The Effectiveness of Occupational Health and Safety Management System Interventions: A Systematic Review

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/223545678
The effectiveness of occupational health and safety management system

interventions: A systematic review
Article in Safety Science · March 2007

DOI: 10.1016/j.ssci.2006.07.003
CITATIONS READS
311 13,183
9 authors, including:
Kimberley L Cullen Amber Bielecky

Memorial University of Newfoundland Institute for Work and Health
33 PUBLICATIONS 1,694 CITATIONS 23 PUBLICATIONS 604 CITATIONS
SEE PROFILE SEE PROFILE
Emma Irvin Anthony John Culyer

Institute for Work and Health The University of York
86 PUBLICATIONS 5,109 CITATIONS 404 PUBLICATIONS 7,385 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Determinants of work particpation in older workers: a systematic review of the quantitative literature View project
Dictionary View project
All content following this page was uploaded by Anthony John Culyer on 20 July 2016.
The user has requested enhancement of the downloaded file.

Safety Science 45 (2007) 329–353
www.elsevier.com/locate/ssci
Review
The effectiveness of occupational health and safety

management system interventions:
A systematic review
Lynda S. Robson a,*, Judith A. Clarke a, Kimberley Cullen a,
Amber Bielecky a, Colette Severin a, Philip L. Bigelow a,b,
Emma Irvin a, Anthony Culyer a,c, Quenby Mahood a
a
Institute for Work and Health, 481 University Ave., Ste. 800, Toronto, ON, Canada M5G 2E9
b
Department of Public Health Sciences, University of Toronto, 6th Floor, Health Sciences Building,
155 College Street, Toronto, ON, Canada M5T 3M7
c
Department of Economics and Related Studies, University of York, Heslington,
York Y01 5DD, England, UK
Received 23 December 2005; received in revised form 12 May 2006; accepted 6 July 2006
Abstract
A variety of OHSMS-based standards, guidelines, and audits has been developed and dissemi-
nated over the past 20 years. A good understanding of the impact of these systems is timely. This
systematic literature review aimed to synthesize the best available evidence on the effects of OHSMS
interventions on employee health and safety and associated economic outcomes. Eight bibliographic
databases covering a wide range of fields were searched. Twenty-three articles met the study’s rele-
vance criteria. Thirteen of these met the methodological quality criteria. Only one of these 13 original
studies was judged to be of high methodological quality; the remainder had moderate limitations.
The studies’ results were generally positive. There were some null findings but no negative findings.
In spite of these promising results, the review concluded that the body of evidence was insufficient to
make recommendations either in favour of or against OHSMSs. This was due to: the heterogeneity
*
Corresponding author. Tel.: +1 416 927 2027; fax: +1 416 927 4167.
E-mail addresses: lrobson@iwh.on.ca (L.S. Robson), jclarke@iwh.on.ca (J.A. Clarke), kcullen@iwh.on.ca (K.
Cullen), abielecky@iwh.on.ca (A. Bielecky), cseverin@iwh.on.ca (C. Severin), pbigelow@iwh.on.ca (P.L.
Bigelow), eirvin@iwh.on.ca (E. Irvin), aculyer@iwh.on.ca, ajc17@york.ac.uk (A. Culyer), qmahood@iwh.on.ca
(Q. Mahood).
0925-7535/$ - see front matter 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ssci.2006.07.003
330 L.S. Robson et al. / Safety Science 45 (2007) 329–353
of the methods employed and the OHSMSs studied in the original studies; the small number of
studies; their generally weak methodological quality; and the lack of generalizability of many of
the studies.
2006 Elsevier Ltd. All rights reserved.
Keywords: Management system; Systematic literature review; Audit; Intervention; Effectiveness
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
1.1. Defining OHSMSs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
1.2. Voluntary and mandatory OHSMS initiatives . . . . . . . . . . . . . . . . . . . . . . . . . 332
1.3. Background literature and research justification . . . . . . . . . . . . . . . . . . . . . . . . 333
1.4. Conceptual framework of the review. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
2. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
2.1. Literature search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
2.2. Selection of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
2.3. Quality appraisal (QA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
2.4. Data extraction (DE). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
2.5. Evidence synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
3. Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
3.1. Publications identified through the literature search and quality assessment . . . . 338
3.2. Quality of the literature investigating OHSMSs . . . . . . . . . . . . . . . . . . . . . . . . 338
3.3. Findings on the implementation and effectiveness of voluntary OHSMSs . . . . . . 340
3.3.1. Description of the studies on the implementation and effectiveness of
voluntary OHSMSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
3.3.2. Findings on the implementation of voluntary OHSMSs . . . . . . . . . . . . . 341
3.3.3. Findings on the intermediate outcomes of voluntary OHSMSs . . . . . . . . 341
3.3.4. Findings on the final outcomes of voluntary OHSMSs . . . . . . . . . . . . . . 342
3.4. Findings on the implementation and effectiveness of mandatory OHSMSs . . . . . 342
3.4.1. Description of the studies on the implementation and effectiveness
of mandatory OHSMSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
3.4.2. Findings on the implementation and effectiveness of the IC legislation
in Norway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
3.4.3. Findings on the effectiveness of simple OHSMS legislation in Canada . . . 343
3.5. Evidence synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
3.5.1. Evidence for the effectiveness of voluntary OHSMS interventions . . . . . . 344
3.5.2. Evidence for the effectiveness of mandatory OHSMS interventions . . . . . 346
4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
4.1. Identifying and addressing research gaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
4.2. Strengths of the review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
4.3. Limitations of the review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
L.S. Robson et al. / Safety Science 45 (2007) 329–353 331
1. Introduction
The concept of an occupational health and safety management system (OHSMS) has
become common over the past 20 years. A variety of OHSMS-based standards, guidelines,
and audits have been developed (British Standards Institution, 1996, 1999; Dalrymple et al.,
1998; Frick et al., 2000; Gallagher et al., 2003; Grayham and del Rosario, 1997; HSE, 1997;
ILO, 2001; Standards Australia and Standards New Zealand, 1997) within the public, pri-
vate and not-for-profit sectors and many have been adopted by workplaces. Some countries,
including Canada, are in the process of developing management standards for occupational
health and safety. An understanding of the impact of these systems is therefore timely.
This systematic literature review investigates the effectiveness of mandatory and volun-
tary OHSMS interventions on employee health and safety and associated economic
outcomes. The scope of the topic was developed through formal and informal consulta-
tions with representatives of employers, labour, and several public and not-for-profit insti-
tutions. A secondary aim of the review is to characterize the content and methodology of
the existing research literature on OHSMSs. The purpose here is to identify gaps and
weaknesses in the literature, whose addressing through future research could help inform
the primary research question.
There are no other systematic literature reviews on the topic of OHSMSs, although some
narrative reviews exist (Frick et al., 2000; Gallagher et al., 2003; Saksvik and Quinlan, 2003;
Walters, 2002). A systematic literature review uses explicit, thorough methods to identify,
select, appraise and synthesize a set of research studies on a well-defined topic. This meth-
odology makes the findings of a systematic review less vulnerable to the biases of a single
researcher than those of a narrative review. Systematic reviews aid decision-makers by sift-
ing through an enormous literature to find the high-quality studies and to synthesize them.
This paper will first discuss some of the concepts and background literature relevant
to the review.
1.1. Defining OHSMSs
There is no consensus on what an OHSMS is and its scope is potentially wide. Some
definitions are simply too vague to be helpful in determining which literature should be
included in a systematic review: e.g., the definition used by the International Labour Orga-
nization (ILO, 2001):
‘‘A set of interrelated or interacting elements to establish OSH policy and objectives,
and to achieve those objectives.’’
It is not clear from the definition whether the management system includes only man-
agement components or technical/operational components as well. This problem of
demarcating the scope of a management system has also been noted by Nielsen (2000):
‘‘OHSM systems are not, of course, a well-defined set of management systems.

Indeed there are not clear boundaries between OHS activities, OHS management,
and OHSM systems.’’
OHSMSs, commonly understood, are distinguishable from traditional OHS programs

by being more proactive, better internally integrated and by incorporating elements of
evaluation and continuous improvement. Some OHSMS documents (e.g., Chemical
Industries Association, 1995; HSE, 1997; ILO, 2001) explicitly ascribe their basic source
as the Plan-Do-Check-Act model of continuous quality improvement made famous by
W. Edwards Deming (Tortorella, 1995). In contrast, traditional OHS programs have rel-
atively less activity corresponding to the Check and Act domains of the Deming Model.
Furthermore, action tends to be in response to workplace accidents, legislation, or
enforcement, rather than proactive.
Redinger and Levine (1998) considered what constitutes an OHSMS in detail. After
reviewing 13 publicly available management system documents for occupational health
and safety, environment, or quality, they selected four of the most comprehensive from
which to construct an integrative, universal OHSMS model containing 27 elements. The
16 primary elements of their model are:
• management commitment and resources,

