Beskæftigelsesudvalget 2022-23 (2. samling)
BEU Alm.del Bilag 67
Offentligt
2665201_0001.png
European Journal of Public Health,
Vol. 32, No. 5, 709–715
ß
The Author(s) 2022. Published by Oxford University Press on behalf of the European Public Health Association.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/),
which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
https://doi.org/10.1093/eurpub/ckac098 Advance Access published on 25 August 2022
.........................................................................................................
Leadership behaviours and health-related early exit
from employment: a prospective cohort study of
55 364 employees
Downloaded from https://academic.oup.com/eurpub/article/32/5/709/6675534 by Faculty of Life Sciences Library user on 21 December 2022
Kathrine Sørensen
1,2
, Jeppe Karl Sørensen
1
, Lars L. Andersen
1
, Julie Eskildsen Bruun
1
,
Paul Maurice Conway
2
, Elisabeth Framke
1,3
, Ida E.H. Madsen
1
, Helena Breth Nielsen
1
,
1,2,4
Mads Nordentoft
1
, Karina G.V. Seeberg
1
and Reiner Rugulies
1
2
3
4
National Research Centre for the Working Environment, Copenhagen, Denmark
Department of Psychology, University of Copenhagen, Copenhagen, Denmark
The Danish Multiple Sclerosis Registry, Copenhagen University Hospital, Copenhagen, Denmark
Department of Public Health, University of Copenhagen, Copenhagen, Denmark
Correspondence:
Kathrine Sørensen, National Research Centre for the Working Environment, Lersø Parkalle 105, DK-2100
´
Copenhagen, Denmark, Tel:
þ4539165200,
e-mail: [email protected]
Background:
Absence of certain leadership behaviours, such as lack of feedback, recognition and involvement in
employee development, has been associated with long-term sickness absence. We tested the hypothesis that
absence of eight specific behaviours predicts health-related early exit from employment, and investigated differ-
ential effects in subgroups to guide future preventive initiatives.
Methods:
Using Cox-proportional hazard mod-
elling, we examined the prospective association between absence of leadership behaviours and health-related
early exit from employment in a sample of 55 364 employees during 4.3 years follow-up. Leadership behaviours
were measured by employee ratings in national surveys from 2012 to 2016. Exit from employment included
disability pension and related measures of health-related early exit, retrieved from a national registry.
Results:
We identified 510 cases of health-related early exit from employment during follow-up. A high level of absence of
leadership behaviours, was associated with an increased risk of exit from employment (hazard ratio: 1.57, 95% CI:
1.31; 1.89). Subgroup analyses showed that the association between absence of leadership behaviours and exit
from employment was similar for women and men and across age groups. The association was stronger for
employees with high level of education than for employees with medium/low education, and the association
was not observed among employees with a prevalent depressive disorder.
Conclusions:
Absence of the eight
leadership behaviours is a risk factor for health-related early exit from employment in the Danish workforce.
More studies are needed to confirm the results.
.........................................................................................................
Introduction
H
ing, is a major challenge in many countries, especially in those
European countries with an ageing society. The old-age dependency
rate is increasing, meaning that for every person in the working-age
group (15–64 years of age), the number of persons aged 65 or older,
who might depend on the working-age population, is rising.
1
Thus, it
has become a major task for public health research to establish risk
factors for health-related early exit from employment that are amen-
able to change, to help identify ways to protect the work-ability of
persons in the working-age group.
Knowledge on the role of psychosocial working conditions in re-
lation to early exit from employment is sparse. A recent systematic
review of the literature on ‘the contribution of psychological, social
and organizational work factors to risk of disability retirement’ con-
cluded that research is mostly limited to a few selected working
conditions, in particular shift work and factors related to the ‘job
strain model’, i.e. the combination of high job demands with low job
control.
2
Other psychosocial working conditions, including
leadership behaviours, have been investigated rarely.
A recent study of the Danish workforce showed that absence of
eight specific leadership behaviours, such as involving employees in
planning of their own work or providing necessary feedback, was
associated with an increased risk of employees’ long-term sickness
absence.
3
One can assume that leadership is also related to exit from
ealth-related early exit from employment, e.g. disability pension-
employment; however, the evidence about such a relationship is
sparse and inconsistent.
Six prospective studies have previously investigated the association
between leadership characteristics and disability pension, a key
measure of health-related exist from employment. Five studies
were conducted with relatively small samples (967–6748 partici-
pants) and provided inconsistent results, with two studies reporting
a significant association
4,5
and three reporting no associations.
