Methods for including non-randomised studies to answer questions about unintended harms of drugs

Article type
Authors
Reeves B1, Shea B2, Wells G3, Loke Y4
1Non-Randomised Studies Methods Group, United Kingdom
2CIET Canada, Canada
3Ottawa Heart Institute, Canada
4University of East Anglia, UK
Abstract
Objective:

The workshop will present a framework for considering harms of interventions and propose methods for including non-randomised studies (NRS) to answer questions about unintended harms of drugs.

Description:

The need to include NRS is likely to arise when suspected harms are serious and rare or occur in the long term, situations which are rarely addressed by randomised controlled trials (RCTs). Nevertheless, including evidence about such harms is important in order to present a balanced review of the effects of an intervention, as highlighted by stakeholders in The Cochrane Collaboration. The facilitator will first present the proposed framework. Participants will work in small groups to discuss the steps identified in the framework: • at the protocol stage, prioritising harms to review; • searching for NRS about the nominated harms; • defining eligible NRS; • evidence synthesis; and • interpreting heterogeneity. Small groups will feedback their conclusions. Then all participants will consider whether there is sufficient consensus about the methods to recommend that they be proposed as guidance to Cochrane Review Groups. Participants interested in incorporating evidence from RCTs on adverse effects may be interested in the workshop, ‘Incorporating data on adverse effects into systematic reviews’.