Reporting of adverse drug reactions in randomised controlled trials - a systematic survey.

Background: Decisions on treatment are guided, not only by the potential for benefit, but also by the nature and severity of adverse drug reactions. However, some researchers have found numerous deficiencies in trial reports of adverse effects. We sought to confirm these findings by evaluating trials of drug therapy published in seven eminent medical journals in 1997.

Methods: Literature review to determine whether the definition, recording and reporting of adverse drug reactions in clinical trials were in accordance with published recommendations on structured reporting.

Results: Of the 185 trials reviewed, 25 (14%) made no mention of adverse drug reactions. Data in a further 60 (32%) could not be fully evaluated, either because numbers were not given for each treatment arm (31 trials), or because a generic statement was made without full details (29 trials). When adverse drug reactions such as clinical events or patient symptoms were mentioned in the reports, details on how they had been recorded were given in only 14/95 (15%) and 18/104 (17%) trials respectively. Of the 86 trials that mentioned severity of adverse drug reactions, only 42 (49%) stated how severity had been defined. The median amount of space used for safety data in the Results and Discussion sections was 5.8%.

Conclusions: Trial reports often failed to provide details on how adverse drug reactions were defined or recorded. The absence of such methodological information makes comparative evaluation of adverse reaction rates potentially unreliable. Authors and journals should adopt recommendations on the structured reporting of adverse effects.