• employee participation,
• occupational health and safety policy,
• goals and objectives,
• performance measures,
• system planning and development,
• OHSMS manual and procedures,
• training system,
• hazard control system,
• preventive and corrective action system,
• procurement and contracting,
• communication system,
• evaluation system,
• continual improvement,
• integration,
• management review.
1.2. Voluntary and mandatory OHSMS initiatives
OHSMS initiatives are either mandatory or voluntary. Mandatory OHSMSs arise from
government legislation and their use is enforced through inspections, fines, etc. Voluntary
OHSMSs arise through private enterprise, employer groups, government and its agencies,
insurance carriers, professional organizations, standards associations and are not directly
linked to regulatory requirements. Their use is not required by governments; instead,
incentives are sometimes offered by governments or insurance carriers to organizations
that voluntarily adopt particular OHSMSs.
Many voluntary OHSMSs, especially those marketed through commercial industries, are
most frequently observed in large companies. They are characterized by being more thor-
oughly specified, and as a result, are considered to be too complex for the majority of (smal-
ler) employers (Frick and Wren, 2000). Voluntary OHSMS schemes marketed through
public agencies, however, target not only large companies but also smaller ones (Frick
and Wren, 2000). These schemes either involve simpler OHSMSs or have a menu of options,
including simple ones, for companies of different sizes or at different stages of OHSMS
development. Mandatory OHSMSs are simpler in terms of the demands placed on organi-
zations, since they are intended for all or most workplaces, including small ones.
1.3. Background literature and research justification
There is a body of research on the correlates of low injury rates in organizations (e.g.,
Cohen, 1977; Habeck et al., 1998; LaMontagne et al., 1996; Reilly et al., 1995; Shannon
et al., 1996). These studies suggest which potential elements of an OHSMS are important
by identifying those that are correlated with low injury rates. Other studies (Mearns et al.,
2003; Simard and Marchand, 1994) have developed researcher-defined measures of
OHSMSs and shown that a more developed OHSMS is correlated with a lower injury rate.
While these studies are a valuable contribution to the literature, they cannot tell us what
the likely effect is of a particular type of OHSMS intervention in a particular type of work-
place. They are limited by their cross-sectional design and the lack of an OHSMS interven-
tion variable.
A separate stream of research and practice related to safety management systems in
high hazard and high reliability operations, such as in the nuclear, chemical process,
and airline, rail and marine transportation industries, has also developed (e.g., Figuera
and López, 2003; Hale, 2003; IAEA, 2005; SAMRAIL Consortium, 2004). The concept
of a safety management system overlaps with that of an OHS management system, but
is generally distinct. For instance, the scope of concern for a safety management system,
unlike that of an OHSMS, extends beyond workers to include the physical work environ-
ment and the surrounding community. Furthermore, the scope of an OHSMS covers a
broad range of workers health concerns, in contrast to that of a safety management system
which focuses on preventing traumatic injuries related to the loss of control of processes.
Many OHS practitioners presume that OHSMS interventions will be effective in lowering
injury and illness. After all, OHSMS standards and guidelines synthesize expert knowledge,
much of which is consistent with the research cited in the first paragraph of this section.
However, the effectiveness of interventions cannot be presumed in all cases. Indeed, the fail-
ure rate of quality management systems has been documented as ranging from 67% to 93%
(Gardner, 2000). There is reason to expect that the failure rate of OHSMSs would be at least
as high. Typically, the level of management commitment to high product or service quality is
higher than to employee health and safety. The effectiveness of mandatory OHSMS strategies
has also been doubted by Quinlan and Mayhew (2000) in light of the current trends of the
globalization of business, the casualization of the labour force and declining unionization.
Some criticisms of OHSMSs have also emerged in more academic circles. Suggested del-
eterious effects of OHSMSs in general or of particular types of OHSMSs have included:
the weakening of external regulatory approaches (Bennett, 2002); a false sense of security
derived from the presence of a formal OHSMS (Gallagher et al., 2003); the development of
blame-the-worker attitudes (Nichols and Tucker, 2000; Wokutch and VanSandt, 2000);
and a shift in the power balance away from workers and toward management (Lund,
2004; Nichols and Tucker).
The success of OHSMSs is likely to be dependent on the nature of the intervention,
characteristics of the workplace and characteristics of the external environment. This
review synthesizes the research results on the effectiveness of a wide range of OHSMS
interventions under a wide range of conditions.
1.4. Conceptual framework of the review
The conceptual framework underlying the review is depicted in Fig. 1. The review
included interventions directed at developing an OHSMS in one or more workplaces. It
therefore included studies of extra-workplace initiatives arising from legislation, or volun-
tary programs arising through the government, its agencies, insurance carriers, groups of
employers, etc. It also included studies of workplace-level initiatives, through which a
workplace might attempt to improve its OHSMS, using either a scheme developed exter-
nally or internally.
A minimalist operational definition of an OHSMS intervention was adopted to screen
the research literature. For a study to be included, an intervention was required to address
two or more of the 27 elements in the Redinger and Levine (1998) universal OHSMS
framework, with at least one of these being a management element.
While the primary focus of the review was on the effectiveness of OHSMSs, evidence
was also sought about implementation at the workplace level. There were two reasons
for this. First, it was anticipated that for some mandatory initiatives there might be mea-
sures only of the OHSMS rather than of its effects in workplaces. Second, implementation
information would allow one to distinguish between two possible explanations for an
absence of effect, namely poor intervention content or poor implementation of the
intervention. It is possible for a well-conceived intervention to fail through poor
implementation.
One could measure implementation in extra-workplace OHSMS initiatives (e.g., legis-
lative initiatives) through indicators of extra-workplace implementation (e.g., labour
inspector orders pertaining uniquely to the legislation). In this review, however, implemen-
tation was considered only at the workplace level; specifically, a change in the state of the
workplaces’ OHSMSs.
Final outcomes were identified by considering the ultimate purpose of OHSMS inter-
ventions. This would be improved employee health and safety for many stakeholders.
For others, the associated economic benefits were also of interest. On these grounds,
two types of final outcomes were studied in the review: OHS outcomes and economic
outcomes. Examples of final OHS outcomes are changes in rates of employee injury or ill-
ness. Examples of economic outcomes are changes in workplace workers’ compensation
premium rates and workplace productivity.
INTERVENTION IMPLEMENTATION INTERMEDIATE FINAL OHS

OHS OUTCOMES OUTCOMES
Workplace or extra- Change in workplace
workplace initiative OHSMS (e.g., safety climate) (e.g., injury rates)
ECONOMIC
OUTCOMES
(e.g., firm insurance
premiums)
Fig. 1. Conceptual framework for the review on OHSMSs.

Intermediate OHS outcomes were considered to be outcomes of secondary interest,

though they are proxies potentially for final OHS outcomes. They are changes in media-
tors between the OHSMS and final OHS outcome, such as safety climate; employee
knowledge, beliefs, values or perceptions; employee behaviours; or OHS hazards.
2. Methods
The methodological steps were: (1) search of the literature; (2) selection of relevant
studies by applying inclusion and exclusion criteria; (3) appraisal of the quality of the
research evidence in the studies; (4) extraction of the higher quality evidence from the stud-
ies; and (5) a synthesis of the higher quality evidence.
2.1. Literature search
Eight electronic databases, abstracting primarily peer-reviewed research journal articles,

were searched from their inception until July 2004: MEDLINE (from 1966), EMBASE
(from 1980), PsycInfo (from 1887), Sociological Abstracts (from 1963), CCInfoWeb (con-
sisting of NIOSHTIC-2, HSELINE, and OSHLINE), Safety Science and Risk Abstracts
(SSRA, from 1981), EconLit (from 1969), and American Business Inform (ABI, from
1918). Since the search terms and language of the databases were found to differ signifi-
cantly, the terms used in the search were customized for each database.
The search strategy combined two sets of keywords using an ‘‘AND’’ strategy. The first
set of keywords focused on a set of 30 OHSMS terms (e.g., health and safety management
system(s), safety management system(s), systematic occupational health and safety man-
agement, OHS program(s), OHSAS 18001, OHSMS, BS8800, International Safety Rating
System(s), safety program(s), occupational health/safety standard(s), safety and health
legislation). The keywords in the second set included 35 evaluation or OHS effect terms
(e.g., evaluat(ing/ion/ions/e/es/ed), program/me evaluat(ing/ion/ions/e/es/ed), implemen-
tat(ing/ion/ed/es), effect(s), impact(s), climate, culture, perception(s), behavio(u)r, work-
place injur(y/ies), injury, occupational health, occupational exposure, occupational
accident, compensation cost(s), compensation claims cost(s), time loss/lost). The full
search strategy is available in Robson et al. (2005) or from the authors. The terms within
each group were combined using an ‘‘OR’’ strategy. When possible, the titles, abstracts,
case registry, and subject headings were searched for keywords. The search strategy was
simplified for CCInfoWeb and ABI, which had simpler search capacities.
After merging the citations from the electronic search of the eight databases and remov-
ing duplicates, 4837 remained for inclusion in the review. Four additional sources were
used to identify citations of potential relevance: the reference lists of all journal articles
meeting the eligibility criteria; the reference lists of pertinent reviews published in the form
of books, book chapters, or journal articles; bibliographies requested from experts in the
area of OHSMSs; and the personal files of the authors.
2.2. Selection of studies
Titles, key words and abstracts of each article were independently screened by two
reviewers. Full text articles were retrieved for those studies appearing to meet the eligibility
criteria, and for those where the information in the title, abstract, and key words was
insufficient for exclusion. Upon retrieval of the full text article, the eligibility of a study
was determined again through a pair of reviewers. A consensus method was used to
resolve any disagreements between the reviewers about inclusion. A third reviewer was
used if the pair could not reach a consensus on eligibility.
The full set of inclusion and exclusion criteria used to determine the eligibility of studies
is described below:
Publication type. Only peer-reviewed journal articles were included.

