Impact of information letters on the reporting rate of adverse drug reactions and the quality of the reports: a randomized controlled study

All 151 primary healthcare units in the Region V？stra G？taland, Sweden, were randomly allocated (1:1) to an intervention (n = 77) or a control group (n = 74). The intervention consisted of one-page ADR information letters administered at three occasions during 2008 to all physicians and nurses in the intervention units. The number of ADR reports received from the 151 units was registered, as was the quality of the reports, which was defined as high if the ADR was to be reported according to Swedish regulations, that is, if the ADR was (i) serious, (ii) unexpected, and/or (iii) related to the use of new drugs and not labelled as common in the Summary of Product Characteristics. A questionnaire was administered to evaluate if the ADR information letter had reached the intended recipient.Before the intervention, no significant differences in reporting rate or number of high quality reports could be detected between the randomization groups. In 2008, 79 reports were sent from 37 intervention units and 52 reports from 30 control units (mean number of reports per unit ± standard deviation: 1.0 ± 2.5 vs. 0.7 ± 1.2, P = 0.34). The number of high quality reports was higher in intervention units than in control units (37 vs. 15 reports, 0.5 ± 0.9 vs. 0.2 ± 0.6, P = 0.048). According to the returned questionnaires (n = 1,292, response rate 57%), more persons in the intervention than in the control group had received (29% vs. 19%, P < 0.0001) and read (31% vs. 26%, P < 0.0001) an ADR information letter.This study suggests that repeated ADR information letters to physicians and nurses do not increase the ADR reporting rate, but may increase the number of high quality reports.Clinical trials contribute greatly to knowledge on drug safety. However, uncommon adverse drug reactions (ADRs) and ADRs in certain patient groups not included in clinical trials, e.g. children and older people with many concomitant diseases and medications, cannot be expected to be detected in these trials.