6–8
These inconsistent results may be due to lack of statistical power,
as early exit from employment is a relatively rare outcome, requiring
large sample sizes for calculating estimates with acceptable precision.
The only large-scale study (40 554 participants) reported no associ-
ation between leadership quality and risk of disability pension,
9
how-
ever, the measurement of leadership was limited to a four-item scale.
The aim of this study was therefore to examine the association
between leadership behaviours and risk of health-related early exit
from employment, including but not limited to disability pension, in
a large-scale prospective study of the Danish workforce. The large
study sample allowed us to estimate the association between leader-
ship behaviour and work exit with a high level of precision and to
conduct analyses in subgroups. Further, we were able to measure
leadership comprehensively, assessing leadership behaviours by
means of eight specific leader behaviours that we deemed as sup-
portive or beneficial for the employees. We hypothesized that the
absence of these behaviours would be associated with an increased
risk of health-related early exit from employment. In addition to this
 BEU, Alm.del - 2022-23 (2. samling) - Bilag 67: Orientering om nye forskningsresultater fra NFA
2665201_0002.png
710
European Journal of Public Health
hypothesis, we also explored whether associations between leader-
ship behaviours and risk of health-related exit from employment
differ in subgroups defined by sex, age, educational level and preva-
lence of depressive disorders at baseline.
Textbox 1
The eight items used to measure leadership behav-
iour
Methods
Study design and participants
We conducted a prospective cohort study by merging data on lead-
ership behaviours from the nationwide questionnaire-based Work
Environment and Health in Denmark (WEHD) survey with national
register data on social transfer payments from the Danish Register
for Evaluation of Marginalization (DREAM).
10,11
A detailed descrip-
tion of WEHD is published elsewhere.
12
Respondents of the WEHD waves from 2012, 2014 and 2016 were
initially selected (n
¼
67 407); of these, 62 289 were working at base-
line. If a participant responded to more than one wave of WEHD, the
first response was chosen. We pre-censored 2775 respondents, be-
cause they emigrated (n
¼
489), retired (n
¼
311) or registered with
health-related early exit from employment (n
¼
1623) before the start
of the follow-up, or were censored between filling in the survey and 1
January 2013 (n
¼
341, for the 2012 wave only). Respondents from
the 2014 and 2016 waves were followed from the day they filled in
the questionnaire until 31 January 2019. Respondents from the 2012
wave were followed not from the day they filled in the questionnaire
but from 1 January 2013, because on this day a major political reform
of the Danish disability pension system came into force
13
and we
wanted to ensure that the same disability pension policies applied to
all participants throughout the follow-up period.
We further excluded 1717 participants who had no leader, and
lastly we excluded participants with missing values on key variables
including sex, age, educational level, depressive state and sample type
(n
¼
2433). The final population consisted of 55 364 participants (see
flowchart,
Supplementary figure SA1).
Each item could be answered with: ‘1
¼
Always’, ‘2
¼
Often’,
‘3
¼
Sometimes’, ‘4
¼
Seldom’, ‘5
¼
Never’ or ‘Have no leader’
(individuals with no leader were excluded from the study).
How often -
Downloaded from https://academic.oup.com/eurpub/article/32/5/709/6675534 by Faculty of Life Sciences Library user on 21 December 2022
•
L1: does your immediate leader explain the company’s
•
•
•
•
•
•
•
objectives so you understand what they mean for your
assignments?
L2: do you have sufficient authority in relation to your
responsibility?
L3: does your immediate leader take the time to engage in your
professional development?
L4: does your immediate leader involve you in the planning of
your work?
L5: does your immediate leader give you the necessary
feedback (favourable and critical)?
L6: is your work recognized and appreciated by the
management?
L7: do you get the necessary help and support from your
immediate leader?
L8: can you rely on announcements from the management?
compare this group with the fourth quartile (high absence of the
leadership behaviours). A more detailed description and validation
of the leadership behaviour scale is provided in the
Supplementary
material.
Covariates
As covariates, we included sex, age, educational level, sampling
method, presence of a depressive disorder at baseline and eligibility
for disability pension (i.e. age
40
years, as younger individuals usu-
ally do not get granted a disability pension in Denmark). Sex, age and
educational level were retrieved from population registers.
14,15
Highest educational attained was categorized into low (less than
high school and no vocational education), medium (high school de-
gree or completed vocational education) and high (post-secondary
education and above). As the population in WEHD was sampled
through two different methods,
12
we included a variable that indi-
cates sampling method.
