BEU, Alm.del - 2021-22 - Bilag 267: Orientering om nye forskningsresultater om arbejdsmiljøindsatsen på virksomheder med et certificeret arbejdsmiljøledelsessystem, fra beskæftigelsesministeren

Beskæftigelsesudvalget 2021-22
BEU Alm.del Bilag 267
Offentligt

Safety Science 152 (2022) 105794

Contents lists available at

ScienceDirect

Safety Science

journal homepage:

www.elsevier.com/locate/safety

Differences in occupational health and safety efforts between adopters and

non-adopters of certified occupational health and safety

management systems

Christian Uhrenholdt Madsen

a, b

, Sannie Vester Thorsen

, Peter Hasle

Line Leonhardt Laursen

, Johnny Dyreborg

b, *

Team Worklife, Høffdingsvej 22, 2500 Valby, Denmark

The National Research Centre for the Working Environment, Lersø Parkalle 105, 2100 Copenhagen

Ø,

Denmark

Department of Technology and Innovation, University of Southern Denmark, Campusvej 55, 5000 Odense C, Denmark

A R T I C L E I N F O

Keywords:

OHSAS18001

ISO45001

Systematic process-related OHS

Decoupling

Window dressing

Denmark

A B S T R A C T

Certified occupational health and safety management systems (COHSMSs) continue to grow in popularity and to

diffuse to new industries. This study investigated differences in occupational health and safety efforts between

adopters and non-adopters of COHSMSs. We used cross-sectional survey data from 4,202 Danish workplaces from

all sectors to compare self-reported occupational health and safety efforts in workplaces with a COHSMS and

workplaces without a COHSMS. The ‘systematic process-related OHS efforts’ and ‘content-related OHS efforts’

were scored on five and seven scales, respectively, for both adopters and non-adopters. The results of linear

regression analysis revealed significantly lower score values for non-adopters than for adopters of COHSMSs,

which means certified workplaces perform better than non-certified workplaces in both process-related and

content-related OHS activities. We conclude that COHSMSs workplaces have a higher overall level of efforts for

both process and content OHS activities. The study therefore supports the assumption that COHSMS adopters

provide a higher level of OHS management than non-adopters, and that using the company’s OHS performance

as merely ‘window dressing’ is not a general feature of adopters. However, the results also indicate that a small

group of COHSMS adopters has a considerably lower level of OHS effort than non-adopters, which implies that

the certification system does not necessarily secure a high level of OHS management for all adopters. Further-

more, a small group of adopters have high process activities and low content activities, suggesting a decoupling

between the systematic OHS processes and the specific preventive activities in the workplace, which could be a

sign of window dressing. Further research is needed to establish the possible effects on health and safety out-

comes, such as lost-time injuries.

1. Introduction

Certified occupational health and safety management systems

(COHSMSs) continue to grow in popularity and to diffuse to new in-

dustries (Lafuente

and Abad, 2018).

Over the last decade, the OHSAS

18001 standard has been adopted globally by organizations as different

as Danish municipal workplaces (Jespersen

et al., 2016),

US metalworks

(Pagell

et al., 2014),

and Iranian consumer goods manufacturing com-

panies (Ghahramani,

2016).

This development is expected to be strengthened by the publication

of the ISO 45001 standard

–

the first ISO-standard on occupational

health and safety (OHS) management. But with COHSMSs becoming

institutionalized as a standard operating procedure in organizations

across the world, how effective these systems really are, has become an

important question for authorities, companies, and researchers alike

(Heras-Saizarbitoria

et al., 2019).

The literature is still not conclusive as

to whether or not the implementation of COHSMSs improves OHS out-

comes in companies. Whereas some studies have shown an overall effect

of reducing the number of fines after inspections (Lo

et al., 2014),

others

show no effect on accident outcomes (Heras-Saizarbitoria

et al., 2019).

* Corresponding author.

E-mail address:

[email protected]

(J. Dyreborg).

https://doi.org/10.1016/j.ssci.2022.105794

Received 1 November 2021; Received in revised form 11 April 2022; Accepted 15 April 2022

Available online 22 April 2022

BEU, Alm.del - 2021-22 - Bilag 267: Orientering om nye forskningsresultater om arbejdsmiljøindsatsen på virksomheder med et certificeret arbejdsmiljøledelsessystem, fra beskæftigelsesministeren

C. Uhrenholdt Madsen et al.

Safety Science 152 (2022) 105794

1.1. Effects of COHSMSs

In their review of OHSMSs

Robson et al. (2007)

found 13 original

studies but only one study was judged to be of high methodological

quality and the study concluded that the body of evidence was insuffi-

cient to make recommendations either in favour of or against OHSMSs.

In the last decade, this lack has been somewhat remedied, although the

number of studies is still low compared to studies on the effects of

certified management systems in related fields, such as environmental

protection or quality management (Heras-Saizarbitoria,

2018).

A num-

ber of studies show that there are some positive effects from COHSMSs

on the health and safety performance of adoptees.

Lo et al. (2014)

showed in a study of 211 US manufacturers that COHSMS adopters fared

better than non-adopters in terms of the relative number of safety vio-

lations. Likewise,

Abad et al. (2013)

could report two positive findings in

their study of 149 COHSMS adopters: that accident rates in the com-

panies decreased, and that this rate further decreased with the length of

time they held the certificate. A related study (Lafuente

and Abad, 2018)

showed that a time-dependent ‘learning effect’ was valid across the

various industry groups, but also that the effect on accidents was greater

in manufacturing companies than on construction sites or in office

settings.

However, there are other studies that find little or no effect of

COHSMSs. In their large-scale study of Spanish organizations and

companies,

Heras-Saizarbitoria et al. (2019)

found that COHSMS

adopters do not perform better on accident rates than non-adopters.

Ghahramani and Summala’s (2015)

study of Iranian manufacturing

companies also shows no effect of COHSMSs.

1.2. Mechanisms of COHSMSs

Even these studies investigated the effects of the adoption of

COHSMSs they have not clearly linked the certification process to

improved OHS, which leaves the role of the certification process

ambiguous in terms of how any effects on health and safety outcomes

were brought about.

A realistic review of the scientific literature on the effects of

COHSMSs suggests that effects are highly contextual and depend on how

systematic OHS practices fit into the overall practice of the adopting

companies (Madsen

et al., 2020).

This is a finding that is mirrored in

empirical studies as well (Lafuente

and Abad, 2018).

The ambiguous

evidence for the positive effects of COHSMSs notwithstanding, there is

still a lack of knowledge about how the systems work

–

in particular the

relationship between what

Øystein

Saksvik et al. (2003)

call ‘processual

claims’ (such as risk assessments, employee satisfaction surveys, and

mandatory top management reviews) and tangible ‘content claims’

(Øystein

Saksvik et al. (2003)).

