What happens when you take the Cochrane methodology and apply it to non-clinical literature?

Article type
Authors
Irvin E
Abstract
Background: The Institute for Work & Health (the Institute) is a Canadian not for profit research organization that seeks to improve employee health by advancing research. The Institute has housed the Cochrane Back Review Group since 1995 and has promoted the Cochrane systematic review methodology in its clinical research division. Until recently the Institute had primarily conducted systematic reviews to answer clinical questions, however, in 2003 our major funder commissioned a systematic review on workplace return to work strategies, launching a four year funded program to conduct systematic reviews in the area of workplace prevention strategies.

Objective: To discuss some of the challenges of using Cochrane systematic review methodology in non-clinical literature.

Methods: We have examined three non-clinical systematic reviews investigating: 1. workplace based return to work strategies, 2. the effectiveness of participatory ergonomics, and 3. the effectiveness of occupational health and safety management systems. We highlighted some of the issues that arose at each stage of the review process when attempting to use a Cochrane Systematic review approach.

Results: Differences in the literature were apparent at the outset, specifically when searching for relevant studies. Non-clinical literature covers a range of databases and a mixture of controlled and non-controlled vocabulary. This literature also raised the question of what to do with the non-peer-reviewed grey literature as much of it is not published through academic channels. In the quality appraisal phase it was clear that some of the questions of rigour would not apply and the scoring system usually employed did not adequately differentiate high and low quality studies. There were a number of issues with evidence synthesis e.g .types of research design available, heterogeneity with population, interventions used, and outcomes.

Conclusion: Further methodological research is required in this area perhaps through a workshop model; however, the lessons learned so far suggest bearing in mind the goals of a systematic review approach and a degree of flexibility.