Presence of a depressive disorder at baseline was ascertained with
the Major Depression Inventory (MDI), a comprehensively validated
self-administered rating scale.
16–18
In accordance with recommenda-
tions in the literature, an MDI-score of
21
points was used to
indicate a probable depressive disorder.
17
For the sensitivity analyses, we included additional covariates from
registers job type and occupational industry, both based on standard
classifications from register data,
19,20
immigration status, organiza-
tion type (private vs. public), and also included self-reported senior-
ity of the participant from WEHD.
Health-related early exit from employment
Data on health-related early exit from employment were retrieved
from DREAM.
10,11
Health-related early exit from employment was
assessed using the DREAM codes for (i) disability pension, (ii) dis-
ability pension while working in a light job with limited work
demands, (iii) enrolment in a work-ability assessment program (a
prerequisite of disability pension) or (iv) starting in a special pro-
tected employment scheme for workers with severe health problems.
Leadership behaviour
Leadership behaviours were measured in WEHD by asking the
respondents to rate how often eight specified behaviours of their
closest leader or the management occurred. See
textbox 1
for a list
of included items. The response categories ranged from ‘Always’
(scored as one) to ‘Never’ (scored as five) and were added together
to form a total score. A higher score on the scale thus indicates more
absence of the leadership behaviours considered.
When values were missing on
<4
items, the sum score was calcu-
lated by replacing the missing items with the average of the non-
missing items. We excluded respondents with missing values on
more than four out of the eight items.
We operationalized leadership behaviour as a continuous variable
(one-point increase of the score on the scale), as well as a categorical
variable with four categories based on quartiles of the distribution,
which allowed us to investigate a dose–response relationship.
Furthermore, we made the
post-hoc
decision to also include a dichot-
omous variable. The dichotomous variable was constructed by col-
lapsing the first three quartiles of the leadership behaviours score
(low, medium low and medium high absence) into one group and
Statistical analysis
Using Cox-proportional hazard models with weeks since baseline as
the underlying time scale, we tested the prospective association be-
tween absence of the leadership behaviours and subsequent risk of
health-related early exit from employment. Participants were fol-
lowed until health-related early exit from employment or censored
due to non-health based retirement, emigration, death or end of
follow-up (31 January 2019), whichever of these occurred first.
 BEU, Alm.del - 2022-23 (2. samling) - Bilag 67: Orientering om nye forskningsresultater fra NFA
2665201_0003.png
Leadership behaviours and risk of health-related early exit from employment
711
We analyzed the association between absence of leadership behav-
iour and risk of health-related early exit from employment for the
continuous, four-category and dichotomous measure of leadership
behaviours. We calculated crude associations and associations
adjusted for age, sex, educational level, type of sample and eligibility
for disability pension (Model 1), and additionally for depressive dis-
order at baseline (Model 2). Further, we tested the associations with-
in the above-mentioned subgroups and for the interaction between
subgroup and category of leadership behaviour, defined as deviation
from multiplicativity.
Finally, we conducted a number of sensitivity analyses; one with a
restricted outcome that only included disability pension, a series
where we adjusted for wave of WEHD, job type and industry,
immigrant status, seniority of the employee, organization type and
lastly by analyzing each of the eight items of the leadership scale
separately.
Results
Table 1
shows the characteristics of the study population and the
number of cases per 10 000 person-years. There were slightly more
women (53.1%) than men (46.9%), the mean age was 45.2 years
(standard deviation: 11.2 years), and most participants had a middle
(44.0%) or high (42.9%) level of education. There were 4168 partic-
ipants (7.5%) with an indication of a depressive disorder.
Downloaded from https://academic.oup.com/eurpub/article/32/5/709/6675534 by Faculty of Life Sciences Library user on 21 December 2022
Table 1
Study population characteristics and incidence of health-related early exit from employment
n
% of population
Cases
Person-years
Cases per
10 000 person
years
21.5
24.4
18.2
7.3
18.0
30.9
35.9
23.6
15.1
67.9
17.8
13.7
15.6
18.5
19.5
32.5
–
17.4
37.2
28.5
–
–
–
22.2
24.7
17.2
7.2
13.7
37.7
21.6
24.0
25.7
–
20.0
23.5
21.2
27.1
19.7
26.2
20.2
17.8
21.4
22.4
Total sample
Sex
Women
Men
Age categories
18–29
30–49
50
Educational level
Low
Middle
High
Depressive disorder at baseline
Yes
No
Job type
a
1: Managers
2: Professionals
3: Technicians and associate professionals
4: Clerical support workers
5: Services and sales workers
6: Skilled agricultural, forestry and fishery workers
7: Craft and related trade workers
8: Plant and machine operators and assemblers
9: Elementary occupations
0: Armed forces occupations
Missing
Industry
a
1: Agriculture, forestry and fishing
2: Manufacturing, mining and quarrying and utility services
3: Construction
4: Trade and transport etc.