Population of interest. Workplaces could be located anywhere in the world.
Nature of intervention. An OHSMS intervention, initiated at either the workplace level
or extra-workplace level was required. An OHSMS intervention was identified by one
of three means:
(i) directly, by a term synonymous with OHSMS or mention of specific types of
OHSMS (e.g., ‘safety and health management system’, OHSAS 18001);
(ii) indirectly, by mention of OHSMS legislation or other extra-workplace OHSMS
initiatives (e.g., European Framework Directive 89/391, Internal Control); or
(iii) indirectly, by a term suggestive of OHSMS, and a description of its components
that demonstrated that it was an OHSMS (e.g., comprehensive occupational health
and safety program). In the case where terms were merely suggestive of an
OHSMS, a description of the OHSMS had to have been reported or referenced
and to have qualified as that of an OHSMS (i.e., two or more system elements
(Redinger and Levine, 1998) were specified, at least one of which was in the man-
agement domain rather than the activity/technical domain).
Multi-faceted management system interventions were included if they had an occupa-

tional health and/or safety component (e.g., a safety, health and environmental manage-
ment system) with primary prevention as the main focus. Management system
interventions focusing on disability or health services were excluded. Extra-workplace
or workplace initiatives designed to address isolated aspects of OHSMSs or particular
risks (e.g., needle-stick injuries in a health-care facility) were not included.
Type(s) of evidence. Studies had to examine either OHSMS implementation or the
effectiveness of OHSMS interventions.
Implementation studies were required to have a quantitative measure of change in the
level or intensity of the OHSMS. Effectiveness studies were required to have a quantitative
measure of one of the following outcomes: intermediate OHS outcomes (e.g., changes in
knowledge, beliefs, values, perceptions, behaviours, hazards, or risks); final OHS out-
comes (e.g., changes in injury/illness statistics or employee quality of life); or economic
outcomes (e.g., changes in the costs associated with employee illness/injury).
Studies were required to make a comparison of outcomes with the presence and absence
of an OHSMS intervention, or between OHSMS interventions of different intensities.
2.3. Quality appraisal (QA)
Studies meeting the eligibility criteria were assessed for methodological quality using a
process developed by the authors based on previous work (Abenhaim et al., 2000; Côté
et al., 2001; Drummond and Jefferson, 1996; Franche et al., 2005; Jadad, 1998; Kuhn et al.,
1999; Schulpher et al., 2000; Tompa et al., 2004; van Tulder et al., 2003; Zaza et al., 2000).
The method developed for this review emphasized parsimony with an aim to streamlining
the consensus procedure. The questions focused on internal validity.
The five quality appraisal questions were:
(i) Are you confident that the means of selecting and maintaining the sample minimized
bias?
(ii) Are you confident that the potential confounders were adequately considered, and
then either well controlled or appropriately discounted as a source of bias?
(iii) Are you confident that the measurement methods did not introduce bias to the
corresponding findings?
(iv) Were appropriate statistical tests applied to the data?
(v) Are you confident that there are no additional potential sources of bias in the esti-
mate of implementation/effectiveness not already captured in the previous questions?
Multiple choice response options were provided (e.g., yes, partially, no, unclear/
unknown from information provided).
Reviewers gave an overall summary rating of the quality of evidence provided by the
study. They selected one of the following four categories:
• very low (serious limitations),

• low (major limitations),
• moderate (moderate limitations), and
• high (no or minor limitations).
The methodological quality of the evidence of each study was rated independently by
two reviewers, who then met for consensus. If consensus could not be reached an addi-
tional reviewer was consulted. The quality assessment was carried out separately for
implementation, intermediate, final OHS and economic outcomes in each study.
2.4. Data extraction (DE)
Evidence rated during the quality appraisal step as moderate or high was extracted for
the evidence synthesis. Pairs of reviewers independently extracted data from the included
studies, using a standard form, and then met to reach consensus. Data were extracted on
the research question, intervention, study design, study population, results, and statistical
analyses.
2.5. Evidence synthesis
Many systematic reviewers choose an explicit algorithm at the outset of the study for
later translation of the findings into a summary statement about the level or strength of
evidence they provide. (Briss et al., 2000; Franche et al., 2005; GRADE Working Group,
2004; Kuhn et al., 1999; Tompa et al., 2004; van Tulder et al., 2003). Criteria for these
algorithms are customarily based on study design, quality of research, consistency of
the results, and number of studies.
This review did not adopt an explicit algorithm at the outset, since there is a lack of
consensus in the field as to which synthesis algorithm is best. In addition, it was thought
premature to base an algorithm upon a newly developed quality assessment tool. Instead,
a summary statement was synthesized in the style of a traditional narrative review, after
considering the same aspects of evidence as do the algorithms. Such an approach is per-
missible in best-evidence syntheses (Slavin, 1995) and has been used in other systematic
reviews in this field (see special issue of American Journal of Preventive Medicine, 2000).
3. Results
3.1. Publications identified through the literature search and quality assessment
Following a review of titles and abstracts, and initial screening of full papers where nec-
essary, 23 studies were identified which met the inclusion criteria. Their citations were
found through the following sources:
• bibliographic databases (18),

• citations in reviews (3),
• authors’ personal files (1),
• suggestion by external reviewer of interim report (1).
The 23 eligible studies were then assessed for methodological quality. Thirteen contained
evidence given an overall rating of ‘‘moderate’’ or ‘‘high’’ and proceeded to the next step in
the review, data extraction. The other ten studies, whose evidence received an overall rating
of ‘‘low’’ or ‘‘very low,’’ (Anonymous, 1993, 1994; Bolton and Kleinsteuber, 2001; Dotson,
1996; Eisner and Leger, 1988; Everley, 1997; Kjellén et al., 1997; Lanoie, 1992; Mikkelsen
and Saksvik, 2004; Nichols, 1990) were excluded from further review. (It should be noted
that this overall rating was determined from the point of view of this review’s question
about OHSMS effectiveness. The studies excluded at this stage were not necessarily low
quality studies from the point of view of their own research questions.)
An overview of the characteristics of the 13 studies proceeding to data extraction is
shown in Table 1. Seven were of voluntary OHSMS interventions, all of which took place
in English-speaking countries. Many of the voluntary interventions were unique to the
workplace(s) under study. Six studies were of mandatory OHSMS interventions, with four
dealing with the Norwegian Internal Control legislation introduced in 1992. The other two
were initiatives of Canadian provincial legislatures in the late 1970s.
While the studies of mandatory OHSMSs were of large numbers of workplaces, the
studies of voluntary OHSMSs were of small numbers of workplaces and often of just
one workplace. The most common effect studied was implementation but there were also
some findings on intermediate OHS outcomes, final OHS outcomes and economic
outcomes.
3.2. Quality of the literature investigating OHSMSs
The methodological quality of the OHSMS intervention literature was generally judged
to be weak. Only the study of implementation by LaMontagne et al. (2004) was assessed as
‘‘high’’ quality.
Table 1
Summary of the 13 studies providing the highest quality evidence
Authors (year of publication) OHSMS interventiona Country Industrial sector Number of organizations Type of data
in analysis
Voluntary interventions
Alsop and LeCouteur (1999) Organization’s own; based Australia Public administration 1 organization; Economic
on international and multiple sites
L.S. Robson et al. / Safety Science 45 (2007) 329–353

national standards
Bunn et al. (2001) Organization’s own United States Manufacturing 1 corporation; Implementation
multiple sites Final OHS
Economic
Edkins (1998) Organization’s own; Australia Transportation, air 1 airline; 2 sites Intermediate OHS
based on
other programs in industry
LaMontagne et al. (2004) Researcher United States Manufacturing 15 work sites Implementation
Pearse (2002) Organization’s Australia Manufacturing 16 companies Implementation
own; adapted
from AS/NZS 4804
Walker and Tait (2004) HSE-based Britain Mixed 24 enterprises Implementation
Yassi (1998) Organization’s own Canada Health care 1 hospital Economic
Mandatory interventions
Dufour et al. (1998) Quebec LSST (1979) Canada Manufacturing All firms in most Economic
manufacturing sectors
Lewchuk et al. (1996) Ontario Bill 70 (1979) Canada Manufacturing; retail 636 workplaces Final OHS
Saksvik and Nytro (1996) Internal Control (1992) Norway All 2092 enterprises Implementation
Intermediate OHS
Final OHS
Nytro et al. (1998) Internal Control (1992) Norway All 1182–2092 enterprises Implementation
Saksvik et al. (2003) Internal Control (1992) Norway All 1182–2092 enterprises Implementation
Torp et al. (2000) Internal Control (1992) Norway Motor vehicle repair services 267 garages Intermediate OHS
Final OHS
a
Abbreviations used in column: AS/NZS 4804, Australian/New Zealand Standard; HSE, Health and Safety Executive (British government).
339
The prevalence of methodological weakness was mainly a consequence of the experi-