In other words, we still do not know

whether the adoption of a management standard for OHS actually leads

to more and improved

systematic process-related OHS efforts

that in turn

lead to improved

specific content-related OHS efforts,

i.e., implementation

of specific preventive OHS measures at the shopfloor.

Taken together, the extant research point to the importance of

knowledge on how certified workplaces are actually affected by the

certification in relation to their process-related OHS activities and their

content-related OHS activities. Also, previous studies have mainly

considered the systematic proces-related OHS activities, and to a smaller

extent the content-related OHS activities at the workplaces (Hohnen

and

Hasle, 2011; Jespersen et al., 2016).

The present study seeks to open this black box and bridge these gaps

in the literature, by investigating whether COHSMSs lead to changes in

both process-related and content-related OHS activities in the work-

place. Knowledge about the impact of the mechanisms inherent in

COHSM on a workplace’s OHS activities is an important link in the

understanding of whether and how these mechanisms lead to positive

outcomes, such as reducing occupational injuries. Until now, most

comparative studies on COHSMSs have studied the difference between

adopters and non-adopters regarding various health and safety out-

comes. Our contribution to the literature is to investigate whether health

and safety practices are on a higher level in organizations that have

adopted COHSMSs compared to non-adopters, both in terms of more

systematic process-related OHS efforts and the subsequent imple-

mentation of tangible preventive measures in the workplace, i.e.,

content-related OHS efforts.

1.3. The programme theory of COHSMSs

It is reasoned by providers of OHS standards (see e.g. the foreword to

the Danish edition of OHSAS18001 (DS,

2010))

that companies by

implementing COHSMSs, are ensured compliance with legal re-

quirements across their extended organizations as well as integration of

OHS into the operational tasks. In its most basic form, the programme

theory behind COHSMSs consists of four steps (Madsen

et al., 2020):

The process of certification will lead the company, with the help of

external auditors, to uncover systems and processes that are not up to

standard. 2) This will lead the company to improve these systems and

processes, such as strengthened management involvement, active

participation of employees, and systematic risk assessment, which in

turn will lead to 3) the improvement of the actual preventive control

measures, such as noise reduction, fall protection, and safety in-

structions. 4) The outcome will be improved OHS performance, such as

reduction of occupational injuries (see

Fig. 1).

To ensure that a COHSMS

is implemented in practice, compliance with the process is verified by

third-party auditors.

However, as we have described in the introduction, most studies of

the effect of COHSMSs leap directly from Step 1 to Step 4, rendering the

actual mechanisms in Steps 2 and 3 as analytical ‘black boxes’ (Pawson

and Tilley, 1997).

Yet, from a previous review (Madsen

et al., 2020)

identified that these various mechanisms play an important role for the

COHSMSs to successfully improve OHS. This includes the following

mechanisms: how well the OHS system is integrated with other business

activities (‘integration of OHS’); the capability to organize and build up

knowledge (‘organisation of OHS’); the priority and engagement of the

top management and line managers in OHS activities (‘prioritisation of

OHS’); how the OHS systems is tailored to the organization’s needs

(‘translation and adaptation’); and finally focused priorities to ensure

that policies and procedures are translated into practical preventive

activities (‘attention’, in the sense of attention to the practical imple-

mentation of OHS efforts).

Although these basic mechanisms in the operation of COHSMSs are

generally well understood, there is still a lack of empirical studies on the

extent to which these mechanisms to a higher degree are present at

workplaces that have adopted a COHSMS compared to non-adopters of

COHSMSs. Following

Fig. 1,

we divide the OSH management efforts into

process-related OSH efforts (Step 2) and content-related OHS efforts

(Step 3).

1.4. Process-related OHS management

Process-related OHS efforts are by their nature non-specific with

regard to particular risks or hazards at work, in much the same way as is

quality management (Frick

and Wren, 2000).

Process-related OHS ac-

tivities denote efforts to monitor workplaces and organizations, and

thereby identify possible hazards for workers’ health and safety.

Bluff

(2003)

suggests that these efforts consist of a number of related activ-

ities, such as ‘integration of OHSM’ into other business activities,

‘management commitment’, ‘OHS policy’, ‘planning and resourcing of

OHS management’, ‘designation of responsibility and mechanisms of

accountability’, ‘procedures and documentation’, ‘risk management’,

‘worker participation’, ‘development of OHS competency’, ‘reporting’,

‘investigating and correcting deficiencies’, and ‘monitoring, auditing

and reviewing OHS performance’ (Bluff,

2003, p. 7).

These related ac-

tivities are well in accordance with the mechanism identified in the

BEU, Alm.del - 2021-22 - Bilag 267: Orientering om nye forskningsresultater om arbejdsmiljøindsatsen på virksomheder med et certificeret arbejdsmiljøledelsessystem, fra beskæftigelsesministeren

C. Uhrenholdt Madsen et al.

Safety Science 152 (2022) 105794

Fig. 1.

Programme theory of COHSMSs effects.

realistic evaluation of COHSMSs, as mentioned previously (Madsen

et al., 2020).

We might assume that adopters of COHSMSs will outperform non-

adopters in implementing process-related OHS management activities,

and thereby empirically support a confirmation of the programme

theory.

1.5. Content-related OHS management

Content-related OHS management covers the tangible control mea-

sures that usually follow a risk assessment (a process activity), which has

considered something hazardous in an organization. Such concrete and

practical control measures can be as diverse as adjustment of working

height and design of workstations to avoid hazardous work postures in

manufacturing settings, the replacement of ladders with scaffolding in

the construction industry or preventing conflicts and bullying in an of-

fice community. Companies across different industries have to deal with

very different work environment problems, but the general character-

istic of content-related OHS management is that it covers the imple-

mentation of tangible control measures, such as changes in work

procedures or the establishment of barriers to reduce hazardous expo-

sures. A logical consequence of systematic process-related efforts is that

the certified workplaces will show better performance of this step in

terms of implementing control measures.

Importantly for the present paper, only two studies have compared

adopters and non-adopters of COHSMSs to examine whether adopters

have actually implemented OHS efforts to a greater extent in comparison

with non-adopters.

Ghahramani and Summala (2015)

investigated OHS

management practices in three adopters and three non-adopters of

COHSMSs in power, oil and gas facilities in Iran. They found that

adopters were more likely to enforce OHS rules and procedures than

non-adopters, and the level of safety training was also higher among

adopters. Furthermore, the study showed a significant positive differ-

ence for most of the 44 key performance indicators included. A study of

safety management practices and safety behaviour in eight chemical

manufacturing plants in India’s Kerala province by

Vinodkumar and

Bhasi (2011)

showed a significant difference between certified and non-

certified companies in terms of safety management activities. The

certified companies performed better on training, management

commitment, communication, rules and procedures, promotion policies

and workers involvement in safety related matters.

