Occupational Medicine 2009;59:454–458
doi:10.1093/occmed/kqp104
Chronic and acute psychological strain in naval
personnel
R. S. Bridger, A. Dew, K. Brasher, K. Munnoch and S. Kilminster
...................................................................................................................................................................................
Background Previous surveys have shown that there is a greater prevalence of psychological strain in Naval personnel than in the general population and have described the main psychosocial stressors associated
with strain.
...................................................................................................................................................................................
Aims
To determine the prevalence of acute strain and of repeated episodes of strain over 6- and 12-month
periods.
...................................................................................................................................................................................
Methods
Six and twelve months after completing a Phase I Work and Well-Being questionnaire, 2596 personnel
were reassessed using a follow-up General Health Questionnaire-12.
...................................................................................................................................................................................
Results
The response rates at 6 and 12 months ranged from 51 to 60%. There was no evidence of response bias
at follow-up. The prevalence of acute strain was 31% at Phase I. After 6 months, approximately half of
strain cases had recovered. Only 10% had strain over the entire period. Change in strain was linked to
change in work role.
...................................................................................................................................................................................
Conclusions Accumulation of strain and recovery occur within 6 months depending on change in work role. Management of strain might best be achieved by management of work demands and deployment length.
Further studies will investigate the rate of accumulation of strain over the course of demanding deployments. Exposure to psychosocial stressors such as effort reward imbalance accounted for much of
the difference between chronic strain sufferers and those with no strain.
...................................................................................................................................................................................
Key words
Acute strain; chronic strain; effort reward imbalance; naval service.
...................................................................................................................................................................................
Introduction
Previous articles [1,2] have described the prevalence of
psychological strain in UK Royal Navy (RN) personnel
and the main psychosocial risk factors such as perceived
effort reward imbalance (ERI). At any time, 30% of the
RN workforce will be experiencing strain in the form of
symptoms of anxiety and depression, as measured by the
General Health Questionnaire (GHQ-12) [3]. Strain can
occur as a result of exposure to occupational psychosocial
stressors and to other stressful life events.
The Health and Safety Executive (HSE) in the United
Kingdom has published management standards for occupational stress prevention which are based on exposureoutcome models such as the Karasek Demand Control
Support model [4,5]. While such initiatives are of value
and provide employers with practical advice for stress management, less is known about how well general models such
as Karasek’s may predict the stresses and strains of working
life in specific occupations. Implementation of the control
measures advocated by the HSE may only be effective if the
model is valid for the occupational group in question.
Much research has been carried out on occupational
stress in a wide range of service professions but fewer longitudinal than cross-sectional studies have been reported
[4]. Working life in the military differs from that in other
occupations because work demands, including the place
of work, change frequently. The Naval Service cohort
study of occupational stress commenced in January
2007. The purpose of the study was to investigate psychological strain in naval personnel over time. By administering the GHQ-12 repeatedly to the same individuals, the
intention was to determine the percentage of individuals
in which acute strain was in evidence and the percentage
experiencing repeated episodes of strain (chronic strain).
Once identified, any differences in work exposures or demographics could be identified, leading to more closely
targeted interventions for particular groups.
Human Factors Department, Institute of Naval Medicine, Crescent Road,
Alverstoke PO12 2DL, UK.
Methods
Correspondence to: R. S. Bridger, Human Factors Department, Institute of Naval
Medicine, Crescent Road, Alverstoke PO12 2DL, UK. Tel: 144 02392 768220;
fax: 144 02392 504823; e-mail: [email protected]
Phases I to III of the cohort study extended from January
to March 2007 until January to March 2008. A ‘Work and
MOD Crown Copyright 2009.
Published with the permission of the Controller of Her Majesty’s Stationery Office
R. S. BRIDGER ET AL.: CHRONIC AND ACUTE PSYCHOLOGICAL STRAIN IN NAVAL PERSONNEL 455
Well-Being Questionnaire’ (WWBQ) described previously [1,2] was circulated to a stratified, random sample
of 5000 personnel during Phase I. In Phases II (June to
August 2007) and III (January to March 2008), Phase I
respondents were reassessed using the GHQ-12 and
asked whether their work role had changed since the last
questionnaire. Space was provided for respondents to describe any changes. The cohort study is covered by ethical
approval obtained in 2006 for the entire project from the
Ministry of Defence Research Ethics Committee. Questionnaires were marked with a respondent ID linked to
employee service number. The key code was stored securely in compliance with the Data Protection Act,
1998 and all completed questionnaires (hard copy) are
being retained in secure storage for 100 years.
