii) whether they had experienced any difficulties in reporting; (iii) whether they remembered the different methods that could be used for reporting purposes; and (iv) whether they attached importance to the individual physician's role in reporting. We followed up physicians to assess ADR reporting rates to the Northern Pharmacovigilance Centre. In terms of relevance, adverse reactions were classified as serious or unexpected. Statistical analysis was performed on an intention-to-treat basis, and generalized linear mixed models were applied using the penalized quasi-likelihood method. The physicians studied were followed up over a period of 20 months.Two hundred physicians underwent the educational intervention. Comparison with the control group showed that the workshop intervention increased the spontaneous ADR reporting rate by an average of 4-fold (relative risk [RR] 3.97; 95% CI 3.86, 4.08; p < 0.001) across the 20 months post-intervention. Telephone interviews, in contrast, proved less efficient since they led to no significant difference (p = 0.052) vis-à-vis the control group in ADR reporting (RR 1.02; 95% CI 1.00, 1.04). The effects of the interventions on the reporting rate of serious and high-causality ADRs indicated that the RRs associated with workshops were 6.84 (95% CI 6.69, 6.98; p < 0.001) for serious ADRs and 3.58 (95% CI 3.51, 3.66; p < 0.001) for high-causality ADRs.Whereas telephone interventions only increased spontaneous reporting in the first 4 months of follow-up, workshops significantly increased both the quantity and relevance of spontaneous ADR reporting for more than 1 year.