However, both studies have rather small sample sizes, evaluate key

performance indicators based on single item (question) without scale

validations, and were limited to larger enterprises in the chemical in-

dustry with a risk of major accidents. In the present study, we seek to

open the black box concerning the difference in OHS activities between

COHSMS adopters and non-adopters across the key sectors of society.

2. Material and methods

The overall design compares non-adopters of COHSMSs with

adopters of COHSMSs to see whether the latter performs better in terms

of process-related OHS efforts and content-related OHS efforts. We used

a national questionnaire ‘Work Environment Activities in Danish

Workplaces’ (WEADW) to evaluate OHS efforts at workplace level

(Thorsen

et al., 2017).

The questionnaire was administered to both

adopters and non-adopters of COHSMSs.

To identify adopters of COHSMSs we obtained a register including all

workplaces registered with a COHSMS by 9 April 2018 from the Danish

Working Environment Authority (WEA). This list included the Central

Business Register numbers for both the main company and for its local

workplace(s), allowing us to obtain information on company name, size,

and addresses at workplace level from the Central Business Register in

Denmark. We consider this list to be rather complete, as workplaces with

a COHSMS was exempted from a visit from the Labour Inspectorate and

furthermore, workplaces were listed with a ‘Smiley’ on the homepage of

the WEA, indicating good OHS performance.

The register of COHSMS adopters also provides information on the

status of the certificate for each workplace (the company’s local units) in

a company. A third-party certification body checks the status of the

implementation of a COHSMS during the audit process, as set out in the

legislation and/or in the terms and conditions applicable to obtaining a

certificate. Workplaces are excluded from the study if they have not

passed this audit process or if they for other reasons do not have an

active certificate in the study period.

The questionnaire was sent to all workplaces in Denmark that were

registered with a COHSMS (N

2596) conforming to Danish law.

However, only those workplaces that answered ‘Yes’ to a question about

whether the company had an OHS committee (n

630) were included in

the analyses (Table

1).

By Danish law all companies with 35 or more

employees are required to have an OHS committee. The questionnaires

were send to one employee representative and one employer represen-

tative at each workplace. If we received two answers from a workplace,

i.e., both an employer and employee questionnaire, we took the average

value unless the question was a ‘yes’/’no’ question, in which case a ‘yes’

overruled a ‘no’.

The comparison group of non-adopters consists of a stratified sample

of workplaces in 108 strata to ensure comparability (N

8100), as there

are far more small than large workplaces in Denmark, and the distri-

bution by size differs greatly in the various sectors. In each stratum, the

workplaces were randomly selected from the Central Business Register

by size (workplaces with 1–9 employees, 10–34 employees, 35 or more

employees) and industry group (36 different groups). The distribution of

workplaces within industry groups and within workplace size groups is

relatively similar, for adopters and non-adopters of COHSMS, respec-

tively (Table

1).

The average response rate for all workplace sizes is 46% and the

average response rate for workplaces with 35 or more employees is 65%,

where the latter represent the workplace sizes included in the present

study. To establish a comparison group of non-adopters of COHSMSs we

excluded adopters of COHSMSs from the stratified random sample and

kept the remaining (about 90%) workplaces, which have not adopted

BEU, Alm.del - 2021-22 - Bilag 267: Orientering om nye forskningsresultater om arbejdsmiljøindsatsen på virksomheder med et certificeret arbejdsmiljøledelsessystem, fra beskæftigelsesministeren

C. Uhrenholdt Madsen et al.

Safety Science 152 (2022) 105794

Table 1

The number of worksites included for each industry group and for each size

group of workplaces and the percentage distribution, for adopters and non-

adopters of COHSMS, respectively.

COHSMS status

Industrial main groups:

Construction

Trade

Manufacturing industry

Office and communication

Agriculture and food

Public service

Private service

Transport

Social and health

Teaching and research

Adopters

630

150

167

24%

27%

100%

33%

34%

20%

100%

Non-adopters

1917

111

499

154

156

191

147

301

174

26%

10%

16%

100%

34%

41%

13%

100%

2.2. Scoring of OHS efforts scales

In order to evaluate the level of OHS efforts in adopters and non-

adopters of COHSMSs, we scored questions on a Likert scale from

lowest to highest answer category where the scores ranged from 0 to 100

points. ‘Do not know’ was scored as ‘no’, that is the lowest category, and

‘not relevant’ was set as ‘missing’. The exceptions were the questions in

the OHS efforts scales, ‘4. Work Place Assessment (WPA) for psychoso-

cial and physical risks’ and ‘5. WPA for chemical risks’, which were

scored as ‘Yes’=100, ‘No’=0, ‘Do not know’=0, and the questions in the

OHS efforts scales ‘10. Efforts to reduce threats and emotional demands’

and ‘11. Efforts to reduce physical risks’, where the answer category,

‘The workplace does not have any workers this is relevant for’, was

scored as a ‘No’=0.

We scored the 12 scales as the average score of the respective

questions in the scale if half of the questions in the scale were answered.

Furthermore, we prepared two overall scales, aggregating scales 1–5

into one scale for process-related OHS efforts, and scales 6–12 into one

scale for content-related OHS efforts. The overall scales were calculated

if answers to at least half of the sub-scales were present.

2.3. Covariates

We adjusted for the following covariates: industry group (10 groups)

and the size of the workplace as number of full-time employees. Both

industry group and number of employees were taken from national

registers (Statistics Denmark). We also controlled for whether it was an

employee or an employer representative, or both, who answered the

questionnaire for the workplaces.

2.4. Statistical analyses

All analyses were performed using SAS 9.4. We performed explor-

atory factor analysis using both unweighted least squares as the

extraction method and principal component analysis with oblique

rotation of the factors to construct the scales for the present study using

‘proc factor’ from SAS (Joliffe

and Morgan, 1992; Osborne and Banja-

novic, 2016).

We provide descriptive statistics of the mean value of the

OHS efforts scales for COHSMS adopters and non-adopters respectively.

We divided the two ‘overall’ scales (process-related OHS efforts,

content-related OHS efforts) along their 40th percentile from the non-

adopters’ dataset to make a cross-table.

By visual inspection, we examined the distribution of all OHS scales.

The 12 individual sub-scales had small or large deviations from a normal

distribution, the two overall scales, ‘process-related OHS’ efforts and

‘Content-related OHS’ efforts, were normally distributed. We therefore

used a non-parametric (Mann-Whitney-Wilcoxon) test to examine if

each OHS sub-scale was different between COHSMS adopters and non-

adopters. We used a parametric student’s

t-test,

with unequal vari-

ances for the overall scales. We used unequal, since the variance of

COHSMS adopters were smaller than non-adopters. We divided the data

up on 10 different industrial sectors (10 different analyses for each

scale). Last, we examined if the COHSMS adopters compared to non-

adopters had higher scores on process-related OHS efforts and

content-related OHS efforts adjusted for industrial sector and number of

employees by using general linear regression with the GLM procedure in

SAS 9.4. Industrial sector was included as a categorical variable and

number of employees as an interval variable.