mental designs typically employed. Only three of the 23 eligible studies (Edkins, 1998;
LaMontagne et al., 2004; Yassi, 1998) employed a comparison group of any kind; the
remaining studies used after-only, cross-sectional, before–after or time series designs.
The quality assessment process rated studies against four main criteria concerned with
the study’s internal validity: sampling, confounding, measurement, and statistics. The
most common weakness in the 23 eligible studies was poor evaluation of and control
for confounders, with 11 studies failing to meet the corresponding criterion at least par-
tially. The second most common weakness, found in six studies, concerned potential bias
introduced through measurement methods.
The difficulty in meeting the confounding criterion was inherent in the studies with no
comparison group. In some cases, the publications lacked explicit consideration of poten-
tial confounders. In single workplace studies, co-interventions were a major concern. For
example, some authors mentioned the existence of coincident interventions such as cost
control (Bunn et al., 2001) or restructuring (Kjellén et al., 1997). Bolton and Kleinsteuber
(2001) stated that ‘‘many other variables, including budgetary considerations and the
scope and nature of programmatic tasks and activities’’ could have contributed to the
injury rate trend they observed. For other single workplace studies, co-intervention was
a concern simply because the authors omitted any explicit consideration and discounting
of it. No doubt in some cases, there were no major, relevant changes, but the lack of any
report caused the study to fail on the criterion concerned with confounding.
Insufficient information was also a reason why some studies failed to satisfy the crite-
rion regarding measurement methods. In other studies, self-report methods were used to
measure OHSMS implementation without any consideration of the bias arising when
people have a vested interest in portraying compliance.
3.3. Findings on the implementation and effectiveness of voluntary OHSMSs
3.3.1. Description of the studies on the implementation and effectiveness of voluntary

OHSMSs
The review’s search for relevant literature resulted in 14 studies of voluntary OHSMSs,
seven of which remained after quality assessment (Alsop and LeCouteur, 1999; Bunn et al.,
2001; Edkins, 1998; LaMontagne et al., 2004; Pearse, 2002; Walker and Tait, 2004; Yassi,
1998). Only one study (Bunn et al., 2001) examined multiple outcomes (implementation,
final OHS outcomes, and economic outcomes). The other studies focused on only one –
implementation (LaMontagne et al., 2004; Pearse, 2002; Walker and Tait, 2004), interme-
diate OHS (Edkins, 1998), or economic (Alsop and LeCouteur, 1999; Yassi, 1998).
The studies involved a variety of interventions and samples. Four studies (Alsop and
LeCouteur, 1999; Bunn et al., 2001; Edkins, 1998; Yassi, 1998) reported on OHSMSs devel-
oped and implemented in a single organization – respectively, a municipal government, an
international manufacturer, a regional airline, and a hospital. The study by Alsop and
LeCouteur (1999) concerned an OHSMS developed in the presence of and integrated with
management systems for quality and the environment. The intervention reported by Bunn
et al. (2001) consisted of the creation of a Health, Safety and Productivity Department,
which integrated all functions concerned with employee health, safety and associated pro-
ductivity. Key features of the intervention were goal-setting, performance measurement
and senior management commitment. Edkins (1998) investigated a new system of hazard
identification, safety information management, and safety communication, overseen by a

new safety manager. The hospital intervention (Yassi, 1998) consisted of the systematic
identification, measurement, and control of OHS risks, as well as program evaluation.
The three other studies of voluntary OHSMSs involved interventions delivered to mul-
tiple organizations. Pearse (2002) described a community intervention with 20 small and
medium-sized metal fabrication companies. It consisted of the development and dissemi-
nation of OHSMS guidelines adapted from the Australian/New Zealand voluntary stan-
dard (AS/NZS 4804), group networking meetings and audits. LaMontagne et al.’s (2004)
research group targeted large manufacturers that were likely to use hazardous substances.
Fifteen worksites completed the intervention, which included a baseline OHS program
assessment and tailored consultation and educational sessions for managers and commit-
tees. Sessions emphasized upper management commitment, employee participation,
improvement in the resources directed towards hazard control, and OHS training for all
employees. Walker and Tait (2004) determined the impact of 90-min consultation sessions
delivered by two non-profit information centres upon the OHSMSs of small- and medium-
sized organizations.
3.3.2. Findings on the implementation of voluntary OHSMSs

The four studies of implementation all showed the sizeable development of OHSMSs,
as determined by audit methods. The Walker and Tait study (2004) showed marked self-
reported changes in terms of a policy statement being present and a risk assessment having
taken place. Forty-six percent of organizations added a policy statement and 79% added a
risk assessment as a result of the intervention. It must be noted that these organizations
had requested assistance with establishing an OHSMS, and that these two OHSMS ele-
ments were emphasized in the information provided to the organizations.
The other three implementation studies reported positive changes in the scores from a
quantitative audit. Bunn et al. (2001) reported an increase from 63% to 79% of available
audit points over a 2 years period for the organization studied. Pearse (2002) and
LaMontagne et al. (2004) included data in their publications that allowed effect sizes to
be calculated (0.48 and 0.68, respectively). These are considered to be medium and med-
ium–large, respectively (Cohen, 1977).
Only one of the four studies of implementation contained information relating these
changes in audit scores to changes in final OHS or economic outcomes (Bunn et al.,
2001). Thus, the ultimate practical significance of most of these reports remains unknown.
In the study by Bunn et al. (2001), described below, there were changes in final OHS and eco-
nomic outcome indicators, but it is difficult to attribute these solely to the changes in the
OHSMS because of the concurrent cost containment initiatives that were also underway.
3.3.3. Findings on the intermediate outcomes of voluntary OHSMSs

As for intermediate outcomes, Edkins (1998) reported greater positive changes in the
intervention group than in the comparison group, both for some self-reported measures
(e.g., safety culture index) and for some objective measures (confidential hazard reports,
organizational actions). Thirteen new organizational actions in safety were attributed to
the intervention group, while none were attributed to the control group. As with the imple-
mentation findings, the significance of these results for final outcomes remains unknown.
Furthermore, the reader is left wondering whether the large achievements in organiza-
tional actions were largely attributable to the personal qualities of the new safety manager
rather than the entirety of the new OHSMS. Indeed, the reader is told that many of the
actions addressed long standing issues.
3.3.4. Findings on the final outcomes of voluntary OHSMSs

Bunn et al. (2001) was the only study of voluntary OHSMSs to report on final OHS
outcomes. They found a 24% decrease in illness/injury frequency and a 34% decrease in
lost-time case rate over 3 years. It should be noted that management became accountable
to the board of directors for improvement in these safety indicators, coincident with the
intervention. One effect of this would have been to ensure critical management commit-
ment for the OHSMS intervention. It could also have led to a shift in the reporting prac-
tices in the organization, such that incidents were less likely to be documented. The
authors refer to ‘‘aggressive return to work programs,’’ which could have contributed to
the reduction in lost-time case rate.
Workers’ compensation costs also decreased markedly in studies of voluntary
OHSMSs. Bunn et al. (2001) found a 13% decrease in workers’ compensation cost per
employee. Yassi (1998) and Alsop and LeCouteur (1999) found decreases of 25% and
52% in premium rates, respectively. However, in each of these studies, there were initia-
tives beyond primary prevention which likely contributed to these results. Yassi (1998)
mentioned a return to modified work program. Alsop and LeCouteur (1999) referred to
‘‘targeted and sustained effort in claims and injury management’’ during the 2 years prior
to the OHS initiative. Bunn et al. (2001) listed various cost-cutting measures including
better case management and aggressive return to work programs.
3.4. Findings on the implementation and effectiveness of mandatory OHSMSs
3.4.1. Description of the studies on the implementation and effectiveness of mandatory

OHSMSs
Nine studies of mandatory OHSMS interventions were identified by the review. Six
remained after the quality assessment screen. Four of these were based on the Norwegian
regulations on Internal Control (IC) of health, safety, and the environment (Nytro et al.,
1998; Saksvik and Nytro, 1996; Saksvik et al., 2003; Torp et al., 2000). Two studies
focused on regulations in the Canadian provinces of Quebec and Ontario, respectively
(Dufour et al., 1998; Lewchuk et al., 1996).
The Canadian regulations were progressive for their time (the late 1970s), but in terms
of contemporary approaches to integrated health and safety management are more limited
and less system-oriented in their requirements. The Norwegian IC regulations, on the
other hand, incorporated systematic management concepts, which were increasingly being
found in best practice models in business at that time.
3.4.2. Findings on the implementation and effectiveness of the IC legislation in Norway

