Occupational Medicine 2003;53:256–264
DOI: 10.1093/occmed/kqg061
Stress in police officers: a study of the origins,
prevalence and severity of stress-related
symptoms within a county police force
P. A. Collins and A. C. C. Gibbs
Background High levels of stress-related illness are causing concern across industry. Against a
background of impending legislative moves to try to improve this situation, there is a
need to identify key work-related stressors. Police work tends to be regarded as
inherently stressful because of the personal risk of exposure to confrontation and
violence and the day-to-day involvement in a variety of traumatic incidents. As a
result, high levels of stress-related symptoms might be expected in this population.
Objective
To examine the sources of stress-related symptoms within police officers and
measure the prevalence of significant associated mental ill-health.
Method
A cross-sectional questionnaire survey of a population of 1206 police officers was
performed to assess levels of strain associated with a series of potential home and
work related stressors. Participants were then split into low and high scoring groups
on the basis of a General Health Questionnaire (GHQ) threshold score in order to
identify those stressors most associated with mental ill-health effects.
Results
Occupational stressors ranking most highly within the population were not specific
to policing, but to organizational issues such as the demands of work impinging
upon home life, lack of consultation and communication, lack of control over
workload, inadequate support and excess workload in general. The high scoring
group constituted 41% of the population and differed significantly from those with
low scores in perception of all stressors, ranking both personal and occupational
stressors more highly, and from personality constraints appeared significantly more
‘stress-prone’. A significant association between gender and mental ill-health was
found, with females more likely to score more highly on the GHQ than males.
Conclusion
This study confirms previous findings of organizational culture and workload as the
key issues in officer stress. Given that the degree of symptomatology appears to be
worsening, management action is required. Further research is indicated within the
police population into a possible increased susceptibility in female officers.
Key words
Gender; mental health; occupation; police; shifts; stress; workload.
Received
23 September 2002
Revised
4 March 2003
Accepted
12 March 2003
& Safety Executive (HSE) [1] suggests that 20% of
workers feel ‘very’ or ‘extremely’ stressed at work, with
percentages rising to 40% in some occupational groups.
The current study is based on police officers, whose
occupation is generally perceived as highly stressful,
caught between the increasing threat of violence on our
streets, high public demand and a mounting focus on
police efficiency and probity. Policing is amongst the top
Introduction
There is a growing preoccupation with stress as a
problem within the workplace. Evidence from the Health
BMI Health Services, Exeter, UK.
Correspondence to: Pamela Collins, BMI Health Services, Bickleigh House,
Park 5, Harrier Way, Exeter EX2 7HU, UK. E-mail: [email protected]
Occupational Medicine, Vol. 53 No. 4,
© Society of Occupational Medicine; all rights reserved
256
P. A. COLLINS AND A. C. C. GIBBS: STRESS IN POLICE OFFICERS 257
three occupations most commonly reported by both
occupational physicians and psychiatrists in the Occupational Disease Intelligence Network (ODIN) system
for Surveillance of Occupational Stress and Mental
Illness (SOSMI) [2]. In other countries, levels of suicide
several times higher than in age-matched populations
are being reported among police officers [3,4]. When
measured in terms of individual morbidity and mortality,
the cost of stress-related illness in police officers would
therefore appear to be high and financially this translates into an increasing burden in the form of reduced
productivity, sickness absence and early retirement [5].
For police forces, 26% of medical retirement is due to
psychological ill-health [6] and this cost is borne by
the working budget, impacting directly upon service
provision. Like the HSE, both Her Majesty’s Inspectorate
of Constabulary (HMIC) and the Association of Chief
Police Officers (ACPO) are pushing police forces to
reduce sickness absence and ill-health retirement resulting from stress-induced illness.
Much work has been done to try to identify the issues
most associated with police stress. The greater part of this
work stems from the USA and is not necessarily directly
comparable to this country. However, an interesting and
perhaps surprising finding within the US data is that it
is not operational aspects, such as the risk of violence
or exposure to traumatic events, which are perceived as
most stressful—but, rather, organizational issues, such as
managerial structure and climate [7,8]. A major difficulty
in any study of stress is that it is based upon individual
perception—if the same issues are viewed very differently
purely as a result of differing personal interpretation, this
will inevitably make objective assessment difficult. Most
studies of police stress exhibit this problem and measure
perceived strain rather than taking an objective measure
of mental health effects. However, two British studies
in particular are notable [9,10], both performed some
10 years ago, which did take a quantitative measure
using recognized mental health scoring systems. Their
findings show 17–22% of officers reaching significant
levels of mental ill-health—associated particularly with
organizational issues.