Checks for response bias were carried out by splitting
the Phase I database [1] into respondents and nonrespondents to Phase II. Non-respondents were coded
‘0’ and respondents ‘1’ and a logistic regression analysis
was carried out.
Change in strain was calculated by subtracting Phase 1
GHQ-12 symptom score from the Phase II and Phase III
scores. Positive values indicated that strain had increased
and negative values that it had decreased. Strain caseness
was defined using a GHQ-12 symptom score of four or
above. Logistic regression analysis was conducted to compare a subgroup of chronic strain sufferers (n 5 78) with a
strain-free group (n 5 345).
Results
The response rate to Phase II was 51% after correcting for
questionnaires returned because they could not be delivered (e.g. because the recipients’ addresses were
unknown)—1207 of 2596 Phase I respondents returned
complete Phase II questionnaires. Non-respondents to
Phase II were significantly younger (33.0 versus 36.1
years of age) and reported slightly fewer health complaints (0.6 versus 0.7). Small, statistically significant
differences were found between respondents and nonrespondents in coping style, work family conflict and
organizational commitment.
The logistic regression analysis demonstrated that only
age and responses to the item ‘Is your work physically demanding?’ were related to whether or not a participant
responded at Phase II. Younger personnel in physically
demanding jobs were less likely to respond. However,
age explained only 3.6% of the variance in response/
non-response and physical work demand only 1.3%, i.e.
the total variance explained by age and physical work
demand combined was 4.9%. None of the other variables
measured at Phase I explained differences between Phase
II respondents and non-respondents.
At Phase III, 2468 questionnaires were sent (128 Phase
I respondents had left the service in the intervening
12 months) and 1305 were returned which equates to
a 61% response rate (after correction for 315 questionnaires that could not be delivered for any reason).
The prevalence rates for strain were similar in those
responding at the three phases as might be expected in
the absence of response bias (31.5% of respondents
had strain at Phase I, 30% had strain at Phase II and
34% had strain at Phase III). However, the composition
of these groups of respondents with strain changed over
time. Of those responding to both Phase I and Phase II in
the first 6 months of the study, 55% of respondents had
no strain on either occasion, 17% had strain on both occasions, 13% were new cases at Phase II and 15% had recovered from strain by Phase II.
Similarly, for respondents to Phases I and III: 50% had
no strain on either occasion, 35% had strain on one of the
phases, but not both, and 15% had strain on both occasions. Figure 1 shows the distribution of change in strain
symptoms from Phase I to Phase III.
There was a statistically significant effect of change in
work role on change in strain ‘caseness’ between Phases I
and II (x2 5 8.92, d.f. 5 1, P , 0.01, relative risk 5
1.45). Those whose work role had changed were more
likely to change caseness. In all, 21% of personnel made
spontaneous comments about work/life changes. The
most common events associated with an increase in strain
were work–family conflict, high workload related to staffing problems and problems with drafting (e.g. being sent
to work in a different location unexpectedly or in a nonpreferred location). The most common events related to
a reduction in strain were changing jobs, changing location (especially sea to shore) and leave.
In Phase III, the association between caseness change
(in either direction) and work role change was not
Figure 1. Change in the number of strain symptoms: Phase I symptoms
subtracted from Phase III. Positive values indicate a worsening of strain
and vice versa.
456 OCCUPATIONAL MEDICINE
statistically significant. However, when the strain change
group (n 5 256) was split into those who became strain
cases and those who recovered, work role change had
a significant effect (x2 5 29.1, d.f. 5 2, P , 0.001).
The relative risk of becoming a strain case if work role
did not change was 2.0. In other words, personnel whose
work role had not changed over the year were twice as
likely to become strain cases as those whose role had
changed, suggesting that a change in role is generally beneficial.
In all, 791 personnel responded to all three phases and
78 of these were chronic strain cases (GHQ-12 cases on
all three phases). Some 345 of the 791 had no strain over
the period. The prevalence rate of chronic strain was
slightly ,10%.
For subsequent analysis, the sample was divided into
chronic strain suffers (n 5 78) and those without strain
(n 5 345). Logistic regression was then used to compare
the two groups using the demographic and psychosocial
factors measured in the WWBQ at Phase I and statistically significant differences between these groups were
found. Those with chronic strain were more likely to
perceive that reward for high effort was lacking, to be
overcommitted to their work role, to lack commitment
to the service, to report lack of support from leaders
and to experience stressful life events. Serving at sea
and being of lower rank were not associated with chronic
strain. Overall, the model explained almost 50% of the
difference between the two groups (Table 1).