The IC studies demonstrated most clearly the effects of the legislation on the use of
OHSMSs. There was an increase from 8% to 47% of workplaces fully implementing the
IC requirements over the period 1 year to 7 years post-intervention (Nytro et al., 1998;
Saksvik and Nytro, 1996; Saksvik et al., 2003). Consistent with these findings, Saksvik
and Nytro (1996) found that, in 1993, a sizeable portion of workplaces credited the legis-
lation with improvements in various aspects of an OHSMS (e.g., 25% attributed clearer
lines of responsibility). Researchers also found an impact on awareness of health, safety
and the environment, an intermediate OHS outcome, with 39% of workplaces attributing
an increase to the legislation.
The attempts by Saksvik and Nytro (1996) and Torp et al. (2000) to find any impact of
the legislation on other intermediate OHS outcomes and on final OHS outcomes were less
convincing. Saksvik and Nytro (1996) carried out multi-variable regression analyses to
relate the degree of self-reported IC implementation after 1 year to self-reported changes
in absenteeism and injury rates. The model for absenteeism was statistically significant, but
explained only 5% of the observed variance. The model for injury rates was not statisti-
cally significant. These relatively weak results might have been partly due to the short
observation period and the crude measure of change used. Respondents were asked to
report on whether rates were lower/stable/higher in 1991 and the first half of 1992, com-
pared with 1990. (The legislation came into effect on January 1, 1992.)
Torp et al. (2000) took a similar analytic approach to survey data gathered 4 years post-
implementation. They found that the majority of intermediate variables tested (e.g., satis-
faction with HES activities in the garage, perceived physical working environment), had a
statistically significant relationship with IC status. For the two final OHS outcomes tested,
musculoskeletal symptoms and sick leave, the relationship was in the expected direction
but weak in both cases (standardized beta coefficients of 0.026 and 0.048, respectively),
and statistically significant in only the first.
The evidence provided by Saksvik and Nytro (1996) and Torp et al. (2000) on the
effects of IC implementation is weak because of the cross-sectional designs. Whether
the IC implementation preceded the outcomes remains unknown and unmeasured factors
related to both the degree of IC implementation and final OHS outcomes (e.g.,
management commitment to OHS) are alternative conceivable explanations for the
findings.
3.4.3. Findings on the effectiveness of simple OHSMS legislation in Canada

Lewchuk et al. (1996) examined the effect of the 1979 introduction of Bill 70 legislation
in Ontario on lost-time injury claim rates. They analyzed claims data from 1976 to 1989 in
six manufacturing and two retail sectors. These two sectors were differently affected by the
legislation, with the requirements for manufacturing being more intensive. The reviewers
therefore categorized the research design as two distinct time series.
Results from regression analyses of the manufacturing sector data (controlled for
employment, union status, time and sector) showed that Bill 70 had a significant effect,
equivalent to an 18% decrease in lost-time injury rate. A separate analysis showed that
the effects were progressively larger for 1980, 1981, and 1982+, suggesting a phasing in
of implementation and perhaps a lag in effects. In contrast, in the retail sector, the effect
of the legislation was small and statistically not significant.
An explanation for the differences observed between sectors might be a difference in the
OHSMS requirements, especially the requirement for a joint-health-and-safety committee
in only the manufacturing sector. Conflicting with this interpretation, however, is the
observation from separate regression analyses that joint-health-and-safety committees
formed in 1980 or later appeared to have no effect on injury rates – and yet Bill 70, intro-
duced in 1979, appeared to have an effect. Thus, other unmeasured, industry-specific con-
textual factors or co-interventions might have played a role instead. For example, it seems
likely that the retail sector would have received less attention from the Ministry of
Labour’s inspectorate because of retail’s lower level of risk.
In the Quebec study (Dufour et al., 1998), pooled time series regression analyses based on
the 3 years of annual data across the 19 manufacturing sectors were used to develop six dif-
ferent specifications of the authors’ theoretical equation explaining variation in total pro-
ductivity growth. All of the specifications had good explanatory power (R-squared values
from 0.54 to 0.77) and regression coefficients were relatively consistent. The regression coef-
ficient for the variable ‘‘percentage of companies having prevention programs’’ (the OHSMS
variable) was positive and statistically significant. The authors interpreted this to mean that
the prevention programs reduced injuries, which in turn enhanced firm productivity.
However, there are plausible alternative explanations for the findings. Although numer-
ous potential confounding variables were included in the model, there remained the pos-
sibility of a common underlying factor (e.g., management competency) that could be
associated at the aggregate level with both (a) the more rapid development of a prevention
program in response to a legislative change and (b) productivity growth. Additionally, the
data on the prevalence of prevention programs in the study relied on the report of the pre-
vention program by the workplace to the public authorities; its accuracy was unknown.
3.5. Evidence synthesis
Sections 3.3 and 3.4 described the findings of the individual studies of voluntary and
mandatory OHSMSs. This section gives a higher level synthesis of each group of studies
and assesses the body of evidence for OHSMS effectiveness.
The heterogeneity of the studies precluded a quantitative pooling (i.e., meta-analysis),
especially because they are so few in number (van Tulder et al., 2003). The evidence was
therefore synthesized qualitatively. The following discussion focuses separately on volun-
tary and mandatory OHSMSs and assesses the evidence available for each.
3.5.1. Evidence for the effectiveness of voluntary OHSMS interventions

The seven studies on voluntary OHSMS interventions all showed positive effects (see
Table 2). One study also had some null findings; there were no negative findings. The posi-
tive effects included:
• increased OHSMS implementation over time;

• intermediate effects (e.g., better safety climate, increased hazard reporting by employ-
ees, more organizational action taken on OHS issues);
• decreases in injury rates; and
• decreases in disability-related costs (e.g., workers’ compensation costs, short- and long-
term disability costs).
The falls in injury rates and workers’ compensation costs would be of practical impor-
tance to stakeholders (declines of 24–34% in injury rates and of 13–52% in workers’ com-
pensation costs).
However, six of the seven studies had moderate methodological limitations. The one
high-quality study (LaMontagne et al., 2004) provided evidence for only implementation.
It is unknown how the implementation effects observed in this particular study would have
affected injury rates and economic costs.
Overall, the evidence provided by these seven studies was sparse. Only four studies had
findings for any of intermediate OHS, final OHS, or economic outcomes. Each involved a
Table 2
Summary of the evidence
Authors (year) Method. quality Study design Number of workplaces Effects of the interventiona
in analysis
L.S. Robson et al. / Safety Science 45 (2007) 329–353

Implementation Intermediate Final OHS Economic
Voluntary interventions
Alsop and LeCouteur (1999) Moderate Time series 1 +
Bunn et al. (2001) Moderate Time series 1 + + +
Edkins (1998) Moderate Non-randomized trial 2 +
LaMontagne et al. (2004) High Randomized trial 15 +/0
Pearse (2002) Moderate Before-after 16 +
Walker and Tait (2004) Moderate After-only 24 +
Yassi (1998) Moderate Time series with 1 +
comparison group
Mandatory interventions
Dufour et al. (1998) Moderate Time series, 19b +
aggregated data
Lewchuk et al. (1996) Moderate Time series 636 +, 0
Saksvik and Nytro (1996) Moderate Cross-sectional 2092 + + +, 0
Nytro et al. (1998) Moderate Before-after 1182–2092 +
Saksvik et al. (2003) Moderate Time series 1182–2092 +
Torp et al. (2000) Moderate Cross-sectional 267 + +, 0
a
‘‘+’’ indicates that the OHSMS had a beneficial effect; statistical significance was not necessarily determined; ‘‘ ’’ indicates an effect in the opposite direction,
statistical significance was not necessarily determined; ‘‘0’’ indicates that a statistical test was conducted and the results were not significant (n.s.; p > 0.05).
b
For the Dufour study, the number of sectors is shown, since sector was the unit of analysis.
345
single enterprise, making the direct applicability of the results to other workplaces
uncertain.
The single enterprise studies must also be regarded cautiously from the point of report-
ing bias (one favouring positive results). Researchers are known to be reticent to publish
findings that demonstrate an intervention has no significant effect. Workplace representa-
tives who champion interventions (and thus have a vested interest in their success) are
likely to have an even greater tendency not to publish such findings. All four reports on
single workplace interventions appeared to have been authored by workplace champions.
The intervention failure rate seen in this small sample of studies (0%) is markedly different
than the rate of 67–93% reported for quality management systems (Gardner, 2000). There
is no reason to expect the failure rate of OHSMSs to be markedly different than that for
quality, since many of the barriers to implementation can be presumed to be the same
(e.g., management commitment, culture change).
Generalizability is also an issue in the three multiple workplace studies, all of which
focused on implementation. The interventions in both Pearse (2002) and LaMontagne
et al. (2004) were delivered by researchers and showed high refusal rates in the recruitment
phase of 67% (Pearse, 2001) and 59%, respectively. The effects seen in the sample of firms
studied are likely to be a biased estimate of the effectiveness that would be seen if the inter-
vention had been delivered to all eligible firms. Firms willing to participate in the study
would also have been likely to have had a higher degree of management commitment to
OHS, which would also be predictive of successful OHSMS implementation. A similar
consideration is at play for the Walker and Tait (2004) study. They reported on an inter-
vention in the ‘‘real world’’, but all of the workplaces in their sample had invited the inter-
vention by seeking help from information centres.
In sum, there is insufficient evidence in the published, peer-reviewed literature on the
effectiveness of voluntary OHSMSs to make recommendations either in favour of or
against them.
3.5.2. Evidence for the effectiveness of mandatory OHSMS interventions

The studies on mandatory OHSMS interventions also indicated consistently positive
effects, although there were also null findings on some of the final OHS outcomes (see
Table 2). The studies suggest that mandatory interventions result in
• increased OHSMS implementation over time;