3. Results

Adopters of COHSMSs have significantly higher scores on 10 of the

12 scales (Table

2).

There is a tendency for adopters to perform partic-

ular well on process-related efforts, which could be expected as this is

the basic requirement in a COHSMS. With controls for size and sector

differences, the scale values for processual OHS efforts are overall

Workplace size groups:

employees 1–9

employees 10–34

employees 35–99

employees 100–249

employees 250–499

employees 500 or more

205

215

126

630

113

653

792

258

1917

COHSMSs (non-adopters), as the comparison group (N

7543). Also in

this sample we only included those that answered ‘Yes’ to a question

about whether the company had an OHS committee (n

1917) (Fig.

2).

The responses for the stratified sample were collected between late

2017 and mid-2018, and the responses from the COHSMSs group were

collected mid-2018.

2.1. Development of scales to measure workplace OHS efforts

We developed 12 scales for the OHS efforts related to process-related

OHS efforts and content-related OHS efforts by use of exploratory factor

analysis (EFA) of the data from the stratified WEADW 2017 survey.

First, we identified 84 questionnaire items related to workplace OHS

efforts from the WEADW questionnaire on the basis of the OHS man-

agement literature (Bluff,

2003; Frick et al., 2000; Saksvik and Quinlan,

2003).

Then we divided these items into process-related and content-

related questionnaire items, and within each of these two groups of

questions we established scales based on their loadings on relevant

factors by use of the EFA. These factors (indicators) are well in accor-

dance with the five important mechanisms for successfully improving

the OHS, that was identified in a previous study (Madsen

et al., 2020):

Prioritization of OHS (‘Commitment’); Integration of OHS with man-

agement activities (‘integrative

aspect’);

Organisation of OHS (‘organisa-

tional learning’);

Work Place Assessment (‘Translation

and adaptation’);

The seven indicators for content-related OHS efforts, i.e., actual pre-

ventive activities (‘attention’).

This led to the construction of 12 scales, five process-related and

seven content-related scales. They were Likert scale questions scored as

1, 2, 3, etc. from lowest to highest answer category, apart from questions

about work place assessment, where only yes/no response categories

were available. ‘Not relevant’ and ‘do not know’ were scored respec-

tively as ‘missing’ and as ‘no’. We tested the internal consistency of the

scales using Cronbach’s alpha, which was 0.7 or higher for ten of the 12

scales, indicating high internal consistency. The remaining two scales

(Organization of OHS and Efforts to prevent accidents) showed a

Cronbach’s alpha above 0.6, indicating moderate, but still acceptable,

internal consistency.

BEU, Alm.del - 2021-22 - Bilag 267: Orientering om nye forskningsresultater om arbejdsmiljøindsatsen på virksomheder med et certificeret arbejdsmiljøledelsessystem, fra beskæftigelsesministeren

C. Uhrenholdt Madsen et al.

Safety Science 152 (2022) 105794

Fig. 2.

Study populations and data collection.

significantly higher among adopters than non-adopters for the combined

scales and for three out of five of the specific scales. In particular, the

integration of work environment activities with other management ac-

tivities is clearly greater in adopters than in non-adopters, indicating

that the certification process helps the workplaces to integrate the OHS

efforts into the management system.

However, the differences between adopters and non-adopters were

not significant for the two scales for WPA. It can also be an artefact, as

this questions was not a Likert scale but only a yes/no answer category,

and therefore difficult to discriminate various levels of performance.

The mean difference between adopters and non-adopters is found

across almost all main sectors in society for both process-related OHS

efforts and content-related OHS efforts (Table

3).

The difference is sig-

nificant with the exception of Trade on both the combined scales and

Construction and Public Service for the content-related OHS efforts.

The linear regression analysis shows that with controls for workplace

size and sector differences, the scale values for both process-related OHS

efforts and for the specific preventive OHS efforts are overall signifi-

cantly higher among adopters of COHSMSs compared to non-adopters

(Table

4).

Certification with third-party verification is expected to ensure that

all adopters maintain a high level of both systematic and preventive

effort. The scatter plots in

Fig. 3

confirm that most adopters have a high

level on both scales, whereas non-adopters are distributed more widely

on both scales. However, a smaller group of adopters are low on either

both general scales or low on one of the two general scales, and thus

have a performance lower than the 40th percentile of non– adopters.

Both

Table 4

and the scatter plots in

Fig. 3

show that adopters do

better than the average non-adopter. In

Table 5,

we use the 40th

percentile of the non-adopters to show how adopters and non-adopters

are distributed on low/high for both process-related OHS efforts and

content-related OHS efforts. The table shows most adopters have a high

performance on both process-related and content-related OHS efforts,

and a small fraction have a low performance on both scales, indicating

that they might be examples of adopters that adopted COHSMSs for

ceremonial reasons. A group of approximately 25 pct. (9.52%

14,76%) of the workplaces scored low on one of the two scales, but

interestingly with a relatively higher scores for process-related OHS

BEU, Alm.del - 2021-22 - Bilag 267: Orientering om nye forskningsresultater om arbejdsmiljøindsatsen på virksomheder med et certificeret arbejdsmiljøledelsessystem, fra beskæftigelsesministeren

C. Uhrenholdt Madsen et al.

Safety Science 152 (2022) 105794

Table 2

Comparison of mean score values for adopters and non-adopters.

Scale

Process-related OHS efforts

Prioritization of OHS activities

Integration of OHS in management and

operations

Organization of OHS

WPA for psycho-social and physical risks

WPA for chemical risks

Process-related OHS efforts (overall scale)

Content-related OHS efforts

Safety guidance and instruction

Efforts to prevent accidents

Psycho-social prevention efforts

Efforts to reduce conflicts and bullying

Efforts to reduce threats and emotional demands

Efforts to reduce physical risks

Efforts to reduce chemical risks

Content-related OHS efforts (overall scale)

Adopters Mean

value

83.1

86.8

93.3

73.4

23.8

72.1

Non-Adopters Mean

value

75.7

60.1

82.3

72.8

24.5

63.1

Difference Adopters-Non

adopters

7.3

26.7

0.6

0.7

P-value

18,4

19,3

17,8

26,8

32,4

14,3

19,4

32,5

31,5

<0.0001

0.4372

0.1444

<0.0001

10.

11.

12.