5: Information and communication
6: Financial and insurance
7: Real estate
8: Other business services
9: Public administration, education and health
10: Arts, entertainment and other services
Missing
Type of organisation
a
Private
Public
Immigration status
a
Danish
Immigrant or descendent of immigrant
Seniority at baseline
a
3
months
3 months to 1 year
1–3 years
3–5 years
5–10 years
>10
years
55 364
29 412
25 952
6129
26 621
22 614
7266
24 353
23 745
4168
51 196
2590
17 670
7674
5000
8380
338
4609
3290
5488
325
0
583
9059
3022
9244
1940
1955
562
4849
22 234
1909
7
31 531
23 831
52 102
3234
1392
4849
8873
7331
11 784
21 073
100.0
53.1
46.9
11.1
48.1
40.8
13.1
44.0
42.9
7.5
92.5
4.7
31.9
13.9
9.0
15.1
0.6
8.3
5.9
9.9
0.6
–
–
16.4
5.5
16.7
3.5
3.5
1.0
8.8
40.2
3.4
–
57.0
43.0
94.1
5.8
2.5
8.8
16.0
13.2
21.3
38.1
510
310
200
20
219
271
109
246
155
119
391
15
120
61
41
119
<5
b
34
49
66
<5
b
–
<5
b
83
33
68
6
12
9
45
231
20
<5
b
269
241
473
37
12
55
78
59
109
195
236 989
127 101
109 887
27 251
121 898
87 840
30 356
104 236
102 396
17 534
219 455
10 978
76 679
32 898
21 024
36 589
1394
19 565
13 171
23 186
1504
–
2357
37 304
13 357
39 542
8306
8739
2388
20 864
96 317
7785
30
134 276
102 707
223 333
13 643
6086
20 959
38 546
33 079
50 861
87 150
a: Covariates are only used in sensitivity analyses shown in
Supplementary material.
b: Due to protection of the individual participants data, number of cases below five cannot be shown for this specific category.
 BEU, Alm.del - 2022-23 (2. samling) - Bilag 67: Orientering om nye forskningsresultater fra NFA
2665201_0004.png
712
European Journal of Public Health
Leadership behaviour and risk of health-related early
exit from employment
During a mean follow-up time of 4.28 (SD 1.8) years, we identified 510
cases of health-related early exit from employment (21.5 per 10 000
person-years) (table
1).
Of the 510 cases, 179, 67 and 264 were due
to disability pension, enrolment in a work-ability assessment program
and special protected employment scheme for workers with severe
health problems, respectively.
Number of cases per 10 000 person-years was higher in women than
men, higher in participants of older age than participants of younger
age, higher in participants with a lower education level than participants
of higher education and higher in participants with a depressive dis-
order than in participants without a depressive disorder.
The risk of health-related early exit from employment in relation to
absence of the leadership behaviours is presented in
table 2.
A one-point
increase on the sum score was associated with a higher risk of health-
related early exit from employment [hazard ratio (HR): 1.03, 95% CI:
1.01; 1.04], after adjustment for age, sex, educational level, type of sam-
ple and eligibility for disability pension (Model 1). After further adjust-
ing for depressive disorder at baseline, the association became
statistically non-significant (HR
¼
1.01, 95% CI: 0.99; 1.02, Model 2).
Using the quartiles of the sum score, we found a higher risk of health-
related early exit from employment among individuals in the high-
quartile group (indicating absence of leadership behaviours) when
compared with those in the low-quartile group, with a HR of 1.57
(95% CI: 1.22; 1.39, Model 1). After further adjustment for depressive
disorder (Model 2), the estimate became statistically non-significant
(HR
¼
1.16; 95% CI: 0.92; 1.48). We found no increased risk among
individuals in the second and third quartile of the sum score.
Using the dichotomized sum score, we found an increased risk of
health-related early exit from employment among individuals in the
fourth quartile with an absence of the leadership behaviours com-
pared to individuals in the other quartile (first, second and third
quartiles combined), with a HR of 1.57 (95% CI: 1.31; 1.89) in
Model 1 and an HR of 1.25 (95% CI: 1.03; 1.51) when adjusting
for depressive disorder in Model 2.