• intermediate effects (e.g., increased HES awareness; improved employee perceptions of
the physical working environment and the psychosocial environment; and increased
workers’ participation in HES activities);
• decreases in lost-time injury rates; and
• increases in workplace productivity.
The size of the observed changes in OHSMS development (an increase from 8% to 47%
of all workplaces with full implementation of OHSMS regulations) and the decline in
injury rate (18%) observed are likely to be of practical importance to stakeholders.
On the other hand, all studies in this group had moderate methodological limitations.
These limitations arose largely from the simple study designs employed, especially the
two cross-sectional studies measuring final OHS outcomes where the direction of causality
was uncertain (Saksvik and Nytro, 1996; Torp et al., 2000). In addition, it was hard
to exclude the possibility that there had been confounding or co-intervention in some stud-
ies. The most convincing finding was that of Saksvik et al. (2003), demonstrating the
implementation of IC regulations in Norway (an increase from 8% to 47% of workplaces
with full implementation). Unfortunately, there is a lack of credible evidence about how
the Norwegian achievements in compliance affected injury rates and other outcomes.
As well as internal validity weaknesses, two of the six studies were concerned with Que-
bec and Ontario legislation from the late 1970s and are therefore not applicable to devel-
oped countries now.
In sum, there is insufficient evidence in the published, peer-reviewed literature on the
effectiveness of mandatory OHSMSs to make recommendations either in favour of or
against them.
4. Discussion
4.1. Identifying and addressing research gaps
The review identified a number of gaps in the research. The most important was the
lack of research whose explicit purpose was to study the effectiveness of voluntary and
mandatory OHSMS interventions on employee health, safety and economic outcomes.
Moreover, the studies were seldom sufficiently rigorous methodologically to allow for
great confidence in the reported findings. Their limitations also inhibit the ability to apply
the results to other workplaces.
The following were common limitations in the studies:
• simple research designs (e.g., lack of comparison group, use of cross-sectional designs),
• lack of consideration or control of confounding (through design or statistical
adjustments),
• lack of information reported about measurement methods and potential biases in
measurement,
• samples which prevented generalization of findings (i.e., several single workplace
studies).
A similar limitation has been found in the quantity and quality of the evidence on the
closely related topic of safety management system effectiveness (Hale, 2003).
The scarcity of high-quality published research on the implementation/effectiveness of
OHSMSs may relate to the difficulties in carrying out such research. Recruitment of work-
places to both intervention and controls is a challenge. Many workplaces are not willing to
make a commitment to a large intervention like an OHSMS, let alone its evaluation.
Among those that are, some might not want to risk being allocated to the comparison
group. Thus, large refusal rates and some withdrawal after allocation are to be expected.
Measurement presents conceptual, logistic and resource challenges because of the com-
plexity of OHSMSs and their environment.
When conducting a controlled trial is not feasible, observational cohort studies are usu-
ally considered to be the next best option by those in the epidemiological field. To answer
questions about OHSMSs using an epidemiological model, one would need to follow a
large sample of workplaces (the cohort) over time, measuring the introduction or upgrad-
ing of OHSMSs and then measuring outcomes of interest at the workplace level and
possibly at the worker-level too. However, such research designs are very expensive and
complex to implement, and as such, are rarely used. In Canada, such a research design
is employed by the national survey organization Statistics Canada (2005) to examine issues
such as organizational productivity, but thus far OHS is not within its scope.
4.2. Strengths of the review
The volume of studies published each year is more than most practitioners or research-
ers can easily keep track of or synthesize. This review has clearly eliminated a huge volume
of work in finding the relatively few studies of interest and to summarize their findings. Its
use of explicit, systematic methods helps ensure that this summary is relatively objective in
its appraisal.
Within the parameters set by the review question and the included sources, the review
team feels confident that the search has been comprehensive and that it is unlikely that
there are other items in the peer-reviewed, published literature that would dramatically
alter the conclusions of the review. The review drew from a broad range of academic dis-
ciplines. The eight databases used in the search represent the fields of occupational med-
icine, occupational safety, risk management, management, occupational psychology, and
sociology.
The review’s extensive search of the current literature confirmed that no other system-
atic review has considered the effectiveness of OHSMSs. Until now there have only been
high-quality narrative reviews available (see Section 1). The present review therefore
makes a unique contribution to the research literature.
4.3. Limitations of the review
Resource and feasibility constraints limited consideration of the evidence to the pub-
lished, peer-reviewed literature identified in eight databases. The usual expectation is that
the literature of highest quality is in peer-reviewed journals. Only a preliminary search and
screen of other literature, i.e., that which is not peer-reviewed and published, was
undertaken.
A simple search of thesis dissertations (through the Dissertations Abstract Interna-
tional database) revealed some sources which were eligible (Ford, 1998; Weems, 1998).
However, a cursory review of their quality suggested that they would not alter the conclu-
sions of this review in any substantial way and might not in fact pass the quality assess-
ment screen.
Other types of ‘‘grey literature’’ were omitted from the review too (e.g., government
reports, conference proceedings, unpublished reports by OHSMS vendors and users). This
research team identified two reports of this type (Hanson et al., 1998; European Agency
for Safety and Health at Work, 2002) during the course of the review. Accessing the grey
literature on OHSMSs in a systematic way would be a great challenge, given the variety of
potential sources and the proprietary nature of some of the knowledge.
Although a large set of search terms related to OHSMSs was used to search the biblio-
graphic databases, it remains possible that it failed to capture some relevant studies. For
example, an intervention described as a ‘‘climate intervention’’ might not have been
detected if other terms like ‘‘safety management system,’’ ‘‘safety program,’’ ‘‘safety
system’’, ‘‘occupational safety standard’ were not used too (see Section 2.1). Such
omissions are not likely to be many. Not only were the results of the database searches
used as a source of potentially relevant citations, but also the following sources: the 26 eli-
gible studies; recent reviews in journals and monographs, including Gallagher et al. (2003),
Frick et al. (2000), Hale (2003) and Kogi (2002); personal files; and reference lists
requested from experts in the field.
There were a few specific limits placed on the scope of the review. First, only studies
that examined a clearly identifiable voluntary or mandatory OHSMS intervention were
included. This excluded an ambitious U.S. study that compared state voluntary initiatives,
including prevention programs, while controlling for several other variables (Smitha et al.,
2001) and another interesting paper on performance indicators (Simpson and Gardner,
2001). In both cases, the potential OHSMS variable was not clearly defined. Also excluded
by this criterion were a few cross-sectional studies investigating the relationship between a
researcher-defined measure of OHSMSs and injury outcomes (e.g., Mearns et al., 2003;
Simard and Marchand, 1994), because there had been no intervention.
The review required the study to involve intervention on at least two OHSMS elements,
since the review’s focus was on systems. This excluded a substantial number of cross-sec-
tional studies of the effectiveness of single OHSMS elements such as joint-health-and-
safety committees (e.g., Cohen, 1977; Habeck et al., 1998; LaMontagne et al., 1996; Reilly
et al., 1995; Shannon et al., 1996).
By requiring studies to be quantitative in nature and published as a journal articles,
publications that were qualitative or theoretical in nature or published in a format outside
of a journal article were excluded. The case study of an Esso gas explosion (Hopkins,
2000), based on a Royal Commission on the event, included critical comments about
the OHSMS in use. The Frick et al. (2000) monograph contains further critique. Such
materials were not included in this review, although some of the issues raised were sum-
marized in Section 1.
5. Conclusions
This systematic review found a relatively small quantity of published, peer-reviewed evi-
dence involving OHSMS interventions, despite the fact that reviewers screened 4837 stud-
ies drawn from eight databases representing diverse disciplines. A qualitative synthesis of
the available research was used, because of the small number of studies and the heteroge-
neity of the studies’ characteristics.
The review’s synthesis of the evidence showed mostly favourable results. There were a
few null findings, but no findings of negative effects.
However, all but one of the studies included in the final synthesis had moderate meth-
odological limitations. The studies were seldom sufficiently rigorous to allow great confi-
dence in the reported findings. In addition, many of the findings concerned only OHSMS
implementation, with no accompanying results for outcomes. Finally, the characteristics
of workplace samples (e.g., single workplaces, large refusal rates) also prevent certainty
about the applicability of the findings to other workplaces.
In conclusion, despite the generally positive results on the effectiveness of OHSMS
interventions in the published, peer-reviewed literature, the evidence is insufficient to make
recommendations either in favour of or against particular OHSMSs. This is not to judge
these systems as ineffective or undesirable; it is merely to say that it would be incautious to
judge either way in the present state of our research knowledge.
Acknowledgements
The Institute for Work and Health (IWH) is an independent, not-for-profit research
organization. It receives core funding from the Ontario Workplace Safety and Insurance
Board. The authors benefited from the assistance of others at IWH: methodological assis-
tance from Jill Hayden, Selahadin Ibrahim and Victoria Pennick; library assistance from
Doreen Day and Dan Shannon; and administrative assistance from Krista Nolan. Several
external scientists gave generously of their time in reviewing an earlier report on this pro-
ject: Kaj Frick, Clare Gallagher, Wayne Lewchuk, Michael Quinlan, Jos Verbeek, and
Stefan Zimmer. Finally, other external, local stakeholders helped orient LR to the subject
matter from their various viewpoints: Andy King, Ron Lovelock, John MacNamara, Ed
McCloskey, Ken Taylor.
References
Abenhaim, L., Rossignol, M., Valat, J.P., Nordin, M., Avouac, B., Blotman, F., Charlot, J., Dreiser, R.L.,
Legrand, E., Rozenberg, S., Vautravers, P., for the Paris Task Force, 2000. The role of activity in the
therapeutic management of back pain. Report of the International Paris Task Force on Back Pain. Spine 25
(Suppl. 4), 1S–33S.
Anonymous, 1993. Best practice case study: Portland aluminium. Journal of Occupational Health & Safety –
Australia & New Zealand 9, 375–378.
Anonymous, 1994. Best practice case study: Herbert Adams Bakeries. Journal of Occupational Health & Safety –
Australia & New Zealand 10, 275–278.
Anonymous, 2000. American Journal of Preventive Medicine 18 (4S).
Alsop, P., LeCouteur, M., 1999. Measurable success from implementing an integrated OHS management system at
Manningham City Council. Journal of Occupational Health & Safety – Australia & New Zealand 15, 565–572.
Bennett, D., 2002. Health and safety management systems: liability or asset? Journal of Public Health Policy 23,
153–171.
Bolton, F.N., Kleinsteuber, J.F., 2001. A perspective on the effectiveness of risk assessment by first-line workers
and supervisors in a safety management system. Human and Ecological Risk Assessment 7, 1777–1786.
Briss, P.A., Zaza, S., Pappaioanou, M., Fielding, J., Wright-De Agéero, L., Truman, B.I., Hopkins, D.P., Mullen,
P.D., Thompson, R.S., Woolf, S.J., Carande-Kulis, V.G., Anderson, L., Hinman, A.R., McQueen, D.V.,
Teutsch, S.M., Harris, J.R., the Task Force on Community Preventive Services, 2000. Developing an
evidence-based Guide to Community Preventive Services – methods. American Journal of Preventive
Medicine 18 (Suppl. 1), 35–43.
British Standards Institution, 1996. BS 8800: guide to occupational health and safety management systems.
British Standards Institution, London.
British Standards Institution, 1999. OHSAS 18001: occupational health and safety management systems –
specification. British Standards Institution, London.