83.5

80.7

61.8

65.3

43.3

55.3

50.3

63.9

17,9

18,5

12,4

32,8

38,5

32,3

14,8

74.3

71.9

58.4

58.5

34.6

45.3

44.3

56.1

22,1

22,2

13,8

33,7

36,7

32,8

32,3

16,2

9.2

8.8

3.4

6.8

8.7

7.8

<0.0001

0.0027

<0.0001

Note: The standard deviation (SD), the mean difference, and significance of difference (P-value). P-value is calculated by the non-parametric Mann-Whitney-Wilcoxon

test (several sub-scales are not normally distributed).

Table 3

Comparison of mean score values for adopters and non-adopters distributed across main sectors.

Sector

Process-related OHS efforts

Construction

Trade

Manufacturing industry

Office and communication

Agriculture and food

Public service

Private service

Transport

Social and health

Teaching and research

Content-related OHS efforts

Construction

Trade

Manufacturing industry

Office and communication

Agriculture and food

Public service

Private service

Transport

Social and health

Teaching and research

Adopters Mean value

77.3

68.9

76.1

72.3

75.9

71.1

68.2

70.4

11,7

7,8

15,1

12,3

13,7

16,6

13,1

13,4

15,2

Non-adopters Mean value

66.1

62.8

63.9

55.2

62.8

58.8

63.8

62.6

18,8

17,2

17,4

16,2

17,9

16,2

18,6

14,8

15,3

15,6

Difference Adopters-Non adopters

11.2

6.1

12.1

17.1

8.9

7.1

13.2

9.4

4.4

7.8

P value

<0.0001

0.0212

<0.0001

0.0055

0.0210

0.0002

<0.0001

0.0015

0.0050

60.2

62.7

61.4

62.7

60.1

67.1

66.4

65.8

61.7

15,8

11,4

12,9

15,6

10,3

17,1

15,6

14,7

15,7

15,6

56.1

56.6

55.9

45.2

55.6

57.8

57.1

55.2

62.2

53.5

17,3

15,6

15,2

16,1

14,7

14,9

17,3

16,6

15,8

15,4

4.1

6.1

8.1

16.3

7.1

2.3

11.3

3.7

8.2

0.1530

0.0615

<0.0001

0.0042

0.4476

0.0022

<0.0001

0.0163

0.0038

The standard deviation (SD), the mean difference, and significance of difference (P-value). P-value is calculated by the parametric student’s

t-test

(unequal variances).

Table 4

Linear regression of the association between score values and adopters/non-

adopters adjusted for sector and number of employees at the workplace.

Scale

Process-related OHS efforts (overall scale)

Content-related OHS efforts (overall

scale)

Estimate

9.3

7.34

95% CI

[7.82–10.78]

[5.92–8.77]

P value

<0.0001

4. Discussion

The results support the program theory for yielding positive effects of

COHSMSs as the regression analyses show that workplaces that have

adopted COHSMSs, overall perform better compared to non-adopters of

COHSMSs, in terms of both process-related and content-related OHS

efforts. Furthermore, for ten of the twelve subscales included in the

analyses, adopters of COHSMSs perform better (had higher scores)

compared to non-adopters. When comparing the mean scores for both

process-related and content-related OHS efforts, the results show that

the adopters performs better than non-adopters in most industrial

sectors.

The five key mechanisms that enables COHSMSs to function effi-

ciently described in the OHS literature (Madsen

et al., 2020)

are sup-

ported by our findings. In particular the indicator ‘integration of OHS in

management and operations’ (the ‘integrative aspect’ mechanism)

The p-value is calculated based on a t-distribution.

efforts but conversely lower scores for content-related OHS efforts.

There is thus a group of adopters, who do well on the process-related

OHS efforts, but who do not follow up with content-related OHS ef-

forts to the same extent.

BEU, Alm.del - 2021-22 - Bilag 267: Orientering om nye forskningsresultater om arbejdsmiljøindsatsen på virksomheder med et certificeret arbejdsmiljøledelsessystem, fra beskæftigelsesministeren

C. Uhrenholdt Madsen et al.

Safety Science 152 (2022) 105794

Fig. 3.

Scatter plots of adopters and non-adopters distributed on the two general scales. Cross-hairs indicate the 40th percentile among non-adopters of COHSMSs of

respectively the systematic OHS process-scale and specific preventive OHS effort-scale.

Table 5

The cut-points for the division in high and low for the two scales are set as the

40th percentile based on the scales’ distribution among Non-adopters.

Adopters

Process-related OHS efforts

Low

High

Low

6.51

14.76

High

9.52

69.21

Non-adopters

Low

24.67

15.28

High

15.28

44.76

Content-related OHS efforts

shows significantly higher scores for adopters compared to non-

adopters. This indicates that adopters integrate OHS efforts with other

business activities to a much greater extent than non-adopters. Adopters

also score higher on ‘prioritisation of OHS activities’ (‘commitment’

mechanism) and ‘organisation of OHS’ (‘organisational learning’

mechanism) compared to non-adopters.

For the two WPA indicators (related to the

’translation

and adapta-

tion mechanism’) representing risk assessment of

’psychosocial

and

physical risks’ and of ‘chemical risks’, respectively, the differences are

small and non-significant. We assume that the latter reflects that WPA is

a basic legal requirement in Denmark, which is relatively easy and un-

ambiguous to enforce for the authorities, and thus contributes to a high

degree of compliance for most of the larger companies, regardless of

certification status. It can also be related to an artefact, as we only use

yes/no for these scales.

In terms of content-related work environment efforts, we include

seven indicators that focus on the extent to which prioritised policies

and procedures are translated into practical prevention activities

(’attention mechanism’). For all of the content-related OHS efforts, the

COHSMS adopters have a significantly higher score, and thus indicate

that adopters to a higher degree translate policies and procedures into

practical preventive OHS efforts. This mechanism works through the

resources of audits and management reviews, which direct stakeholders

to focus on particular OHS elements, pertinent to the risk assessment

activities (WPA indicators).

It has been highlighted in the scientific literature that psychosocial

risk factors are not very well captured by certification systems, such as

OHSAS18001, and that the psychosocial risks factors might be unfa-

miliar to the auditors and thus are more difficult for auditors to assess

compared to other risk factors, such as accidental risks (Hohnen

and

Hasle, 2011; Jespersen et al., 2016).

According to this, there could be a

risk that certification systems contribute to workplaces downgrading

their focus on specific important risk factors, such as the psychosocial

risk factors. However, our results do not indicate that adopters have less

focus on the psychosocial risk factors, but on the contrary, our results

show that adopters give higher priority to psychosocial prevention ef-

forts, as well as more specific efforts aimed at reducing conflicts and

bullying in the workplace and efforts to reduce threats and emotional

demands at work.

