Background: Investigating harms of interventions is as important as investigating benefits, with the aim of providing a benefit/harm ratio. Methods for evaluating harms are still under discussion. Objective: To investigate the methods applied in HTAs/systematic reviews (SRs) for evaluating the harms of drug interventions, specifically the study types used as data sources and the harm outcomes analyzed. Methods: HTAs/SRs on 14 drug interventions assessed by the Institute for Quality and Efficiency in Health Care (IQWiG) between 2008 and March 2009 were retrieved from the websites of the Agency for Healthcare Research and Quality (AHRQ), Canadian Agency for Drugs and Technologies in Health (CADTH), National Institute for Clinical Excellence (NICE) and from the Cochrane Database of Systematic Reviews. (The interventions had to refer to the same indications assessed by IQWiG.) Study types considered in the evaluation of harms as well as harm outcomes were extracted by one reviewer and checked by another. Results: Overall, 29 HTAs/SRs including 45 evaluations of the 14 drug interventions assessed by IQWiG were identified (AHRQ: 6 evaluations; CADTH: 5, IQWiG: 14, NICE: 9, Cochrane: 11). All evaluations in the included HTAs/SRs used randomized controlled trials (RCTs) as sources for harms data (45/45, 100%). However, only 9/45 (20%) and 6/45 (13%) considered harms data from SRs/meta-analyses and non-RCTs, respectively. Regarding the type of outcome analyzed, overall adverse event (AE) rates were reported in 34/45 (76%) evaluations (serious AEs: 33/45 (73%); withdrawals due to AEs: 34/45 (76%); specific AEs: 42/45 (93%). The impact of the choice of outcomes on the conclusions about the harms of the interventions will be presented. Conclusions: All HTAs/SRs included in our analysis used RCTs as data sources for harm outcomes. However, data from non-RCTs, SRs or meta-analyses were rarely considered. Harm outcomes included both general and specific outcomes.