BACKGROUND: Recent studies have shown that adverse drug reactions (ADRs) are underreported. This may be particularly true of ADRs associated with complementary and alternative medicine (CAM). Data on CAM-related ADRs, however, are sparse.Objective was to evaluate the impact of an educational intervention and monitoring programme designed to improve physician reporting of ADRs in a primary care setting. METHODS: A prospective multicentre study with 38 primary care practitioners specialized in CAM was conducted from January 2004 through June 2007. After 21 month all physicians received an educational intervention in terms of face-to-face training to assist them in classifying and reporting ADRs. The study centre monitored the quantity and quality of ADR reports and analysed the results.To measure changes in the ADR reporting rate, the median number of ADR reports and interquartile range (IQR) were calculated before and after the educational intervention. The pre-intervention and post-intervention quality of the reports was assessed in terms of changes in the completeness of data provided for obligatory items. Interrater reliability between the physicians and the study centre was calculated using Cohen's kappa with a 95% confidence interval (CI). We used Mann Whitney U-test for testing continuous data and chi-square test was used for categorical data. The level of statistical significance was set at P < 0.05. RESULTS: A total of 404 ADRs were reported during the complete study period. An initial 148% increase (P = 0.001) in the number of ADR reports was observed after the educational intervention. Compared to baseline the postinterventional number of ADR reportings was statistically significant higher (P < 0.005) through the first 16 months after the intervention but not significant in the last 4-month period (median: 8.00 (IQR [2.75; 8.75]; P = 0.605). The completeness of the ADR reports increased from 80.3% before to 90.7% after the intervention. The completeness of the item for classifying ADRs as serious or non-serious increased significantly (P < 0.001) after the educational intervention. The quality of ADR reports increased from kappa 0.15 (95% CI: 0.08; 0.29) before to 0.43 (95% CI: 0.23; 0.63) after the intervention. CONCLUSION: The results of the present study demonstrate that an educational intervention can increase physician awareness of ADRs. Participating physicians were able to incorporate the knowledge they had gained from face-to-face training into their daily clinical practice. However, the effects of the intervention were temporary.