Our data is cross-sectional, and we are therefore unable to draw firm

conclusions about the causality of adopting COHSMSs and a subsequent

increase in the performance of OHS activities. Nevertheless, the results

confirm the expectation that certification would correlate with higher

performance on both systematic process-related and content-related

OHS efforts. Our study therefore helps explain the causal mechanisms

behind earlier results (Lafuente

and Abad, 2018; Lo et al., 2014),

which

show that certification leads to better OHS outcomes. In other words, the

certified companies are better at systematically monitoring risks and

having updated policies in place when it comes to, for instance, ma-

chinery and other workplace equipment, and in turn translate this into

content-related OSH efforts.

It is to be expected even for adopters of COHSMSs that the perfor-

mance on both systematic and preventive efforts will have a normal

distribution, but certification with third-party verification ought to

ensure that the really bad performers are either corrected or have their

certificates suspended. It is therefore surprising that our results indicate

that there exists a small group of companies that perform as badly as the

poorest performers among the non-adopters. This finding suggests that

the audit verification system is not able to identify and remedy all bad

performers.

We cannot provide a firm explanation for this outcome but the

literature points towards so-called ‘window dressing’ (Blewett

and

O’Keeffe, 2011; Brunsson et al., 2012; Rocha and Granerud, 2011)

and

‘bureaucratization’ as risks in the application of COHSMSs. Window

dressing implies that companies adopt programmes to appear legitimate

in the eyes of customers, regulators, and competitors (Heras-Saizarbi-

toria et al., 2013).

It has furthermore been argued that such adoption of

COHSMSs purely for legitimacy reasons leads to a decoupling of

COHSMSs from the central decisions and strategies in the adopting

companies (Hasle

et al., 2021),

resulting in very small, if any at all, ef-

fects on the actual OHS practices in the companies.

Notwithstanding the methodological limitations of cross-sectional

studies, the results from this study can refute the window dressing as-

sumptions to some extent as the large majority of adopters perform

considerable better than non-adopters. If window dressing was the

dominant factor, there would not be significant differences between

certified and non-certified companies on preventive OHS efforts. Our

study therefore does not support the findings of

Heras-Saizarbitoria et al.

(2019),

which indicated that the search for legitimacy is a dominant

feature among certified companies. On the contrary, our study suggests

that adopters of COHSMSs seem to have a higher level of both process-

BEU, Alm.del - 2021-22 - Bilag 267: Orientering om nye forskningsresultater om arbejdsmiljøindsatsen på virksomheder med et certificeret arbejdsmiljøledelsessystem, fra beskæftigelsesministeren

C. Uhrenholdt Madsen et al.

Safety Science 152 (2022) 105794

related and content-related OHS efforts compared to non-adopters, and

that adopting OHSMS is an indicator of a high level of OHS activities,

and thereby provide a useful tool for companies to improve their OHS

activities.

However, while the large majority of certified companies perform

better than the non-adopters, our results do show that a small fraction

performs poorly. While it is not possible from the data to assess to what

extent certification requirements are fulfilled, it seems likely that the

6.5% of adopters (the same percentage is 25 % for non-adopters) per-

forming at a low level for both systematic and preventive efforts will

have problems with certification compliance. Furthermore, 15%

perform at a high level on systematic processes, but at a low level on

content-related OHS efforts. These results support a tendency to

decoupling in practice where companies either generally have a low

level of OHS management or make up ‘paper work’ that does not support

tangible preventive efforts. While we cannot draw any conclusions

about motives from our data, the results do suggest a decoupling be-

tween the certificate and practice in this small group of companies, and

an element of window dressing to look legitimate is a likely explanation

for a least some of this behaviour.

Another part of the explanation for the low performance in some

workplaces on content-related OHS efforts may relate to the criticism of

certification for being bureaucratic and superfluous when it comes to

tangible control of risks in the companies. Some sources criticize a

general tendency towards bureaucratization in OHS-management

(Dekker,

2014),

claiming that systematization creates ‘number games’

and shifts focus to known and expected patterns in health and safety,

which prevents bureaucratized organizations from reacting to unex-

pected incidents or to multi-factor issues that do not fit the dominant

understanding of OHS (Madsen

and Hasle, 2017; Nielsen, 2000).

This

echoes a widespread view on certifications and systems pointing out that

only visible and ‘auditable’ incidents and factors are identified and

prevented (Power,

1997).

This perspective has inspired studies that

describe how complex risks such as psychosocial factors affecting well-

being were not being detected and prevented adequately by certified

management systems (Hohnen

et al., 2014; Hohnen and Hasle, 2011).

Bureaucratization does generally not seem to be a problem as the

great majority of adopters perform better on both process and content-

related OHS efforts. Yet, the results show that for a small group of

workplaces there might be a decoupling between the systematic pro-

cesses and the practical preventive efforts. While this will not be

considered a problem in a majority of companies with fairly well-

functioning organizations, the problem can arise if the adopting com-

pany does not possess the capabilities to run a systematic organization to

begin with and/or focuses more on signals to external stakeholders than

the internal outcomes of the management system.

4.1. Implications for practice

Our study has three major implications for practitioners. First of all,

while the majority of certified workplaces actually do quite well both in

terms of content and process OHS activities, we can see that there is still

room for improvement. It is important that OHS-professionals and

managers in certified companies make sure they follow up on the risk

assessments and ensure that the actual preventive activities identified

through the risk assessment, are implemented. The results of the study

suggest that there are some certified workplaces where this is not the

case. Secondly, certification bodies and their auditors need to be aware

that there are some certified workplaces who have a very low OHS

performance, and that it is the responsibility of the auditors to identify

and remedy this problem. Thirdly, the tendency to include COHSMSs in

national legislation implies that law makers and regulatory agents

should pay attention to poorly performing certified companies and

develop tools and processes to find the ‘bad apples’.

4.2. Implications for future research

This study also indicates four distinct future avenues of research,

which can further enhance our understanding of the effects of COHSMSs

on OHS performance and management.

Further research is needed on the causality in the relationship be-

tween certification and improved OHS management activities. It is

important for future research to apply longitudinal data from certified

companies to establish whether they improve their OHS activities over

time or whether workplaces that adopt COHSMSs simply have a higher

performance when first certified. A related factor is whether certifica-

tion can provide continuous improvements or the effect of a COHSMS

wanes after fixing ‘low-hanging fruits’ in terms of OHS issues. Only

research that has access to historical or prospective follow-up data on

OHS activities from a large group of companies will be able to answer

these questions.

Furthermore, it is important to establish whether the stronger man-

agement performance of certified companies also means that actual OHS

outcomes improve in certified companies. Have the risks of accidents,

fatal accidents, and lost workdays due to occupational health problems

actually been reduced in these companies over time since certification?