In the intervening decade, a more violent and blameorientated culture appears to have developed within our
society, which may have increased the frequency and
intensity of stressful operational exposures. Police forces
have also changed, by introducing measures such as
employee assistance programmes, improved recruitment
selection, better performance recognition, equal opportunities and sexual equality training to try to mitigate the
effects of the organizational stress–strain reaction.
Ten years on, this study reviews the evidence, to
determine which factors within policing are currently
most associated with a high risk of mental ill-health.
Method
A cross-sectional survey was conducted by administering
a postal questionnaire to 1206 constables and sergeants
within a county police force. These ranks were selected
for the survey on the basis both of their predominance
within the organization, constituting 85% of the force,
and indications from previous studies [9,10] of increased
liability to strain. The survey included all officers within
the force holding these ranks, spread across the geographical sectors of the county, each of which constitutes
a separate managerial division of the workforce.
The survey was preceded by a pilot study using
senior officers to test both questionnaire design and the
delivery system. Draft copies were also circulated to the
Police Federation, Welfare Department and Human
Resources for comment and feedback, which was duly
noted and acted upon.
The questionnaire was designed to assess all aspects of
the stress–strain cycle and, in addition to demographic
information, collected details of:
grading of the severity of perceived occupational
· stress
from equal numbers of organizational (relating
·
to workload and work climate) and operational
(aspects of front-line duty) issues, which were drawn
from previous studies [9,10]; the list of stressors so
generated can be seen in Tables 2 and 3;
grading of the severity of perceived personal life stress
constructed from the Holmes and Rahe life events
scale [11]; these included relationship problems,
problems relating to children or to family illness/
bereavement, financial worries and moving house.
A five-point grading scale was used to link participants’
verbal descriptions of perceived stress to a numerical
scoring system with responses of ‘not at all’, ‘slightly’,
‘fairly’, ‘considerably’, or ‘extremely’ stressful being given
scores of 1–5, respectively.
Exposure patterns were measured by entering ‘nonapplicable’ in place of a perceived stress rating for
stressors which were either not inherent in the work role
or were inappropriate to personal home circumstances.
This was felt to be important to avoid any reporting bias
from differing exposure, particularly in specific police
roles.
Details were also collected of:
constraints—personality traits as a measure of
· ‘stress-proneness’
taken from the Eysenck personality
·
inventory [12], a widely used measure of self-reported
personality assessment (as the complete inventory
would have been too lengthy to incorporate into the
questionnaire, a selection of questions was included to
allow for a form of assessment of individual susceptibility to stress);
moderators—factors which may buffer individual
258 OCCUPATIONAL MEDICINE
·
·
reactions to stressor exposure, including home support, support from colleagues, workload management
techniques such as delegation and limitation of work
and exercise activity; data were also gathered on
maladaptive behaviour patterns such as smoking and
alcohol consumption;
details of shift patterns, both as potential stressors in
their own right and as potential confounders of other
issues; officers were asked to state whether they worked
regular hours with no shifts, early and late shifts but no
night duty, full rotating shifts including night duty, or
to give details of any other shift pattern;
individual feelings about continuing a career within
policing as an index of job satisfaction and enjoyment.
The General Health Questionnaire (GHQ 12) was
then used to identify two groups: a high scoring group
with a threshold of 3 or more, with the remainder
constituting a low scoring comparison group. Goldberg’s
terminology of describing the high scoring group as
‘cases’ and the low scorers as ‘non-cases’ is used for
simplicity throughout the study. It is, however, important
to emphasize that the term ‘case’ is used simply to define
a statistically significant probability of psychological
morbidity at the point of study [13,14], rather than as a
diagnostic or epidemiological term.
Details of awareness of stress-related symptoms and
resulting usage of primary care facilities and prescribed
medication were also gathered, together with data on
sickness absence, as a means of confirming the validity of
the GHQ threshold score within the population.