There was a small, but statistically significant difference between the chronic strain rates in males and
females, with females more often suffering chronic strain
(x2 5 4.9, d.f. 5 1, P , 0.05). In all, 12% of 319 females
had chronic strain as opposed to 8% of 472 males. No
differences in chronic strain were found between those
serving at sea or ashore or between officers and ratings
(all respondents).
Table 1. Differences between those with chronic strain (n 5 78)
versus no strain (n 5 345): results of logistic regression analysis
Predictor
variable
ba
d.f.
Cumulative %
variance explained
ERI
Overcommitment
Leader support
Organizational commitment
Negative mood
Stressful life events
0.19*
0.32***
0.18**
0.12**
0.52***
0.19**
1
1
1
1
1
1
19
34
40
45
47
50
*P , 0.05, **P , 0.01, ***P , 0.001. d.f., degrees of freedom.
a
High scores indicate greater negativity, so positive values of b indicate that strain
is greater when there is a lack of organizational commitment and a lack of support
from the leaders.
Discussion
The prevalence of acute psychological strain in Naval personnel at Phases II and III of this study were similar to the
strain prevalence at Phase I and similar to the rate found
in previous surveys in 1999 and 2004. Phases II and III of
the study had response rates of 52 and 61%. Logistic regression analysis of WWBQ data from Phase II respondents and non-respondents indicated that response bias
was minimal.
Large variations in strain levels were found to have occurred over the 12-month period separating Phases I and
III. Only 15% of respondents had strain at both phases
12 months apart, similar to the percentage having strain
at both Phases I and II (6 months apart). This suggests
that strain is an acute reaction to work demands in many
personnel and that many recover from strain episodes
within 6 months. That the prevalence of strain at all three
phases was the same is indicative of new cases replacing
those who had recovered. Change in work role was associated with change in strain. Those without strain and
those with chronic strain were less likely to report
a change in work role over the period. The data indicate
that change in work role can be positive. Personnel whose
work role had not changed over the year were twice as
likely to become strain cases as those whose role had
changed. This finding is partially supported by the findings of Rona et al. [7]. UK armed forces personnel deployed on active service for more than 13 months in
a 3-year period were at increased risk of developing mental health problems although the number of deployments
in the same 3-year period was not consistently related to
mental health problems.
These findings are remarkably similar to those of
Stansfield et al. [8] in the Whitehall Study of over
10 000 civil servants. ERI at Phase I of their study predicted future psychological strain whereas a change in
work role predicted change in strain in the expected direction (less demanding roles were associated with lower
strain and vice versa).
The logistic regression analysis showed large psychological differences between personnel with no strain
and chronic strain sufferers (strain on all three phases).
Perceptions of an imbalance between effort and reward
accounted for almost 19% of the variance between these
two groups. This perceived imbalance is only partly a perception of high work demands. The ERI scale in the
WWBQ also measures perceptions of an imbalance between the effort expended, the recognition received
and the promotion opportunities. Thus, chronic strain
results not just from high workload but also from unsatisfactory social conditions (unsatisfactory to the sufferer). It
is well known that chronic strain can foster disease through
its effects on hypothalamic–pituitary–adrenocortical (HPA)
axis function. HPA function is determined by the subjective
response to work demands [6]. In the present research,
R. S. BRIDGER ET AL.: CHRONIC AND ACUTE PSYCHOLOGICAL STRAIN IN NAVAL PERSONNEL 457
this response (strain) is measured using the GHQ-12.
The adverse reaction to job stress measured by the
GHQ-12 questionnaire is the psychological component
of the broad-based stress response, sometimes
described as the ‘fight or flight reflex’. Personnel experiencing chronic strain would be expected to have elevated
levels of the so-called ‘stress hormones’, cortisol and epinephrine [10], raised blood sugar levels, blood pressure
and suffer psychomatic complaints such as back pain
and indigestion and have metabolic and haemostatic risk
factors for heart disease [11,12]. One possible explanation for the finding that ERI accounts for a large percentage of the variance in chronic strain is that raised cortisol
levels are thought to influence the way threats are perceived in the environment [13]. Chronic strain sufferers
tend to have a heightened expectation of and attention to
threats in the environment and to exhibit signs of social
dominance and hostility [13]. This tendency to misperceive the behaviour of others and to see other people
as a threat may explain why the main difference between
chronic strain sufferers and personnel without strain is
that the former perceive that their efforts are not recognized by others. This may be a misperception of the true
state of affairs.
