When a body of literature does not allow a direct comparison, one needs recourse to indirect comparisons and this can, under certain conditions, be a valid procedure (Song et al., 2003). In this review, valid indirect comparisons would require the voluntary and mandatory OHSMS studies to have been conducted in similar contexts, with similar populations, outcome measures, and contrasts. Unfortunately, the studies found do not meet these conditions.
In the course of synthesizing evidence on a particular topic, systematic reviews pool the results from homogeneous groups of studies (i.e., homogeneous with respect to study population, intervention, outcome measures, and even research design and context). If studies are very homogeneous and the number of higher quality studies sufficient, a quantitative pooling (meta-analysis) of the measured effects of the interventions is appropriate (van Tulder et al., 2003). Where higher quality studies are sparse and heterogeneous, qualitative pooling should be done.
Among the nine studies included in this systematic review, there is little in common in terms of population, intervention or outcome measures (see Tables 4.3.1 and 4.4.1). Indeed, looking within each of the voluntary and mandatory groups of studies, there is only one instance in which more than one study provided data on the same outcome (workers’ compensation premium rates in the Yassi (1998) and Alsop and LeCouteur (1999) studies). The findings of this review, therefore, require a qualitative synthesis.
The following discussion focuses separately on voluntary and mandatory OHSMSs, and describes the levels of available evidence for each.
In systematic reviews, the algorithms used for grading the level of evidence typically involve the following considerations: study design, the quality of the research, the amount of available evidence, and the consistency of the results (i.e., GRADE Working Group 2004; van Tulder et al., 2003; Briss et al., 2000; Franche et al., 2004; Kuhn et al., 1999; Tompa et al., 2004). Although this study did not adopt an explicit algorithm at its outset (for reasons described in section 3.5), similar considerations are used below to characterize the literature.
4.5.1 Evidence for the effectiveness of voluntary OHSMS interventions The studies on voluntary OHSMS interventions are relatively few, but they all show positive effects – that is, when grouped together, the studies suggest that voluntary OHSMS interventions result in desirable outcomes. These include:
increased OHSMS implementation over time (as assessed by a
Effectiveness of Occupational Health & Safety Management Systems: A Systematic Review