The sensitivity analyses yielded results similar to the main analysis
(Supplementary
table SA1).
The results of the sensitivity analyses
considering each of the eight leadership behaviour item analyzed
separately, are presented in
Supplementary table SA2.
The analysis
showed that all items presented statistically significant associations
with early exit from employment in Model 1.
defined by sex, age, education and depressive disorder at baseline,
including tests for interaction. We found no effect modification, meas-
ured by multiplicative interaction, in terms of sex or the age groups
(below or above 50 years of age). The association between the leader-
ship behaviours and health-related early exit from employment was
stronger in those with a level of high education than in those with a
low or medium level of education (P for interaction 0.06 and 0.07,
respectively). Among those without depressive disorder, there was a
statistically significant association between absence of the leadership
behaviours and risk of health-related early exit from employment (HR:
1.49; 95% CI: 1.21; 1.85), whereas there was no statistically significant
association among those with depressive disorder (HR: 0.77; 95% CI:
0.53; 1.10) (P
<
0.01 for interaction).
Downloaded from https://academic.oup.com/eurpub/article/32/5/709/6675534 by Faculty of Life Sciences Library user on 21 December 2022
Discussion
In this nationwide, prospective cohort study in Denmark, we found
that employees reporting an absence of eight leadership behaviours
from their supervisors were at increased risk of health-related early
exit from employment. The association was similar in women and
men and in those of younger and older age, but tended to be stronger
among those with a higher level of education than among those with
low or medium level of education. Participants with a prevalent de-
pressive disorder at baseline had, as expected, a higher risk of health-
related early exit from employment. However, in this subgroup, ab-
sence of the leadership behaviours did not appear to further increase
the risk of health-related early exit from employment.
In the study, we treated the sum score of leadership behaviours in
three different ways; as continuous score, categorized into quartiles and
dichotomized (highest quartile as indicator of exposure vs. the lower
quartiles). All three operationalizations showed that higher absence of
the leadership behaviours was associated with an increased risk of early
exit from employment before adjustment for prevalent depressive dis-
order at baseline. After adjustment for prevalent depressive disorder at
baseline, the estimates attenuated and statistical significance was lost
for the continuous score and the score categorized in quartiles, whereas
the dichotomized score retained a statistical significant association
with risk of health-related exit from employment. This suggests that
the group that bears the risk is mainly the ‘extreme’ group, wherein the
leadership behaviours is absent to a large extent.
Other studies have investigated the association between leadership and
the risk of disability pension. Our result is different from the only large-
scale study (40 554 participants) that found no association between a
four-item leadership quality scale and the risk of disability pension.
9
The stratum consisting of individuals with depressive disorder
contained 4168 (7.5%) of the participants; however, the group also
contained 119 of 510 cases (23.3%). We assumed that depression
Subgroup analyses
Table 3
shows the association between the dichotomized exposure and
risk of health-related early exit from employment in the subgroups
Table 2
Association between absence of the leadership behaviours and health-related early exit from employment
Crude
n
Person-years
Cases
Cases pr.
10 000 person
years
HR
(95% CI)
Model 1
a
HR
(95% CI)
Model 2
b
HR
(95% CI)
Absence of leadership behaviour
55 364
236 989
Per 1 increase on sum score
c
Absence of the leadership behaviours, in quartiles
Low
15 062
63 975
Medium low
11 980
51 894
Medium high
14 607
63 080
High
13 715
58 040
Absence of the leadership behaviours, dichotomized
Low/medium low/medium high
41 649
178 949
High
13 715
58 040
510
125
87
122
176
334
176
21.5
19.5
16.8
19.3
30.3
18.7
30.3
1.03
1
0.86
0.99
1.55
1
1.62
(1.01; 1.04)
ref.
(0.65; 1.13)
(0.77; 1.28)
(1.24; 1.96)
ref.
(1.35; 1.95)
1.03
1
0.89
1.02
1.53
1
1.57
(1.01; 1.04)
ref.
(0.68; 1.18)
(0.79; 1.31)
(1.22; 1.93)
ref.
(1.31; 1.89)
1.01
1
0.86
0.93
1.16
1
1.25
(0.99; 1.02)
ref.
(0.65; 1.13)
(0.72; 1.19)
(0.92; 1.48)
ref.
(1.03; 1.51)
a: Model 1: adjusted for sex, age, educational level, type of sample and eligibility for disability pension (age 40 or above).
b: Model 2: as Model 1 and further adjusted for depressive disorder at baseline.
c: High sum score is equivalent to a high degree of absence of leadership behaviour.