The main weakness of the present study is that change
in work role was not investigated in detail—participants
noted only whether their role had changed over the previous 6 months and made spontaneous comments about
the changes that had occurred. In the next phase of the
study, participants will be asked to state the date when
a change in work role last occurred and more details of
the changes that had occurred, to provide a better understanding of the rates of accumulation of strain and recovery. The response rates to Phases II and III of the survey
were low to moderate at 51–60%. Clearly, this is also
a weakness of the study. However, the conclusions regarding change in strain prevalence over time are likely to be
robust because any non-response bias appears, from the
logistic regression analysis, to be due to age and not due to
strain caseness (i.e. strain sufferers and non-sufferers
were equally likely to respond to subsequent GHQ-12
questionnaires).
Much of the research on occupational stress in
military personnel has focussed on the role of psychosocial stressors measured using self-report methods. In the
RN (and very likely in other professions as well), work
roles and job demands change over time. The present
findings suggest the need to shift the focus of the current
research from measuring psychosocial risk factors for
strain to measuring work exposure factors, such as
deployment length and length of recovery periods in
relation to strain. The rate of accumulation of strain,
and that of general feelings of tiredness, will be investigated in the next phase so that the workplace dynamics
underlying the psychosocial risk factors can be better
understood.
Our findings to date indicate that chronic psychological strain is far less prevalent in RN personnel than acute
strain. At any one time, 30% of personnel will be suffering strain. For many of these, the strain is a temporary
reaction to work demands, which resolves within a few
months, as circumstances change. Less than 10% experiencing strain at Phase I still had strain 12 months later
(chronic strain). Chronic strain sufferers differed greatly
from non-sufferers according to the predictions of the
ERI model [14]. To understand best the occupational exposures that cause strain in RN personnel, the focus of
future studies will shift from measuring psychosocial risk
factors for strain to measuring work exposure factors,
such as deployment length and length of recovery periods
in relation to strain. Manning systems that rotate personnel rapidly through demanding sea and shore deployments may be of value in preventing chronic strain
from occurring (as long as these systems are not stressful
in other ways). Understanding the rate of accumulation
of strain would be useful for determining the optimum
length of deployments and recovery periods for demanding work.
Key points
• Approximately 30% of Royal Navy personnel have
job strain at any time, but the strain persisted for
12 months in less than one third of cases.
• Both recovery from and accumulation of strain occur quite rapidly when work role changes.
• Although psychosocial factors accounted for much
of the variance when chronic strain sufferers were
compared to strain-free individuals, further progress in understanding strain in Royal Navy personnel will depend on the measurement of work
exposure factors such as deployment length.
Conflicts of interest
None declared.
References
1. Bridger RS, Kilminster S, Slaven G. Occupational stress
and strain in the Naval service: 1999 and 2004. Occup
Med (Lond) 2007;57:92–97.
2. Bridger RS, Brasher K, Dew A, Kilminster S. Occupational
stress and strain in the Naval service 2007. Occup Med
(Lond) 2007;58:534–539.
3. Goldberg D, Williams P. A User’s Guide to the General Health
Questionnaire. London: NFER Nelson, 1988.
4. Mackay CJ, Cousins R, Kelly PJ, Lee S, McCaig RH.
Management standards and work-related stress in the UK:
policy background and science. Work Stress 2004;18:91–112.
458 OCCUPATIONAL MEDICINE
5. Karasek RA, Theorell T. Healthy Work: Stress, Productivity
and the Reconstruction of Working Life. New York: Basic
Books, 1990.
6. Dollard MF, Winefield AH, Winefield HR. Occupational
Stress in the Service Professions. London: Taylor and Francis,
2003.
7. Rona RJ, Fear NT, Hull L et al. Mental health consequences
of overstretch in the UK armed forces: first phase of a cohort
study. Br Med J 2007;335:603.
8. Stansfield SA, Fuhrer R, Shipley MJ, Marmot MG. Work
characteristics predict psychiatric disorder: prospective results from the Whitehall II Study. Occup Environ Med
1999;56:302–307.
9. Miller GE, Chen E, Zhou ES. If it goes up, must it
come down? Chronic stress and the hypothalamic-
10.
11.
12.
13.
14.
pituitary-adrenocortical axis in humans. Psychol Bull 2007;
133:25–45.
Lundberg U. Methods and applications of stress research.
Technol Health Care 1995;3:3–9.
Vrijkotte TGM, van Doornen LJP, de Geus EJC. Work
stress and metabolic and hemostatic risk factors. Psychosom
Med 1999;61:796–805.
Bosma H, Peter R, Siegrist J, Marmot M. Two alternative
job stress models and the risk of coronary heart disease. Am
J Public Health 1998;88:68–75.
Wargo E. Understanding the have nots. Observer.
2007;20:18–23. Association for Psychological Science.
Siegrist J. Adverse health effects of high effort/low
reward conditions. J Occup Health Psychology 1996;1:
27–41.