The study was anonymous in order to encourage
participation and frank completion. Numbers not
responding in terms of gender, rank and division were
calculated by subtraction of responders from the total
target population. Reminder notices were delivered at 2
and 4 weeks after the questionnaires were dispatched.
Statistical analysis
An initial analysis of data relating to the whole population
was made, followed by comparison between the two
groups created by the GHQ threshold. The Mann–
Whitney non-parametric test was used to establish the
degree of difference between the two groups in their
responses to the graded answers given to enquiry about
stressors and moderators and for answers of an ordinal
nature in the questionnaire. The χ2 test, or Fisher’s test
where appropriate, was used for answers that were
categorical in nature from other sections of the
questionnaire. Forward stepwise logistic regression was
used to select those stressors that combined significantly
to predict whether a subject was a case. Comparisons
between the groups were based on statistical significance
tests. This means that causal relationships cannot necessarily be inferred.
Results
General results from analysis of the total
population
Demographics
Of the 1206 questionnaires sent out, 873 were returned,
giving a response rate of 72%. The ratios of police
constables to sergeants in the population and of male to
female officers were both in the region of 5:1.
Mental health
The range of scoring on the GHQ 12 is shown in
Figure 1, with 41% deemed to be ‘cases’ as defined by the
study [14]. This percentage was higher than anticipated
from previous studies [9,10]. However, the GHQ 12
threshold was that recommended by the author [14] and
its validity is supported not only by previous studies [15],
but also by the significant difference (P < 0.05) between
the high and low scoring groups in their responses to all
other questions relating to mental health.
Response pattern
Constables were less likely to take part in the survey than
sergeants, bringing the ratio down from 5:1 in the
population to 4:1 in the response group. Female officers
had a higher response rate at 81%, compared with 70%
in males. This differential response rate may introduce a
degree of bias in the higher proportion of female cases,
with a ratio of male to female cases of 3:1, compared with
the population ratio of 5:1 male to female officers. The
response rates from the different geographical divisions
within the organization varied from 67 to 75%. Details of
responders and subsequent analysis of the differences in
responses from cases and non-cases are given in Table 1.
Figure 1. GHQ scores for the total population.
P. A. COLLINS AND A. C. C. GIBBS: STRESS IN POLICE OFFICERS 259
Stressors
Organizational stressors were perceived as more stressful
than operational issues by the population as a whole.
Nine organizational stressors had a median score of 3
(fairly stressful), whereas no operational stressor had a
median score >2 (slightly stressful). Issues ranking most
highly within the population centred upon workload
and time constraints and are identified in Table 2 (see
footnote to table).
Comparisons between groups
Demographic differences
Cases tended to be slightly, but not significantly, older
(P = 0.14), with proportionately more cases found in the
40–49 year age group. However, there was no increase in
caseness in the >50 year age group.
Female gender appeared to be associated with caseness
(P = 0.03) and there were significant differences in
marital status (P = 0.0001), with cases more likely to be
divorced or separated. Whilst being unsupported in this
way was associated with caseness, being single was not.
No difference was found between the two groups in
relation to parenthood, either in number or age of
children. There was no evidence of association between
either rank or length of service and caseness. The number
of cases differed significantly between the managerial
divisions within the organization (P = 0.03). This finding
could not be explained by differences in the response rate,
age, rank, or gender pattern and may relate to differing
management styles in the various divisions.
Stressor analysis
As shown in Table 2, cases found all organizational
stressors significantly more stressful than non-cases
(P ≤ 0.005), except for lack of work, which was an issue to
neither group, with a median score for both groups of 1
(not at all stressful). All operational stressors were also
perceived as more stressful by cases than non-cases at the
conventional level of significance (P ≤ 0.05), as shown in
Table 3.
Whilst the median values marked in Tables 2 and 3
show clearly the greater perception of stress associated
with organizational issues than operational issues and give
some indication of the relative significance of particular
stressors, no clear individual ranking order for those most
associated with measurable ill-health is possible from
these data. This was examined firstly by comparison of the
individual Z-value given by the statistical test. It is notable
that the top 13 stressors derived by this means are all
organizational issues and are shown in descending order
of significance in Table 2.