 BEU, Alm.del - 2022-23 (2. samling) - Bilag 67: Orientering om nye forskningsresultater fra NFA
2665201_0005.png
Leadership behaviours and risk of health-related early exit from employment
713
Table 3
Association between absence of the leadership behaviours (dichotomized) and health-related early exit from employment in
subgroups
Model 1
a
n
Person-years
Cases
Cases per
10 000
person-
years
HR
(95% CI)
P-value
on
interaction
term
Model 2
b
HR
(95% CI)
P-value
on
interaction
term
By sex
Women
Low/medium low/medium high
High
Men
Low/medium low/medium high
High
By age
Age
<50
Low/medium low/medium high
High
Age
50
Low/medium low/medium high
High
By educational level
Low and medium educational level
Low/medium low/medium high
High
High educational level
Low/medium low/medium high
High
By depressive disorder
No depressive disorder
Low/medium low/medium high
High
High depressive disorder
Low/medium low/medium high
High
0.48
22 371
7041
19 278
6674
96 873
30 228
82 076
27 812
210
100
124
76
21.7
33.1
15.1
27.3
1
1.49
1
1.70
ref.
(1.17; 1.89)
ref.
(1.27; 2.26)
0.28
24 789
7961
16 860
5754
113 149
35 999
65 800
22 040
163
76
171
100
14.4
21.1
26.0
45.4
1
1.41
1
1.74
ref.
(1.07; 1.85)
ref.
(1.36; 2.22)
0.06
23 288
8331
18 361
5384
99 395
35 198
79 554
22 842
237
118
97
58
23.8
33.5
12.2
25.4
1
1.41
1
2.03
ref.
(1.13; 1.76)
ref.
(1.47; 2.81)
<0.01
39 624
11 572
2025
2143
170 350
49 104
8599
8935
270
121
64
55
15.8
24.6
74.4
61.6
1
1.49
1
0.77
ref.
(1.21; 1.85)
ref.
(0.53; 1.1)
1
1.14
1
1.53
ref.
(0.91; 1.44)
ref.
(1.09; 2.15)
1
1.16
1
1.37
ref.
(0.87; 1.53)
ref.
(1.06; 1.78)
1
1.18
1
1.35
ref.
(0.92; 1.52)
ref.
(1.01; 1.82)
0.28
Downloaded from https://academic.oup.com/eurpub/article/32/5/709/6675534 by Faculty of Life Sciences Library user on 21 December 2022
0.26
0.07
a: Model 1 is adjusted for sex, age, educational level, type of sample and eligibility for disability pension (age 40 or above) except the
variable used for stratification.
b: Model 2 like Model 1 and further adjusted for depressive disorder.
would be a confounder due to this strong association between de-
pressive disorder at baseline and later early exit from employment
(table
1)
and the possible association between depressive disorder
and the reporting of absence of leadership behaviours. Indeed, the
analyses showed that adjusting for baseline depressive disorder
strongly attenuated the association between absence of the leadership
behaviours and risk of health-related exit from employment (table
2).
However, when we stratified by depressive disorder, instead of
adjusting for depressive disorder (table
3),
we found that the HR
for the association between absence of leadership behaviours and
risk of health-related early exit from employment was 1.49 among
those without a baseline depressive disorder and 0.77 among those
with a baseline depressive disorder. A multiplicative interaction anal-
yses showed that this difference was statistically significant. It is
notable that we found that the association is not only different in
the two strata, but actually points into different directions.
In the stratum consisting of individuals with depressive disorder at
baseline a possible explanation for the lack of an association between
absence of the leadership behaviours and increased risk of health-
related early exist of employment could be that individuals with a
depressive disorder are already at a high risk for health-related early
exit from employment and this pull away from employment may be
so strong that absence of the leadership behaviours that we investi-
gate in this study play an insubstantial role for this group.
Overall, the results suggest that baseline depressive disorder may
act more as an effect modifier than a confounder. However, it cannot
be ruled out that, in some cases, absence of the leadership behaviours
may have contributed to the onset of a depressive disorder, meaning
that a depressive disorder may, at least partly, operate also as a me-
diator for the association between absence of the leadership behav-
iours and risk of health-related early exit from employment. Further
studies with repeated measures of both leadership behaviour and
depressive disorder are needed to examine to what extent a depres-
sive disorder is a confounder, an effect modifier or a mediator in the
association between absence of leadership behaviour and risk of
health-related early exit from employment.
