A psychosocial and biomedical comparison between men in ... pc shelter/ill scans/theorell... · that waiters, the group reporting the most unfavourable working conditions, tended

WORK & STRESS, 1990, VOL. 4, NO. 1, 51--63

Preliminary Findings

A psychosocial and biomedical comparison between men in six contrasting service occupations

TORES THEORELL, GUNNEL AHLBERG-HULTEN, FILlS SIGALA, ALEKSANDER PERSKI and MAUD SODERHOLM

National Institute of Psychosocial Factors and Health, Stockholm, Sweden

ANDERS KALLNER

Department of Clinical Chemistry, Karolinska Hospital, Stockholm, Sweden

PETER ENEROTH

Department of Endocrinology, Research Centre, Huddinge Hospital, Sweden

A study of psychosocial job factors and physiological states was conducted using 150 men, aged between 25 and 60, working in six widely different occupations - freight handlers, aircraft mechanics, air traffic controllers, physicians, waiters and symphony musicians. The aim of the study was to characterize any differences in physiological parameters in relation to differences in psychosocial job characteristics. Analyses of variance were used to test differences between groups. The results indicated that waiters, the group reporting the most unfavourable working conditions, tended to have many cardiovascular risk factors. Physicians, who had relatively few cardiovascular risk factors, reported high decision latitude and many emotional reactions in their daily life and had relatively high cortisol and low prolactin levels in plasma.

Keywords:job stress; Cortisol; Prolactin; Testosterone; Smoking; Blood pressure; Cholesterol; Demands; Skill utilization; Authority over decisions; Emotional states.

1. Introduction Previous research regarding psychosocial work conditions has mostly been performed either as in-depth descriptions of the conditions associated with one occupation or as macroepidemiological descriptions of the conditions of a large number of subjects working in many different occupations. Both these strategies have advantages and drawbacks. However, in order to usefully extend our knowledge, we need in-depth descriptions of all occupations. Since this is an unrealistically vast undertaking, a middle way would be to make in-depth descriptions of a small number of occupations representing widely different work activities. This is what has been attempted in the present study of six service occupations. The occupations that were selected for study differ on a number of important psychosocial dimensions, for instance physically strenuous/non-strenuous, rushed/nonrushed, stimulating/boring and controllable/uncontrollable. Comparisons between subjects

0267-8373/90 $3·00 © 1990 Taylor & Francis Ltd.

T. Theorell et al. 52

in these occupations were made in terms of their physiological state, and their reports on

their work environment, habits and emotional states. Epidemiological studies by our group as well as by others (Bohm-Audorf and Siegrist

1983, Swedish Central Statistical Bureau 1981, Olesen and Sondergaard-Kristensen 1988, Norwegian Central Statistical Bureau 1976) have indicated that morbidity and mortality differ dramatically in different occupations. For instance, waiters and transportation workers, two groups that are represented in the present study, have been shown to have short longevity and high incidence of cardiovascular disease. Physicians (also participating in this study) have been shown to have a low incidence of cardiovascular illness but a high incidence of suicide (Arnetz et al. 1987). National surveys have shown that subjects who work in the two former groups frequently report poor decision latitude in their work (Karasek et al. 1988). The aim of the present study was to expand on these observations.

2. Method

2.1. Subjects Six contrasting occupations were chosen for study. The selection of individual subjects was made from employment lists by the researchers, who had no previous knowledge of the people involved. The samples were stratifted in such a way that for all groups there would be an equal proportion of subjects in the young (25-30), intermediate (4o-45) and older (55-60) age ranges. The proportion of men selected varied between the groups, because in three (air traffic controllers, physicians and waiters) equal numbers of women were included while in the other groups they were not. This paper reports the data collected from the men

in the six occupational groups.

2.2. Occupations Musicians employed by the Swedish Radio Symphony Orchestra. This highly professional orchestra has a contract with the Swedish government which states that a high percentage of its productions should involve contemporary music, which is considered to be technically very difficult but is frequently looked upon by the musicians as meaningless. Group interviews indicated that symphony musicians in general experience unusually high qualitative demands as well as having very little influence over decisions regarding their job

(Theorell et al. 1987). 78% (n = 30) participated at least once in the study. Freight handlers at the Arlanda Airport in Stockholm. Group interviews revealed that this

group has a physically strenuous and periodically hectic job with very little skill utilization

(Theorell et al. 1987). 58% (n=30) participated at least once in the study. Aircraft mechanics at the Arlanda Airport. Group interviews revealed that workers in this

group also have a physically strenuous job but in general experience a high degree of skill utilization. Periodically their job may be extremely hectic. They have immediate responsibility for the lives of other people (Theorell et al. 1987). 59% (n = 30) participated at

least once in the study. Air traffic controllers are a highly selected group of employees. They obtain their licence

after having passed through a special school. However, difficult tests at entry as well as during the course make it difficult for those who are interested in air traffic-control to get a job. Group interviews reveal that air traffic controllers view their work as demanding, responsible and sometimes hectic, and guarded by many regulations which diminish their authority over decisions (Theorell et al. 1987). 66% (n = 20) participated at least once in the

study.