Secondly, in order to determine which combination
of stressors best predicted which individuals were cases
Table 1. Basic data (percentage) for cases and non-cases
Non-case
Case
Total
Years of service
<2
2–4
5–9
10–19
20–29
30+
Total
Mann–Whitney test
37 (7)
65 (13)
73 (14)
183 (36)
150 (29)
4 (1)
512
Z = –1.61
15 (4)
41 (11)
44 (12)
142 (40)
113 (32)
2 (1)
357
P = 0.11
52 (6)
106 (12)
117 (13)
325 (37)
263 (30)
6 (1)
869
Rank
Constable
Sergeant
Total
Fisher’s test
408 (80)
104 (20)
512
P = 0.67
289 (81)
68 (19)
357
697 (80)
172 (20)
869
69 (19)
64 (18)
52 (15)
55 (16)
54 (15)
41 (12)
13 (4)
6 (2)
354
P = 0.03
141 (18)
139 (16)
114 (13)
179 (21)
135 (16)
109 (13)
32 (4)
11 (1)
860
272 (76)
85 (24)
357
693 (80)
176 (20)
869
Division
Eastern
72 (14)
Western
75 (15)
Northern
62 (12)
Southern
124 (25)
Traffic & Operations
81 (16)
Criminal Investigation 68 (13)
Criminal Justice
19 (4)
Headquarters
5 (1)
Total
506
df = 7
χ2 = 15.33
Sex
Male
Female
Total
Fisher’s test
421 (82)
91 (18)
512
P = 0.03
n = 873. Where one or more variable is missing they have been omitted from
the table.
and which were not, a forward stepwise logistic regression
was carried out, using the numerical answers to the
stressor questions as possible predictors. The seven
stressors that were significantly predictive of being a case
are given in Table 4 in order of importance. Five of
the seven are organizational stressors, including the two
most significant, which stand out very clearly in the
marked degree of statistical significance: ‘demands of
work impinging on home’ and ‘not enough support from
superior officers’.
Despite only reaching a median score within cases of
‘slightly stressful’, two operational stressors were found to
be of significant predictive importance: dealing with a
drunk and feeling at risk of exposure to HIV/hepatitis B
infection.
Life events
Eleven of the 12 life stressors were perceived as
significantly more stressful by cases than non-cases
(P < 0.05), despite a lack of significant difference in the
levels of exposure to life stressors between the two
groups.
260 OCCUPATIONAL MEDICINE
Table 2. Comparison of organizational stressor scores between cases and non-cases
Median score
Organizational stressor
Cases
Non-cases
Zb
P
Demands of work impinging on homea
Lack of consultation/communicationa
Not enough support from senior officersa
Working long hours
Pressure to get resultsa
Urgent requests preventing completion of planned worka
Not enough control over work
Deadlines/time pressuresa
Too much worka
Paperworka
Subject to complaints investigation
Working shiftsa
Working unpredictable hours
Not enough support from fellow officers
Recalled when off duty
Slow career progression
Not enough scope for initiative
Uncertainty about house move
Not enough work
4
3
4
3
3
3
3
4
3
4
2
3
3
2
2
2
2
1
1
2
2
2
2
2
2
1
3
3
3
2
2
2
1
1
1
1
1
1
–9.5
–9.1
–8.9
–8.6
–8.2
–8.2
–8.2
–8.1
–7.9
–7.6
–7.1
–7.1
–7.1
–5.8
–5.4
–5.3
–4.3
–2.8
–1.1
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
0.00001
0.005
0.28
a
Overall population median score = 3 (‘fairly stressful’).
b
Z-value from the Mann–Whitney test.
Table 3. Comparison of operational stressor scores between cases and non-cases
Median score
Operational stressor
Cases
Non-cases
Za
P
Dealing with someone who is drunk
Verbal aggression from the public
Having to use force to restrain
Physical aggression from the public
Answering call for officer assistance
Dealing with a drug addict
High-speed driving
Attending a domestic dispute
Searching for a missing person
Giving evidence in court
Adult victim of violence or abuse
Attending a serious road traffic accident
Being at risk of hepatitis or AIDS
Administering first aid
Interview suspect of serious crime
Crowd control or riot duty
Attending a sudden death
Child victim of violence or abuse
Informing a relative of a death
2
2
2
3
3
2
2
2
1
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
1
1
2
1
2
1
2
2
1
1
2
1
2
2
–6.4
–6.3
–6.1
–5.7
–5.7
–5.6
–5.4
–5.2
–4.4
–4.2
–4.0
–3.8
–3.7
–3.4
–3.2
–2.7
–2.7
–2.3
–1.9
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
<0.000001
0.00001
0.00002
0.00008
0.0002
0.0002
0.0007
0.001
0.006
0.008
0.02
0.05
a
Z-value from the Mann–Whitney test.