Our results suggest level of education as a further effect modifier,
as the association between absence of the leadership behaviours and
risk of health-related early exit from employment was stronger
among participants with a high level of education than among par-
ticipants with a low or medium level of education. This is in agree-
ment with a previous Danish study showing that high workplace
social capital, a construct that includes measures of leadership be-
haviour, was more strongly associated with a decreased risk of long-
term sickness absence among employees of high occupational grade
than among employees of low occupational grade.
21
A possible ex-
planation could be that a higher level of complexity in jobs requiring
a high level of education also makes the presence of leadership
behaviours of supervisors more important.
Strengths and limitations
A strength of the study is the size of the study population, providing
the necessary statistical power for a detailed analysis of both main
 BEU, Alm.del - 2022-23 (2. samling) - Bilag 67: Orientering om nye forskningsresultater fra NFA
2665201_0006.png
714
European Journal of Public Health
associations and interactions. Furthermore, the study employs a lon-
gitudinal design, with exposure (self-reported) and outcome (regis-
ter-based) variables measured by two different methods, thereby
reducing the risk of common method variance.
A limitation of this study is that we only measured absence of the
leadership behaviours at one point in time, instead of using repeated
measures, and that we measured absence of the leadership behaviour
and depressive disorder with self-reported questionnaires, which
may have led to misclassification. Possible fluctuations in the rating
of leadership behaviours during follow-up are therefore not
accounted for in this study. Leadership behaviour was rated by the
employee, meaning that our measure captured the individual per-
ception of absence of leadership behaviour. Therefore, we lack infor-
mation on leader behaviours as observed by a third-party observer or
on the leaders’ perception of their own behaviours. Further, we do
not know the reasons why leaders displayed an absence of the lead-
ership behaviours. It is possible that some leaders lacked leadership
skills and therefore their leadership behaviours were absent. It is also
possible that absence of leadership behaviours was not due to lack of
leadership skills but to lack of resources in the organization, which
may also affect the work environment as a whole, for instance in
terms of high job demands and low control. Some leaders may have
been overwhelmed with other work tasks and consequently did not
have the time resources to give feedback, recognition and support to
the employees.
Whether different reasons for absence of the leadership behaviours
could have a different impact on risk of health-related early exit from
employment should be examined in further studies.
Another limitation of the study is that we did not include other
work environment factors of the employee, or other indicators of the
relationship between the leader/management and the employee.
03 and 10-2019-03). The funder had no further role in the study
design, the collection, analyses and interpretation of data; the writing
of the manuscript; or the decision to submit the manuscript for
publication.
Conflicts of interest:
None declared.
Downloaded from https://academic.oup.com/eurpub/article/32/5/709/6675534 by Faculty of Life Sciences Library user on 21 December 2022
Key points
•
In this prospective study in the Danish workforce, absence of
specific leadership behaviours predicted an increased risk of
health-related early exit from employment.
•
In a subgroup of employees with prevalent depressive disorder
at baseline, which had a markedly increased risk of health-
related early exit from employment, absence of the leadership
behaviours did not further increase the risk.
•
Whether facilitating more of the leadership behaviours may
reduce risk of health-related early exit from employment
should be investigated in workplace intervention studies.
References
1
Eurostat. Old-Age Dependency Ratio Increasing in the EU. 2020. Available at:
https://ec.europa.eu/eurostat/en/web/products-eurostat-news/-/ddn-20200713-1 (10
January 2022, date last accessed).
Knardahl S, Johannessen HA, Sterud T, et al. The contribution from psychological,
social, and organizational work factors to risk of disability retirement: a systematic
review with meta-analyses. BMC Public Health 2017;17:1–31.
Sørensen JK, Framke E, Clausen T, et al. Leadership quality and risk of long-term
sickness absence among 53,157 employees of the Danish workforce.
J Occup Environ
Med
2020;62:557–65.
Sinokki M, Hinkka K, Ahola K, et al. Social support as a predictor of disability
pension: the Finnish Health 2000 study.
J Occup Environ Med
2010;52:733–9.
Sterud T. Work-related psychosocial and mechanical risk factors for work disability:
a 3-year follow-up study of the general working population in Norway.
Scand J
Work Environ Health
2013;39:468–76.
Støver M, Pape K, Johnsen R, et al. Work environment and disability pension—an
18-year follow-up study in a Norwegian working population.
Scand J Public Health
2013;41:587–96.
Hinkka K, Kuoppala J, Vaananen-Tomppo I, Lamminpaa A. Psychosocial work
¨¨ ¨
¨¨
factors and sick leave, occupational accident, and disability pension: a cohort study
of civil servants.