A psychosocial and biomedical study of contrasting occupations 53

Physicians were sampled in~ hospital and in primary-care centres in one area of reater Stoc~~o1m. They w~re all c~tmcally active, that is treating patients. In national s!rve phystctans have descnbed thetr work in Sweden as psycholo · 11 d d . · .ys,

1 1 f k.l] ·1· · giCa Y eman mg wtth a htgh

eve o s 1 ut1 tzatiOn. In general they feel the h d h · · · 1 h h

. Y ave goo aut onty over deCiston-mak· a t oug certam aspects of their work mainly work ho d 'ffi 1 mg ( 20)

· . ' urs, are 1 ICU t to control 53<Y< n = . partiCipated at least once in the study. ·

0

Watters were.sampled in three. different locations in Stockholm and in several kinds of restaurant . In natiOnal surveys walters report th · k b d · · f h . . . , etr wor to e emandmg wtth a low level o aut onty over declSlons and a low level of skill utilization 50<Y< (n- 20) · · d least once in the study. ·

0 - parttetpate at

2.3. Participation Participation rates as well as mean ages of the d'ffi · . h . 1 erent groups are presented m table 1. The mean ages m t e dtfferent groups were quite similar with one exception· air traffi controllers had a lower mean age than the oth Th. · d · tc . a: er groups. ts ts ue to the fact that ver few atr traniC controllers were available in the olde t 0 · · y · s age group. ue to thts dtfference all grou compansons have been adjusted for age As ind· t d · h bl ' p b f . . . . tea e m t e ta es, there are reduced num ers o parttetpants for particular variables. The largest internal drop-o t d £i blood pressure. u occurre or

2.4. Measures Measures. were taken ~pproximately once every third month during a working F each subject and studted variable there could be between one a d fi year. or depending on how often the subjects participated About 80; four ~e~sureme( nhts

P 0

t' · d · o o part1c1pants t e r por ton vane somewhat between the variables) had at 1, t t b 3301 h d 11 eas wo measurements and

a out ; o a a four measurements. Whenever a subiect had more tha 1 J n one measurement an average was ca culated, and this average was used as the subject's value in these analyses.

2.5. Blood pressure Blo~~ pressure was measured by means of self-triggered equipment (Cardiocare 2000) Th

~arttetpants ·~ler)edasked to measure their blood pressure once an hour (or as close to on~e a~ our as poss1 .e urmg an ordmary working day from waking up until going to bed The

a~erage systolic and diastolic blood pressure was calculated for work· h d. fi letsure h A 1 h mg ours an or avera o;~s. t east t ree a~ceptable readings were required for the calculation of an av ge. e m~ans for. workmg hours and for leisure hours have been based u on on th e:age 7·7 readings dunng work and 6·4 readings during leisure per recorded wo~k day· ye:rmean average' for each subject is based upon 1-4 work days distributed over the who!~

Table 1. Number of participants, participation rate and mean age in different occupation groups.

Symphony musicians Air traffic controllers Aircraft mechanics Freight handlers Physicians Waiters

Number of participants

31 20 30 30 20 20

Participation rate(%)

78 66 59 58 53 50

Mean age+SD

41·7±12·8 36·8±8·0 40·4±9·9 44·6± 10·7 41·8±8·7 41·2+9-4

T. Theorell et al. 54

The self-recorded blood pressures were compared with a standard mercury manometer reading performed by a trained nurse immediately after the self-record~ng. For healthy persons the correlation was high (r=0·85) for systolic ~lt~ough not_as htgh (r=0·73) for diastolic blood pressure. The systolic levels were not stgmficantly dtfferent be~ween selfmeasurement and conventional recording, but the diastolic self-monitored readmgs tended to be slightly lower than the conventional ones. The first-day measurement effect ~as comparable across the studied groups. In this report the blood pressur~ averages dunng working hours are presented as well as the systolic blood pressure dtfference between

working hours and leisure.