Personality constraints
Whilst it should be noted that using an extract of the
Eysenck inventory would undoubtedly have reduced the
validity of personality assessment, analysis of personality
constraints revealed that the two groups have very
different constraint patterns (P ≤ 0.001 for five out of the
six questions). Cases believe themselves to be tense or
nervous individuals by nature, with a tendency to worry
about things in general and in particular about mistakes
or things they feel they should not have done or could
have done better. The only issue where no difference was
found between the groups was in the tendency to stop
and think things over before making a decision, which
was common to all and may relate to the legal exactness
required of police work.
Given the increased proportion of female cases, the
P. A. COLLINS AND A. C. C. GIBBS: STRESS IN POLICE OFFICERS 261
Table 4. Significant stressors predictive of caseness in the logistic
regression
Question
χ
2
Demands of work impinging on home
19.81
Not enough support from senior officers 16.23
Dealing with someone who is drunk
10.99
Subject to a complaints investigation
8.78
Being at risk of hepatitis or AIDS
8.00
Not enough control over work
7.99
Urgent requests preventing planned work 5.36
Table 5. Association between personality traits and gender
Male (%)
df
P
1
1
1
1
1
1
1
<0.00001
0.00006
0.0009
0.003
0.005
0.005
0.02
association between personality constraints and gender
was also examined. However, as can be seen in Table 5,
the evidence for any association is mixed, reaching
statistical significance in only 50% of traits.
Moderating behavioural patterns
The majority of behavioural patterns were significantly
different between the two groups (P ≤ 0.006). Cases
appear to respond to feeling stressed with negative/
withdrawal behaviour patterns of working harder, taking
work home and keeping things to themselves, rather
than taking breaks, delegating to others, or talking to
colleagues. Cases were less likely to use exercise to release
tension and smokers likely to increase their consumption,
with some non-smokers seeming to be prepared to adopt
the habit. Cases were also more likely to take their stress
out on colleagues or the public.
Female (%)
Pa
Would you say you are a tense or nervous person by nature?
Yes
103 (15)
32 (18)
0.30
No
587 (85)
145 (82)
Do you often worry about things you should not have done or said,
or things you feel you could have done better?
Yes
474 (69)
135 (76)
0.05
No
215 (31)
42 (24)
Do you tend to stop to think things over before doing anything?
Yes
592 (86)
143 (83)
0.28
No
95 (14)
30 (17)
Has your mood always tended to go up and down a lot?
Yes
166 (24)
55 (31)
0.07
No
524 (76)
122 (69)
Does it worry you if you know there are mistakes in your work?
Yes
584 (85)
163 (92)
0.01
No
106 (15)
14 (8)
Are you a worrier?
Yes
292 (42)
105 (60)
<0.0001
No
397 (58)
69 (40)
a
Fisher’s test.
Table 6. Comparison between groups—job perception
Job perception
Job perception was significantly more negative for cases
(P < 0.005), with a clear association between wishing to
leave policing altogether in 55% of the high scoring
group. Thirty-five per cent of non-cases also expressed a
willingness to move away from policing.
The findings in relation to job perception are given in
Table 6.
Shift pattern
Although cases scored shiftwork as significantly more
stressful than non-cases, there was no significant
difference between the shift patterns worked by the two
groups. There was therefore no evidence that working full
rotating shifts (including night work), with the potential
disruption of circadian rhythms, carried any increased
association with mental ill-health.
Discussion
Despite societal and organizational change, levels of
stress-related mental ill-health in police officers do not
appear to have improved over the past 10 years. Indeed,
from this study it would appear that proportions of
officers with measurable ill-health have doubled since
that time [9,10] to a level of 41%. This is well above the
levels being noted in most occupational groups, but
comparable with high measurable mental ill-health in
managerial occupations [16] and in line with the highest
levels of perceived occupational stress from the HSE
Bristol study [1].
