|| Checking for direct PDF access through Ovid
It is generally felt that randomised studies are difficult to conduct in the work-environment. Then again, nonrandomised studies (NRS) are considered to provide less reliable evidence for intervention effects. However, these are included in Cochrane reviews, despite discouragement. There has been no evaluation of when and how these designs are used. Therefore we conducted an overview of current practice.We included all Cochrane reviews that considered NRS. We conducted study screening and data extraction in duplicate.Of the included 202 reviews, 114 (56%) did not cite a reason for including NRS. In reviews that do cite a reason, the reasons were divided into two major categories: NRS were included because randomised trials (RCTs) are wanted (n=81, 92%) but not feasible, lacking, or insufficient alone, or because RCTs are not needed (n=7, 8%).Review authors included a range of study designs with controlled before after studies being the most common.Most interventions evaluated in Cochrane reviews incorporating NRS were non-pharmaceutical and the settings non-medical.For risk of bias assessment, most review authors (38%) used Cochrane Effective Practice and Organisation of Care (EPOC) Group’s checklists while others used a variety of checklists and self-constructed tools.Most Cochrane reviews do not justify including NRS. Where they do, the majority is not in line with Cochrane recommendations. Risk of bias assessment varies across reviews and needs improvement. We provide an algorithm for when it is useful to include NRS in systematic reviews. If interventions are implemented at the group level as is often the case in occupational health studies, it is useful to also include NRS.