An avenue for future studies would also be to gain a deeper under-

standing of the reasons some companies are more successful than others

in implementing COHSMSs. Are there any patterns in the mechanisms

applied that make some companies successful with COHSMSs while

others fall into the decoupling trap? This would be a question of interest

to practitioners and regulators alike, as certification plays an increasing

role in supplier selection, tenders, and regulatory schemes across

Europe. This avenue of research should also study the integration with

other management systems in particular ISO 9001 and ISO 14001. A

new line of research is increasingly pointing towards the advantages for

application of integrated systems (Wilkinson

Dale, 1999; Kauppila

et al., 2015).

Finally, earlier literature suggested that OHS management would

have a positive impact on economic performance (Pagell

et al., 2014),

research is needed to confirm this relationship, also with regard to

productivity, quality, and the environment. It would be particularly

interesting to illuminate the mechanisms behind such positive impacts.

5. Limitations

One limitation of this study was that we used data from an existing

national workplace-survey, ‘Work Environment Activities in Danish

Workplaces’ that was not specifically developed to investigate differ-

ences between adopters and non-adopters of COHSMSs. Also, the

questions in the survey were not designed to capture all elements of

systematic OHS management, even though it could capture workplace

responses on the most important OHS efforts, which we used to establish

scales for the analyses. On certain dimensions only a limited number of

questions covered the scale, which can give more uncertain measures of

a scale.

Another limitation is that the OHS efforts are self-reported which can

introduce some bias in the responses, as workplaces want to indicate

that OHS is under control. However, we have used answers from both

employer and employee representatives to evaluate OHS at each work-

place. We assume that this has alleviated the bias of self-reporting.

Even we excluded workplaces with COHSMSs from the comparison

group, some workplaces might use OHSMS without being certified and

also using certification systems that are not approved by Danish au-

thorities, but still might work well. However, this would bias our results

towards the null, which means that we would see a smaller differences

between adopters and non-adopters. For the two scales for WPA

(workplace assessment) we only had the answer category yes or no to

distinguish between adopters and non-adopters, making it more difficult

to discriminate between various levels of adoptions. This will also bias

our results towards the null and might be a plausible explanation for not

BEU, Alm.del - 2021-22 - Bilag 267: Orientering om nye forskningsresultater om arbejdsmiljøindsatsen på virksomheder med et certificeret arbejdsmiljøledelsessystem, fra beskæftigelsesministeren

C. Uhrenholdt Madsen et al.

Safety Science 152 (2022) 105794

having a significant difference between adopters and non-adopters for

these two scales.

Finally, we have no information on the quality of the implementation

of the content-related OHS efforts, and neither the process-related OHS

efforts. For this reason, it can be uncertain to what extent that OHS ef-

forts translate into tangible outcomes, such as occupational injuries.

6. Conclusion

This paper contributes to several important aspects of the discussion

on the effects of COHSMSs. First of all, the results clearly indicate that

COHSMS adopters have a better performance on both process-related

OHS effort and content-related OHS efforts, and it is a performance

which is found across industrial sectors and workplace sizes. The study

thereby is supportive of earlier findings of a positive effect on OHS

efforts.

While the superior performance with regard to process-related OHS

efforts is to be expected given that COHSMSs are designed to improve

these processes, it is even more important that our results show that

adopters of COHSMSs also outperform non-adopters in terms of content-

related OHS efforts. Furthermore, our results show that COHSMS

adopters give higher priority to psychosocial prevention efforts, as well

as more specific efforts such as those aimed at reducing threats and

emotional demands at work. This is interesting as previous studies have

suggested that complex OHS issues, such as psychosocial OHS risks, are

deprioritized or simplified by COHSMSs.

Also, we found no support for the frequent criticism that COHSMSs

mainly increases paperwork and bureaucracy in safety without

improving the quality of the tangible control of OHS risks. While our

study seems to refute window dressing and bureaucratization as a gen-

eral tendency, we did, however, find a small percentage of certified

companies where certification generally show low performance for all

dimensions of OHS efforts or little or no apparent effect on the level of

content-related OHS efforts. There is a small minority of adopters, whose

performance is so poor in general that it should have been identified by

third-party audits.

With this conclusion, further research is needed to fully capture the

effect of COHSMSs and also to find out how the risk of window dressing

and bureaucratization can be avoided to a larger extent. For instance,

analyses of historical data would be necessary to finally prove a causal

link between certification and the quality of activities. And more case

studies will be needed to investigate the mechanisms that lead to suc-

cessful and unsuccessful adopters, respectively.

Credit authorship contribution statement

Christian U. Madsen:

Conceptualization, Methodology, Writing -

Original draft preparation.

Sannie V. Thorsen:

Data curation, Formal

analysis, Writing - Review

Editing.

Peter Hasle:

Conceptualization,

Methodology, Writing - Review

Editing, Funding acquisition.

Line L.

Laursen:

Data curation, Formal analysis, Comments and Writing - Re-

view

Editing.

Johnny Dyreborg:

Conceptualization, Methodology,

Writing- Review

Editing, Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence

the work reported in this paper.

Acknowledgement

This research was funded by the Danish Work Environment Research

Fund (Grant no. 40-2017-03).

References

Abad, J., Lafuente, E., Vilajosana, J., 2013. An assessment of the OHSAS 18001

certification process: objective drivers and consequences on safety performance and

labour productivity. Saf. Sci. 60, 47–56.

https://doi.org/10.1016/j.

ssci.2013.06.011.

Blewett, V., O’Keeffe, V., 2011. Weighing the pig never made it heavier: Auditing OHS,

social auditing as verification of process in Australia. Saf. Sci. 49, 1014–1021.

https://doi.org/10.1016/j.ssci.2010.12.010.

Bluff, L., 2003. Systematic management of Occupational Health and Safety, in: Working

Paper 20. National Research Centre for OHS Regulation, Canberra.

Brunsson, N., Rasche, A., Seidl, D., 2012. The Dynamics of Standardization: Three

Perspectives on Standards in Organization Studies. Organization Stud. 33, 613–632.

https://doi.org/10.1177/0170840612450120.

Dekker, S.W.A., 2014. The bureaucratization of safety. Saf. Sci. 70, 348–357.

https://doi.

org/10.1016/j.ssci.2014.07.015.

DS, 2010. Arbejdsmiljøledelsessystemer, 2nd ed. Dansk Standard.

Frick, K., Langaa Jensen, P., Quinlan, M., Wilthagen, T., 2000. Systematic Occupational

Health and Safety Management - Perspectives On An International Development.

Emerald Group Publishing.

Frick, K, Wren, J., 2000. Reviewing occupational health and safety management:

multiple roots, diverse perspectives and ambiguous outcomes, in: Frick, Kaj,

Quinlan, M., Jensen, P.L., Wilthagen, T. (Eds.), Systematic Occupational Health and

Safety Management - Perspectives On An International Development. Emerald Group

Publishing Limited, pp. 527–527.