The study examined the pattern of stressors most
strongly associated with symptoms of mental ill-health
and confirms that the principal issues remain centred
upon work structure and climate. As such, despite
exposure to job-specific hazards, police officers do not
differ from other occupational groups in their perception
of the sources of work-related stress [17,18].
Given that this finding has been reported across
continents for some years, it is interesting to consider why,
despite this knowledge, the level of associated ill-health
appears to have worsened rather than improved. There are
a number of possible reasons for this. It may be either that
262 OCCUPATIONAL MEDICINE
the message is simply not getting through to police
managers, or that it is not being effectively translated into
action. If the message is getting through, it may be that
lack of staff to respond to increasing public demands and
increasing workload is simply negating the effect of any
response measures. Certainly, for the population in
question, the complement of officers has been static over
the past 10 years and has only recently, with political
sanction, been increased. It may also be that response
measures are misdirected; the main focus of stress
reduction in policing over the last decade has tended to
centre upon the provision of welfare support and
employee assistance programmes. Whilst such measures
do appear to be beneficial in reducing perceived strain in
some occupational groups [19] and may, indeed, be
having some effect in this population, the study shows
that further action is needed. In some respects, this is
perhaps predictable since these are secondary interventions designed to moderate the strain response rather
than fundamentally to eliminate or reduce exposure to
stressors. The findings in the current study point to a need
for resources to be directed towards exploring effective
means of modifying the organization of workload and
improving the management climate, rather than into
further medical or welfare interventions.
Although the population size was large and the
response rate was high, the study was performed in a
relatively small county constabulary. It could be argued
that in metropolitan forces, where operational exposures
are likely to be far greater in both frequency and intensity
and undoubtedly exposure to the only two operational
issues that were predictive of caseness more common, the
balance of factors provoking the stress–strain reaction
may differ. However, evidence from the study of inner city
police does not appear to support this. On the contrary,
the findings of a study of Greater Manchester police [20],
revealing the same high ranking for issues such as time
pressures, poor support from senior ranks and too much
paperwork, make it likely that the findings in the current
study should be applicable to other police forces.
A subsidiary aim of the study, using the stress model
identified, was to assess personality constraints and the
effect of moderating factors, providing information about
areas where secondary interventions might be of greatest
benefit. The apparent personality predisposition to caseness is marked, supporting the premise that the two
groups are constitutionally different, either through
inherent or acquired behavioural patterns. Previous
evidence does support personal predisposition. As far
back as 1978, Cherry [21] published data which showed
that those predisposed to anxiety are more likely to report
strain independent of any work factors. Since that time,
much has been written, including extensive work by
Davidson and Cooper [22] and by Payne [23] suggesting
that this may be linked to particular personality types,
notably type ‘A’, who appear more stress-prone. It is
possible that an increased proportion of particular
personality types may enter police employment, both by
self-selection and by recruitment selection. Type A
personalities, for example, are likely to be attractive to
police forces by dint of the other common characteristics
they share—high levels of drive, competitiveness and
achievement. It is also possible that the development of
type A behaviour is positively encouraged by police
culture and that some of these traits may be culturally
acquired. To consider screening out these high-striving
achievers, albeit potentially stress-prone, would not only
be to screen out some of those with the best career
potential, but would automatically exclude some 50% of
the population [24] and is clearly non-viable. The only
practicable means of translating this knowledge into
action is to consider a behaviour modification programme
to address potentially damaging styles, both inherent and
acquired, for which there is some evidence of benefit [24].