Bunn III, W.B., Pikelny, D.B., Slavin, T.J., Paralkar, S., 2001. Health, safety, and productivity in a
manufacturing environment. Journal of Occupational and Environmental Medicine 43, 47–55.
Chemical Industries Association, 1995. Responsible care management systems for health, safety and
environment. Chemical Industries Association, London.
Cohen, A., 1977. Factors in successful occupational safety programs. Journal of Safety Research 9, 168–178.
Côté, P., Cassidy, J.D., Carroll, L., Frank, J.W., Bombardier, C., 2001. A systematic review of the prognosis of
acute whiplash and a new conceptual framework to synthesize the literature. Spine 26, E445–E458.
Dalrymple, H., Redinger, C., Dyjack, D., Levine, S., Mansdorf, Z., 1998. Occupational health and safety
management systems: review and analysis of international, national, and regional systems and proposals for a
new international document. International Labour Organization, Geneva.
Dotson, K., 1996. An international safety and health measurement strategy: corporate programs, systems and
results. Journal of Occupational Health & Safety – Australia & New Zealand 12, 669–678.
Drummond, M., Jefferson, T., 1996. Guidelines for authors and peer reviewers of economic submissions to the
British Medical Journal. British Medical Journal 313, 275–283.
Dufour, C., Lanoie, P., Patry, M., 1998. Regulation and productivity. Journal of Productivity Analysis 9, 233–247.
Edkins, G.D., 1998. The INDICATE safety program: evaluation of a method to proactively improve airline
safety performance. Safety Science 30, 275–295.
Eisner, H.S., Leger, J.P., 1988. The international safety rating system in South African mining. Journal of
Occupational Accidents 10, 141–160.
European Agency for Safety and Health at Work, 2002. The use of occupational safety and health management
systems in the member states of the European Union: experiences at company level. European Agency for
Safety and Health at Work, Luxembourg.
Everley, M., 1997. Setting the standard. Health and Safety at Work 19, 10–12.
Figuera, J., López, L., 2003. Final report: review of the safety management and reporting practices. SAMRAIL
Consortium. Available from: <http://samnet.inrets.fr>.
Ford, A.C., 1998. Statistical analysis of the effectiveness of health and safety programs [dissertation]. University
of Louisville, Kentucky.
Franche, R.-L., Cullen, K., Clarke, J., Irvin, E., Sinclair, S., Frank, J., and the Institute for Work & Health
Workplace-Based RTW Intervention Literature Review Research Team, 2005. Workplace-based return-to-
work interventions: a systematic review of the quantitative literature. Journal of Occupational Rehabilitation
15, 607–631.
Frick, K., Wren, J., 2000. Reviewing occupational health and safety management: multiple roots, diverse
perspectives and ambiguous outcomes. In: Frick, K., Jensen, P.L., Quinlan, M., Wilthagen, T. (Eds.),
Systematic Occupational Health and Safety Management: Perspectives on an International Development.
Pergamon, Amsterdam, pp. 17–42.
Frick, K., Jensen, P.L., Quinlan, M., Wilthagen, T. (Eds.), 2000. Systematic Occupational Health and Safety
Management: Perspectives on an International Development. Pergamon, Amsterdam.
Gallagher, C., Underhill, E., Rimmer, M., 2003. Occupational safety and health management systems in
Australia: barriers to success. Policy and Practice in Health and Safety 1, 67–81.
Gardner, D., 2000. Barriers to the implementation of management systems: lessons from the past. Quality
Assurance 8, 3–10.
GRADE Working Group, 2004. Grading quality of evidence and strength of recommendations. British Medical
Journal 328, 1490–1497.
Grayham, D.A., del Rosario, V.O., 1997. The management of health and safety at work regulations 1992. Journal
of the Royal Society of Health 117, 47–51.
Habeck, R.V., Hunt, H., van Tol, B., 1998. Workplace factors associated with preventing and managing work
disability. Rehabilitation Counselling Bulletin 42, 98–143.
Hale, A.R., 2003. Safety management in production. Human Factors and Ergonomics in Manufacturing 13, 185–201.
Hanson, M.A., Tesh, K.M., Groat, S.K., Donnan, P.T., Ritchie, P.J., Lancaster, R.J., 1998. Evaluation of the
six-pack regulations 1992. Report No. 177/1998. HSE Books, Sudbury, UK.
Health and Safety Executive, 1997. Successful health and safety management: HSG65. Health and Safety
Executive, Sudbury, UK.
Hopkins, A., 2000. Lessons from Longford: the Esso Gas plant explosion. CCH Australia, Sydney.
International Atomic Energy Agency (IAEA), 2005. The management system for facilities and activities: draft
safety requirements, DS338 Draft 10. Available from: <http://www-ns.iaea.org/downloads/standards/drafts/
ds338.pdf>.
International Labour Office, 2001. Guidelines on Occupational Safety and Health Management Systems.
MEOSH/2001/2(Rev.). International Labour Office, Geneva.
Jadad, A., 1998. Randomised Controlled Trials: a User’s Guide. BMJ Books, London.
Kjellén, U., Boe, K., Hagen, L., 1997. Economic effects of implementing internal control of health, safety and
environment: a retrospective case study of an aluminium plant. Safety Science 27, 99–114.
Kogi, K., 2002. Work improvement and occupational safety and health management systems: common features
and research needs. Industrial Health 40, 121–133.
Kuhn, M., Doucet, C., Edwards, N., Public Health Research, E.a.D.P.P., Ontario Ministry of Health, P.H.B.,
1999. Effectiveness of coalitions in heart health promotion, tobacco use reduction, and injury prevention: a
systematic review of the literature 1990–1998. Public Health Research, Education and Development Program,
Dundas (ON).
LaMontagne, A.D., Rudd, R.E., Mangione, T.W., Kelsey, K.T., 1996. Determinants of the provision of ethylene
oxide medical surveillance in Massachusetts hospitals. Journal of Occupational and Environmental Medicine
38, 155–168.
LaMontagne, A.D., Barbeau, E., Youngstrom, R.A., Lewiton, M., Stoddard, A.M., McLellan, D., Wallace, L.M.,
Sorensen, G., 2004. Assessing and intervening on OSH programmes: effectiveness evaluation of the Wellworks-
2 intervention in 15 manufacturing worksites. Occupational and Environmental Medicine 61, 651–660.
Lanoie, P., 1992. Safety regulation and the risk of workplace accidents in Quebec. Southern Economic Journal
58, 950–965.
Lewchuk, W., Robb, A.L., Walters, V., 1996. The effectiveness of Bill 70 and joint health and safety committees
in reducing injuries in the workplace: the case of Ontario. Canadian Public Policy 22, 225–243.
Lund, H.L., 2004. Strategies for sustainable business and the handling of workers’ interests: integrated
management systems and worker participation. Economic and Industrial Democracy 25, 41–74.
Mearns, K., Whitaker, S.M., Flin, R., 2003. Safety climate, safety management practice and safety performance
in offshore environments. Safety Science 41, 641–680.
Mikkelsen, A., Saksvik, P., 2004. The relationship between systematic OHS management and sick leave. Journal
of Occupational Health & Safety – Australia & New Zealand 20, 169–179.
Nichols, T., 1990. Industrial safety in Britain and the 1974 Health and Safety at Work Act: the case of
manufacturing. International Journal of the Sociology of Law 18, 317–342.
Nichols, T., Tucker, E., 2000. OHS management systems in the UK and Ontario, Canada: a political economy
perspective. In: Frick, K., Jensen, P.L., Quinlan, M., Wilthagen, T. (Eds.), Systematic Occupational Health
and Safety Management: Perspectives on an International Development. Pergamon, Amsterdam, pp. 285–
310.
Nielsen, K., 2000. Organization theories implicit in various approaches to OHS management. In: Frick, K.,
Jensen, P.L., Quinlan, M., Wilthagen, T. (Eds.), Systematic Occupational Health and Safety Management:
perspectives on an International Development. Pergamon, Amsterdam, pp. 99–124.
Nytro, K., Saksvik, P., Torvatn, H., 1998. Organizational prerequisites for the implementation of systematic
health, environment and safety work in enterprises. Safety Science 30, 297–307.
Pearse W., 2001. Implementing OHS management systems in small to medium fabricated metal product
companies. In: Proceedings of the First National Conference on Occupational Health and Safety
Management Systems, Crown Content, Melbourne, pp. 83–100.
Pearse, W., 2002. Club zero: implementing OHSMS in small to medium fabricated metal product companies.
Journal of Occupational Health & Safety – Australia & New Zealand 18, 347–356.
Quinlan, M., Mayhew, C., 2000. Precarious employment, work-re-organisation and the fracturing of OHS
management. In: Frick, K., Jensen, P.L., Quinlan, M., Wilthagen, T. (Eds.), Systematic Occupational Health
and Safety Management: Perspectives on an International Development. Pergamon, Amsterdam, pp. 175–
198.
Redinger, C.F., Levine, S.P., 1998. Development and evaluation of the Michigan occupational health and safety
management system assessment instrument: A universal OHSMS performance measurement tool. American
Industrial Hygiene Association Journal 59, 572–581.
Reilly, B., Paci, P., Holl, P., 1995. Unions, safety committees and workplace injuries. British Journal of Industrial
Relations 33, 275–288.
Robson, L., Clarke, J., Cullen, K., Bielecky, A., Severin, C., Bigelow, P., Irvin, E., Culyer, A., Mahood, Q., 2005.
The effectiveness of occupational health and safety management systems: a systematic review. Institute for
Work & Health, Toronto, Canada.
Saksvik, P.O., Nytro, K., 1996. Implementation of internal control (IC) of health, environment and safety (HES)
in Norwegian enterprises. Safety Science 23, 53–61.
Saksvik, P.O., Nytro, K., Torvatn, H., 2003. Systematic occupational health and safety work in Norway: a
decade of implementation. Safety Science 41, 721–738.
Saksvik, P.O., Quinlan, M., 2003. Regulating systematic occupational health and safety management: comparing
the Norwegian and Australian experience. Relations Industrielles/Industrial Relations 58, 33–59.
SAMRAIL, 2004. Guidelines for the safety management system. SAMRAIL Consortium. Available from:
<http://samnet.inrets.fr>.
Schulpher, M., Fenwick, E., Claxton, K., 2000. Assessing quality in decision analytic cost-effectiveness models: a
suggested framework and an example of application. Pharmacoeconomics 17, 461–477.
Shannon, H.S., Walters, V., Lewchuck, W., Richardson, J., Moran, L.A., Haines, T., Verma, D., 1996.
Workplace organizational correlates of lost-time accident rates in manufacturing. American Journal of
Industrial Medicine 29, 258–268.
Simard, M., Marchand, A., 1994. The behaviour of first-line supervisors in accident prevention and effectiveness
in occupational safety. Safety Science 17, 169–185.
Simpson, I., Gardner, D., 2001. Using OHS positive performance indicators to monitor corporate OHS
strategies. Journal of Occupational Health & Safety – Australia & New Zealand 17, 125–134.
Slavin, R.E., 1995. Best evidence synthesis: an intelligent alternative to meta-analysis. Journal of Clinical
Epidemiology 48, 9–18.
Smitha, M.W., Kirk, K.A., Oestenstad, K.R., Brown, K.C., Lee, S.D., 2001. Effect of state workplace safety laws
on occupational injury rates. Journal of Occupational and Environmental Medicine 43, 1001–1010.
Standards Australia and Standards New Zealand, 1997. AS/NZS 4804: occupational health and safety
management systems – general guidelines on principles, systems and supporting techniques. Standards
Australia; Standards New Zealand, Homebush (NSW); Wellington.
Tompa, E., Trevithick, S., McLeod, C., 2004. A systematic review of the prevention incentives of insurance and
regulatory mechanisms for occupational health and safety. Working paper #213. Institute for Work and
Health, Toronto.
Torp, S., Riise, T., Moen, B.E., 2000. Systematic health, environment and safety activities: do they influence
occupational environment, behaviour and health? Occupational Medicine (Oxford) 50, 326–333.
Tortorella, M.J., 1995. The three careers of W. Edwards Deming. Siam News. Available from: <http://
www.deming.org/theman/articles/articles_threecareers01.html>.
van Tulder, M., Furlan, A., Bombardier, C., Bouter, L., Editorial Board of the Cochrane Collaboration Back
Review Group, 2003. Updated method guidelines for systematic reviews in the Cochrane Collaboration Back
Review Group. Spine 28, 1290–1299.
Walker, D., Tait, R., 2004. Health and safety management in small enterprises: an effective low cost approach.
Safety Science 42, 69–83.
Walters, D. (Ed.), 2002. Regulating health and safety management in the European Union: a study of the
dynamics of change. P.I.E.-Peter Lang, Bruxelles.
Weems, L.D., 1998. The Effect of Participation in the Occupational Safety and Health Administration’s
Voluntary Protection Programs on Injury/Illness Rates [dissertation]. Louisiana State University, Louisiana.
Wokutch, R.E., VanSandt, C.V., 2000. OHS management in the United States and Japan: the DuPont and the
Toyota models. In: Frick, K., Jensen, P.L., Quinlan, M., Wilthagen, T. (Eds.), Systematic Occupational
Health and Safety Management: Perspectives on an International Development. Pergamon, Amsterdam, pp.
367–390.
Yassi, A., 1998. Utilizing data systems to develop and monitor occupational health programs in a large Canadian
hospital. Methods of Information in Medicine 37, 125–129.
Zaza, S., Wright-De Aguero, L.K., Briss, P.A., Truman, B.I., Hopkins, D.P., Hennessy, M.H., Sosin, D.M.,
Anderson, L., Carande-Kulis, V.G., Teutsch, S.M., Pappaioanou, M., 2000. Data collection instrument and
procedure for systematic reviews in the Guide to Community Preventive Services. Task Force on Community
Preventive Services. American Journal Preventive Medicine 18, 44–74.
View publication stats