Ghahramani, A., 2016. Factors that influence the maintenance and improvement of

OHSAS 18001 in adopting companies: A qualitative study. J. Cleaner Prod. 137,

283–290.

https://doi.org/10.1016/j.jclepro.2016.07.087.

Ghahramani, A., Summala, H., 2015. A study of the effect of OHSAS 18001 on the

occupational injury rate in Iran. Int. J. Injury Control Safety Promotion 24 (1),

78–83.

Hasle, P., Madsen, C.U., Hansen, D., 2021. Integrating operations management and

occupational health and safety: A necessary part of safety science! Saf. Sci. 139,

105247

https://doi.org/10.1016/j.ssci.2021.105247.

Heras-Saizarbitoria, I. (Ed.), 2018. ISO 9001, ISO 14001, and new management

standards, Measuring operations performance. Springer, Cham.

Heras-Saizarbitoria, I., Boiral, O., Arana, G., Allur, E., 2019. OHSAS 18001 certification

and work accidents: Shedding light on the connection. J. Saf. Res. 68, 33–40.

https://doi.org/10.1016/j.jsr.2018.11.003.

Heras-Saizarbitoria, I., Cilleruelo, E., Zamanillo, I., 2013. Adoption of ISO 9000 Practices

in Manufacturing Companies: The Perspective of the Shop-Floor Workers: Adoption

of ISO 9000 in Manufacturing Companies. Hum. Factors Man. 23, 311–321.

https://

doi.org/10.1002/hfm.20312.

Hohnen, P., Hasle, P., 2011. Making work environment auditable–A ‘critical case’ study

of certified occupational health and safety management systems in Denmark. Saf.

Sci. 49, 1022–1029.

https://doi.org/10.1016/j.ssci.2010.12.005.

Hohnen, P., Hasle, P., Jespersen, A.H., Madsen, C.U., 2014. Hard Work in Soft

Regulation: A Discussion of the Social Mechanisms in OHS Management Standards

and Possible Dilemmas in the Regulation of Psychosocial Work Environment. Nordic

J. Working Life Stud. 4, 13–30.

Jespersen, A.H., Hohnen, P., Hasle, P., 2016. Internal audits of psychosocial risks at

workplaces with certified OHS management systems. Saf. Sci. 84, 201–209.

https://

doi.org/10.1016/j.ssci.2015.12.013.

Joliffe, I., Morgan, B., 1992. Principal component analysis and exploratory factor

analysis. Stat Methods Med Res 1, 69–95.

https://doi.org/10.1177/

096228029200100105.

Kauppila, O., Harkonnen, J., Vayrynen, S., 2015. INTEGRATED HSEQ MANAGEMENT

SYSTEMS: DEVELOPMENTS AND TRENDS. Int. J. Quality Res. 12.

Lafuente, E., Abad, J., 2018. Analysis of the relationship between the adoption of the

OHSAS 18001 and business performance in different organizational contexts. Saf.

Sci. 103, 12–22.

https://doi.org/10.1016/j.ssci.2017.11.002.

Lo, C.K.Y., Pagell, M., Fan, D., Wiengarten, F., Yeung, A.C.L., 2014. OHSAS 18001

certification and operating performance: The role of complexity and coupling.

J. Oper. Manage. 32, 268–280.

https://doi.org/10.1016/j.jom.2014.04.004.

Madsen, C.U., Hasle, P., 2017. Commitment or Compliance? Institutional Logics of Work

Environment Management. Nordic J. Working Life Stud. 7, 17–38.

https://doi.org/

10.18291/njwls.v7iS2.96688.

Madsen, C.-U-, Kirkegaard, M.L., Dyreborg, J., Hasle, P., 2020. Making occupational

health and safety management systems ‘work’: A realist review of the OHSAS 18001

standard. Saf. Sci. 129, 104843

https://doi.org/10.1016/j.ssci.2020.104843.

Nielsen, K.T., 2000. Organization Theories Implicit in Various Approaches to OHS

Management. In: Frick, K., Wilthagen, T., Quinlan, M., Langaa Jensen, P. (Eds.),

Systematic Occupational Health and Safety Management - Perspectives On An

International Development. Emerald Group Publishing Limited, p. 515.

Osborne, J., Banjanovic, E., 2016. Exploratory factor analysis with SAS. SAS Institute.

Øystein

Saksvik, P., Torvatn, H., Nytrø, K., 2003. Systematic occupational health and

safety work in Norway: a decade of implementation. Saf. Sci. 41, 721–738.

https://

doi.org/10.1016/S0925-7535(02)00020-6.

Pagell, M., Johnston, D., Veltri, A., Klassen, R., Biehl, M., 2014. Is Safe Production an

Oxymoron? Production and Operations Management 23, 1161–1175.

https://doi.

org/10.1111/poms.12100.

Pawson, R., Tilley, N., 1997. Realistic Evaluation. Sage Publications (CA).

Power, M., 1997. The Audit Society: Rituals of Verification. Oxford University Press,

Oxford.

Robson, L.S., Clarke, J.A., Cullen, K., Bielecky, A., Severin, C., Bigelow, P.L., Irvin, E.,

Culyer, A., Mahood, Q., 2007. The effectiveness of occupational health and safety

BEU, Alm.del - 2021-22 - Bilag 267: Orientering om nye forskningsresultater om arbejdsmiljøindsatsen på virksomheder med et certificeret arbejdsmiljøledelsessystem, fra beskæftigelsesministeren

C. Uhrenholdt Madsen et al.

management system interventions: A systematic review. Saf. Sci. 45, 329–353.

https://doi.org/10.1016/j.ssci.2006.07.003.

Rocha, R.S., Granerud, L., 2011. The search for legitimacy and organizational change:

The agency of subordinated actors. Scand. J. Manag. 27, 261–272.

https://doi.org/

10.1016/j.scaman.2011.01.001.

Saksvik, P.Ø., Quinlan, M., 2003. Regulating Systematic Occupational Health and Safety

Management: Comparing the Norwegian and Australian Experience. RI 58, 33–59.

Doi: 10.7202/007368ar.

Thorsen, S.V., Madsen, I.E.H., Flyvholm, M.-A., Hasle, P., 2017. Associations between the

workplace-effort in psychosocial risk management and the employee-rating of the

Safety Science 152 (2022) 105794

psychosocial work environment

–

a multilevel study of 7565 employees in 1013

workplaces. Scand. J. Public Health 45, 463–467.

https://doi.org/10.1177/

1403494817696377.

Vinodkumar, M.N., Bhasi, M., 2011. A study on the impact of management system

certification on safety management. Saf. Sci. 49, 498–507.

https://doi.org/10.1016/

j.ssci.2010.11.009.

Wilkinson, G., Dale, B.G., 1999. Integrated management systems: an examination of the

concept and theory. The TQM Magazine 11, 95–104.

https://doi.org/10.1108/

09544789910257280.