Predisposition may also play a part in the finding of a
significant association between female gender and mental
ill-health, for which a number of reasons may be postulated. First, women may be predisposed by personality
type; Davidson and Cooper [24] postulate that working
women differ in personality from their non-working
counterparts and that, whereas type A personality is
generally less common in women than in men, this is not
true of the working population. However, whilst female
officers did differ significantly in their tendency to worry
more, the evidence to support gender-specific personality
predisposition was mixed, reaching statistical significance
in only half of the traits tested. Secondly, women may
report higher levels of distress because they are more
ready to admit to personal difficulties [25]. Certainly, a
higher proportion of females than males participated in
the study and this differential reporting could influence
the association if female cases in particular are more likely
to respond. Thirdly, a confounding variable is present in
that increased numbers of female officers—twice the
percentage of their male colleagues—were divorced or
separated, a status itself associated with caseness. Lastly,
despite the fact that considerable effort has been put into
equal opportunities training by police forces nationwide,
female officers may still be exposed to additional genderspecific stressors in being treated differently both by their
organization and by male colleagues [25]. A gender
association has been noted in other occupational groups,
with higher levels of reported strain noted in female civil
servants in the Whitehall II study [26] and higher levels of
suicide recently in female doctors [27]. Previous police
studies have often failed conclusively to address gender
patterns because of the considerable male predominance
in study populations. However, current evidence shows
that a proportionately higher number of female officers
are retiring on the grounds of psychological illness than
P. A. COLLINS AND A. C. C. GIBBS: STRESS IN POLICE OFFICERS 263
male officers [6] and it would seem that further study of
female officers is indicated to explore gender issues
further.
Turning to moderators, the significantly more negative
behaviour patterns evidenced by cases in managing their
workload under strain indicate an area where stress
awareness and management training may be able to
impact. Whilst it is always difficult to know whether
behavioural issues such as exercise, smoking and drinking
provide evidence to support a need for health promotion
or merely reflect the pattern of deteriorating well-being, it
may be reasonable to consider life style adjustment as part
of a combined programme. Evidence has been produced
[28,29] to show that regular physical training can both
reduce strain and improve mental well-being; this would
sit well with the need for good general fitness for the
physical demands of policing and the current interest in
routine physical fitness assessment for police officers [30].
Although cases ranked shift work as fairly stressful, the
lack of association between working any particular form
of shift and caseness was interesting; shift work, which
includes night duties and disrupts circadian rhythms,
might have been expected to be associated with strain
effects. Previous studies have shown shift work as an issue
commonly identified by officers as a source of stress
[9,10,31], but, as in this study, there appears to be no firm
evidence of associated mental ill-health. This may, nevertheless, be an area where simple sleep hygiene advice
may be considered to improve tolerance [32] and reduce
subjective effects on well-being.
Other interesting but inconclusive findings are raised
by the study. First, it is interesting to note the absence of
any simple association of increased likelihood of mental
ill-health with increasing age or length of service, as might
have been predicted simply by increased exposure. It is
interesting to note similar findings in the Bristol study [1]
and to speculate that this may be an example of the
‘survivor effect’, with those remaining in service into their
later years being those who are better able to cope, but
this would require further study. Secondly, given that
there were indications of significant inter-divisional differences, which were not explained by non-response bias,
there may be merit in examining further the reasons for
this at local level.
A weakness of the study was the inability specifically to
identify non-responders in order to allow both a second
mailing and a more in-depth analysis of responder bias.
It is possible that non-response was associated with
increased negativity about work issues and may, in itself,
have been associated with caseness, a concern that could
have been addressed in the response to a second mailing.
However, this weakness was acknowledged at the design
stage from the results of the pilot study, which suggested
that any means of identifying those who had participated,
even in complete dissociation from the questionnaire
itself, would be likely to impact on the response rate
because of fear of identification.
It is not surprising that more than half of cases
expressed a desire to leave policing. It is difficult, in the
absence of similar data from other forms of employment,
to judge whether the proportion of non-cases also
expressing this desire is unusual in itself, but it is likely
that the general level of job disillusionment will be of
concern to senior police managers.
References
1. Health & Safety Executive (HSE). The Scale of Occupational
Stress. The Bristol Stress and Health at Work Study, Contract
Research Report 265/2000:212-218. HSE, 2000.
2. Centre for Occupational & Environmental Health.
Occupational Physicians Reporting Activity. Quarterly Report.
Manchester: University of Manchester, March 2000.
3. Schmitdke A, Fricke S, Lester D. Suicide among German
federal and state police officers. Psychol Rep 1999;84:
157–166.
4. McCafferty F, McCafferty E, McCafferty M. Stress and
suicide in police officers: paradigm of occupational stress.
South Med J 1992;85:223–243.
5. Confederation of British Industry (CBI). Focus on Absence.
Absence and Labour Turnover Survey. London: CBI, 1999;
5–9.
