Realistic evaluation as a new way to design and evaluate occupational safety interventions

Recent debates regarding the criteria for evaluating occupational health and safety interventions have focused on the need for incorporating qualitative elements and process evaluation, in addition to attempting to live up to the Cochrane criteria. Reflecting fundamental epistemological conflicts and shortcomings of the Cochrane criteria in evaluating intervention studies, the debate challenges the traditional (quasi-) experimental design and methodology, which are often used within safety research. This article discusses a revised ‘realistic evaluation’ approach as a way to meet these challenges.Evidence from the literature as well as examples from an integrated (leader-based/worker-based) safety intervention study (2008-2010) in a large wood manufacturing company are presented, with focus on the pros and cons of using randomised-controlled-trials and a revised realistic evaluation model.A revised realistic evaluation model is provided which includes factors such as role behaviour, leader and worker motivation, underreporting of accidents/injuries, production pressure, unplanned organisational change and accounting for multilayer effects. These can be attained through qualitative and/or quantitative methods, allowing for the use of realistic evaluation in both large and small scale studies, as well as in systematic reviews. The revised realistic evaluation model offers a promising new way of designing and evaluating occupational safety research. This model can help safety science move forward in setting qualitative and/or quantitative criteria regarding context, mechanisms and processes for single studies and for reviews. Focus is not limited to whether the expected results appear or not, but include suggestions for what works for whom, under what circumstances, in what respects and how.