Effect of an educational intervention to improve adverse drug reaction reporting in physicians: a cluster randomized controlled trial

Abstract

Background: The yellow-card scheme continues to be one of the principal methods for signal generation in pharmacovigilance. Nevertheless, under-reporting, one of its disadvantages, delays alert signals and has a negative influence on public health. Educational interventions in pharmacovigilance may have a positive impact on the spontaneous reporting of adverse drug reactions (ADRs).

Objectives: To assess the duration of the effect and effectiveness of an educational intervention in pharmacovigilance designed to improve ADR reporting in a robust pharmacovigilance system.

Methods: A spatial, cluster randomized controlled trial was conducted covering all National Health System physicians in the northwest of Spain and targeting those who were actively engaged in clinical practice (n = 7,498). Of these, 2,120 were assigned in three spatial clusters to the intervention group (six hospitals and 138 primary care centers) and 3,614 in four clusters to the control group (seven hospitals and 267 primary care centers). The educational intervention consisted of two complementary approaches--one active (group sessions), the other passive (educational material, reporting form)-implemented from November 2007 to December 2008, with a follow-up period of 8 months.

Results: Intervention participation was 53.7 % in a hospital setting and 60.5 % in primary care settings. ADR reporting in the intervention group increased by 65.4 % (95 % confidence interval [CI]: 8.2-153.4) across the follow-up. The ADR reporting rate per 1,000 physicians/year in the intervention group rose from 28.1 to 39.6 following the intervention (51.7 and 27.4 in the first and second 4-month period, respectively). For the intervention group, relative risk (RR) was 2.31 (95 % CI: 1.46-3.68) and 1.04 (95 % CI: 0.61-1.77) in the first and second 4-month period, respectively adjusted to baseline values. There was an increase in unexpected ADR reporting (RR 2.06, 95 % CI 1.19-3.55).

Conclusions: Pharmacovigilance educational interventions that have proved effective can be successfully applied in different geographical areas. A high baseline notification rate could account for the educational program having a moderate effect.