J Occup Environ Med
2013;55:191–7.
Andersen LL, Villadsen E, Clausen T. Influence of physical and psychosocial
working conditions for the risk of disability pension among healthy female eldercare
workers: prospective cohort.
Scand J Public Health
2020;48:460–7.
Clausen T, Burr H, Borg V. Do psychosocial work conditions predict risk of dis-
ability pensioning? An analysis of register-based outcomes using pooled data on
40,554 observations.
Scand J Public Health
2014;42:377–84.
2
Practical implication
Our results suggest providing a higher amount of the investigated
leadership behaviours may contribute to the prevention of health-
related early exit from employment. As this is an observational study,
such a conclusion needs to be verified in workplace intervention
studies. We did not find evidence that absence of leadership behav-
iour increased the risk of health-related early exit from employment
for employees with a depressive disorder in our study, suggesting
that an intervention based on the leadership behaviours that we
investigated here might not be beneficial in terms of reducing early
exit from employment among this group.
3
4
5
6
Conclusion
In conclusion, this study suggests that absence of leadership behav-
iours is a risk factor of health-related early exit from employment in
the Danish workforce. The risk was similar for women and men and
for younger and older employees, while it was somewhat higher for
employees with a high level of education than for employees with a
medium or low level of education. Employees with a prevalent de-
pressive disorder had a higher risk of health-related early exit from
employment, but a higher level of absence of leadership behaviours
was not associated with a further increase of the risk of health-related
early exit from employment. More studies are needed to confirm the
results, e.g. studies with more comprehensive measures of leadership
behaviours.
7
8
9
10 Burr H, Pedersen J, Hansen JV. Work environment as predictor of long-term
sickness absence: linkage of self-reported DWECS data with the DREAM register.
Scand J Public Health
2011;39:147–52.
11 Hjollund NH, Larsen FB, Andersen JH. Register-based follow-up of social benefits
and other transfer payments: accuracy and degree of completeness in a Danish
interdepartmental administrative database compared with a population-based sur-
vey.
Scand J Public Health
2007;35:497–502.
12 Johnsen NF, Thomsen BL, Hansen JV, et al. Job type and other socio-demographic
factors associated with participation in a national, cross-sectional study of Danish
employees.
BMJ Open
2019;9:e027056.
13 Mathisen J, Jensen N, Bjorner J, et al. Disability pension among persons with
chronic illness: differential impact of a Danish policy reform. European Journal of
Public Health 2021;31(1):186–192.
Supplementary data
Supplementary data
are available at
EURPUB
online.
Funding
The study was supported by grants from the Danish Working
Environment Research Fund (grant numbers: 10-2016-03, 01-2018-
 BEU, Alm.del - 2022-23 (2. samling) - Bilag 67: Orientering om nye forskningsresultater fra NFA
Leadership behaviours and risk of health-related early exit from employment
14 Pedersen CB. The Danish civil registration system.
Scand J Public Health
2011;39:
22–5.
15 Jensen VM, Rasmussen AW. Danish education registers.
Scand J Public Health
2011;
39:91–4.
16 Bech P, Rasmussen NA, Olsen LR, et al. The sensitivity and specificity of the Major
Depression Inventory, using the Present State Examination as the index of diag-
nostic validity.
J Affect Disord
2001;66:159–64.
17 Bech P, Timmerby N, Martiny K, et al. Psychometric evaluation of the Major
Depression Inventory (MDI) as depression severity scale using the LEAD
(Longitudinal Expert Assessment of All Data) as index of validity.
BMC Psychiatry
2015;15:190.
715
18 Olsen LR, Jensen DV, Noerholm V, et al. The internal and external validity of the
Major Depression Inventory in measuring severity of depressive states.
Psychol Med
2003;33:351–6.
19 Torma K, Simbold C, Sørensen A, et al. DB07, Copenhagen: Statistics Denmark,
2007.
20 International Labour Organization. International Standard Classification of
Occupations - ISCO 08. Geneva: International Labour Office, 2018. http:
//www.ilo.org/public/english/bureau/stat/isco/isco08/.
21 Rugulies R, Hasle P, Pejtersen JH, et al. Workplace social capital and risk of long-
term sickness absence. Are associations modified by occupational grade?
Eur J Public
Health
2016;26:328–33.
Downloaded from https://academic.oup.com/eurpub/article/32/5/709/6675534 by Faculty of Life Sciences Library user on 21 December 2022