6. Her Majesty’s Inspectorate of Constabulary (HMIC). Lost
Time. The Management of Sickness Absence and Medical
Retirement in the Police Service, Thematic Inspection Report.
HMIC, 1997; 59–75.
7. Violanti JM, Aron F. Ranking police stressors. Psychol Rep
1994;75:824–826.
8. Kirkcaldy B, Cooper CL, Ruffalo P. Work stress and
health in a sample of U.S. police. Psychol Rep 1995;76:
700–702.
9. Brown JM, Campbell EA. Sources of occupational stress in
the police. Work Stress 1990;4:305–318.
10. Alexander D, Walker L, Innes G, Irving B. Police Stress at
Work, The Police Foundation Series 1993; 15–38, 74–77.
11. Holmes TH, Rahe RH. The Social Readjustment Rating
Scale. J Psychosom Res 1967;11:213–218.
12. Eysenck H, Eysenck S. Eysenck Personality Inventory.
Sevenoaks: Hodder & Stoughton, 1964; 5–21.
13. Goldberg D. The Detection of Psychiatric Illness by
Questionnaire. Oxford: Oxford University Press, 1972;
87–92.
14. Goldberg D, Williams P. A User’s Guide to the General Health
Questionnaire. Basingstoke: NFER-Nelson, 1991; 3–4.
15. Goldberg D, Gater R, Sartorius N, et al. The validity of two
versions of the GHQ in the WHO study of mental illness in
general health care. Psychol Med 1997;27:191–197.
16. Hobson J, Beach JR. An investigation of the relationship
between psychological health and workload among
managers. Occup Med (Lond) 2001;50:518–522.
17. Cooper CL. Identifying stressors at work: recent research
developments. J Psychosom Res 1983;27:369–376.
18. Creed F. Mental health problems at work. Br Med J
1993;306:1082–1083.
264 OCCUPATIONAL MEDICINE
19. Nadolski JN, Sandonato CE. Evaluation of an employee
assistance program. J Occup Med 1987;29:32–37.
20. Crowe G, Stradling SG. Dimensions of perceived stress in
a British police force. Policing Society 1993;3:137–150.
21. Cherry N. Stress, anxiety and work: a longitudinal study.
J Occup Psychol 1978;51:259–270.
22. Davidson MJ, Cooper CL. A model of occupational stress.
J Occup Med 1981;23:564–574.
23. Payne R. Individual differences in the study of occupational stress. In: Cooper CL, Payne R, eds. Causes, Coping
and Consequences of Stress at Work. Chichester: John Wiley &
Sons, 1988; 209–232.
24. Davidson MJ. Cooper CL. Type A coronary-prone
behaviour in the work environment. J Occup Med
1980;22:375–383.
25. Brown JM, Campbell EA. Special groups in the police
service. In: Stress and Policing: Sources and Strategies.
Chichester: John Wiley & Sons, 1994; 5: 81–90.
26. Fuhrer R, Stansfeld SA, Chemali J, Shipley MJ. Gender,
social relations and mental health: prospective findings
27.
28.
29.
30.
31.
32.
from an occupational cohort (Whitehall II study). Soc Sci
Med 1999;48:77–87.
Hawton K, Clements A, Sakarovitch C, Simkin S, Deeks JJ.
Suicide in doctors: a study of risk according to gender,
seniority and specialty in medical practitioners in England
and Wales, 1979–1995. J Epidemiol Community Health
2001;55:296–300.
Norris R, Carroll D, Cochrane R. The effects of aerobic
and anaerobic training on fitness, blood pressure and
psychological stress and well being. J Psychosom Res
1990;34:367–375.
Kobasa SC, Maddi SR, Puccetti MC. Personality and
exercise as buffers in the stress–illness relationship. J Behav
Med 1982;5:391–404.
Fogg E. Fit to serve. Police Rev 1999;11:16–18.
Ottmann W, Karvonen MJ, Schmidt KH, Knauth P,
Rutenfranz J. Subjective health status of day and
shift-working policemen. Ergonomics 1989;32:847–854.
Glazner LK. Shiftwork: its effect on workers. Assoc Am
Occup Health Nurse J 1991;39:416–421.