The Effectiveness of Occupational Health and Safety Management System Interventions: A Systematic Review

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Effectiveness of Occupational Health and Safety Management System Interventions: A Systematic Review

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

The effectiveness of occupational health and safety management system

Article in Safety Science · March 2007

Kimberley L Cullen Amber Bielecky

SEE PROFILE SEE PROFILE

Emma Irvin Anthony John Culyer

SEE PROFILE SEE PROFILE

Dictionary View project

The user has requested enhancement of the downloaded file.

The eﬀectiveness of occupational health and safety

Keywords: Management system; Systematic literature review; Audit; Intervention; Eﬀectiveness

1.1. Deﬁning OHSMSs

‘‘OHSM systems are not, of course, a well-deﬁned set of management systems.

OHSMSs, commonly understood, are distinguishable from traditional OHS programs

• management commitment and resources,

1.2. Voluntary and mandatory OHSMS initiatives

1.3. Background literature and research justiﬁcation

1.4. Conceptual framework of the review

INTERVENTION IMPLEMENTATION INTERMEDIATE FINAL OHS

Fig. 1. Conceptual framework for the review on OHSMSs.

Intermediate OHS outcomes were considered to be outcomes of secondary interest,

2.1. Literature search

Eight electronic databases, abstracting primarily peer-reviewed research journal articles,

2.2. Selection of studies

Publication type. Only peer-reviewed journal articles were included.

Multi-faceted management system interventions were included if they had an occupa-

2.3. Quality appraisal (QA)

• very low (serious limitations),

2.4. Data extraction (DE)

2.5. Evidence synthesis

• bibliographic databases (18),

3.2. Quality of the literature investigating OHSMSs

L.S. Robson et al. / Safety Science 45 (2007) 329–353

The prevalence of methodological weakness was mainly a consequence of the experi-

3.3. Findings on the implementation and eﬀectiveness of voluntary OHSMSs

3.3.1. Description of the studies on the implementation and eﬀectiveness of voluntary

identiﬁcation, safety information management, and safety communication, overseen by a

3.3.2. Findings on the implementation of voluntary OHSMSs

3.3.3. Findings on the intermediate outcomes of voluntary OHSMSs

3.3.4. Findings on the ﬁnal outcomes of voluntary OHSMSs

3.4. Findings on the implementation and eﬀectiveness of mandatory OHSMSs

3.4.1. Description of the studies on the implementation and eﬀectiveness of mandatory

3.4.2. Findings on the implementation and eﬀectiveness of the IC legislation in Norway

3.4.3. Findings on the eﬀectiveness of simple OHSMS legislation in Canada

3.5. Evidence synthesis

3.5.1. Evidence for the eﬀectiveness of voluntary OHSMS interventions

• increased OHSMS implementation over time;

L.S. Robson et al. / Safety Science 45 (2007) 329–353

3.5.2. Evidence for the eﬀectiveness of mandatory OHSMS interventions

• increased OHSMS implementation over time;

4.1. Identifying and addressing research gaps

4.2. Strengths of the review

4.3. Limitations of the review

View publication stats

You might also like