Published randomized controlled trials of drug therapy for dementia often lack complete data on harm.

Objective: The objective of the study was to determine the extent to which published randomized controlled trials (RCTs) report data on harm.
Study Design and Setting: A systematic search strategy was used to identify RCTs published between 1996 and 2005 on the use of cholinesterase inhibitors or atypical antipsychotics in patients with dementia. A structured abstraction form was used to determine if data on mortality or serious adverse events were reported and if the articles followed Consolidated Standards of Reporting Trials format for reporting harm.
Results: Thirty-three RCTs were identified (27 on cholinesterase inhibitors and 6 on atypical antipsychotics). Nineteen trials (58%) had explicit data on mortality and only four (12%) reported regulatory-agency-defined serious adverse events. Most abstracts (31, 94%) stated that harm was studied but few studies (9, 27%) provided a clear definition of the measures of harm.
Conclusions: Although most published RCTs state that they examine harm, many failed to provide data on mortality and most lacked clear definitions or detailed analyses of harm. Better reporting of harm would provide timely and important information that could help physicians and the public to make more informed decisions.