2.6. Blood chemistry . . Blood specimens were drawn by a registered nurse at the work site when the su?Je~t arnv.ed in the morning. All subjects had been instructed to abstain from eating and drmkmg (wtth the exception of water) for ten hours before blood was drawn. Cholesterol was analysed using a standard enzymatic method and prolactin, cortisol and testosterone by means ?f radioimmunoassay (Diagnostic Products Corporation 1985 a, b, c). Glycated haemoglobm (Hb A

1 C) was measured according to Jeppson et al. (1983) a~d .the liver e_nzyme glut~~yl

transferase (GT) according to the Scandinavian Society for Chmcal Chemtstry and Chmcal Physiology (1976). GT levels tend to be elevated in subjects who drink exce~stvely .• but also reflect other sources ofhepatic disorder, while elevated Hb A 1 C levels occur m subjects who have a long-lasting elevation of blood glucose. Cortisol and prolactin were expressed as logarithms due to the skewed distributions of these varia~les in .the st~tistical analyses but will also be presented in the final results as the correspondmg anttloganthms for means and

upper and lower limits of the standard errors of means.

2.7. Work environment and smoking habits Psychological demands, skill utilization and authority over decisions were m~a~ured using a Swedish version of a questionnaire introduced by Karasek et al. (1988). Posltlve factors at work (social support at work) were analysed by means of a short questionnaire developed for this project. It included six statements regarding 'atmosphere' at work. For each statement there were four response categories ranging from 'Not at all' to 'Completely true'. Factor analysis has shown that the total index has an eigenvalue of 3·48 with eigenvectors ranging from 0·30 to 0·47. The results are expressed as the average item

score (1--4). Smoking habits were recorded by means of two simple questions, namely 'Do you

smoke <;:igarettes?' and if yes 'How many a day?' Five response categories ranging from 0

(none) to 4 (more than 20 cigarettes a day) were given.

2.8. Emotional states Emotional states were assessed by means of an adjective checklist describing a wide variety of moods and asking 'How do you feel right now?' The respondent was asked to note which adjective most adequately described his mood at the moment. A four-point response scale (Q-3) was provided for each mood. Subsequently, subjects' responses were grouped

according to the following system:

Sadness. If the respondent marked at least scale points 2 and/or 3 for adjectives 'sad'

and/or 'depressed', the occasion was operationally defined as 'sad' . . . ,. , Joy . If the respondent marked at least scale points 2 and/or 3 for the adjecuves JOyful

and/or 'happy' the occasion was operationally defined as 'joyful'.


Worry. If the respondent marked scale points 2 and/or 3 for the adjectives 'tense' 'nervous' and/or 'worried' the occasion was operationally defined as 'worried'. '

Anger. If the respondent marked scale points 2 and/or 3 for the adjectives 'angry' and/or 'irritated' the occasion was operationally defined as 'angry'.

These categ~ri~ations were partly based upon previous studies using factor analysis of responses to adjecttve checklists (Sjoberg et al. 1979). For each day of measurement, the percentage of'sad', 'joyful', 'worried' and 'angry' occasions was calculated using the total number of measurement occasions that day as the denominator. Subsequently, the average per~entage of each of these four emotional states was calculated for each participant using all avatlable measurement days for the subject.

2.9. Sleep disturbance ~ sleep disturbance index was used which had been constructed in advance according to Akerstedt and Torsvall (1977). It included four items (scaled Q-3) dealing with difficulties relating to waking-up, nightmares, and tiredness during the day. The data were expressed as summed scores (Q-12). On the morning of the examination, before work started, the participants were asked to rate whet~er they felt rested or tired after the previous night. A five-point response scale was used (Akerstedt and Torsvall 1978).

2.1 0. Statistical methods Subjects' ages were recorded and the effect of age on each of the variables was studied by means of linear regression performed on the whole group.

One-way analysis of covariance was used in order to test the degree to which the variation between groups was larger than the variation between individuals. Since several variables were age-dependent, age was used as a covariate throughout all these analyses.

3. Results

3.1. Age trends

Table 2 shows the results of the analyses of age trends for the different variables. Several physiological variables showed the expected age trends : increasing levels with increasing age, namely cholesterol, triglycerides, glutamyl transferase (GT), HbA 1C and maximal as well as average blood pressure during working hours. The difference between blood pressure at work and during leisure on the other hand did not increase with in~re~sing age. For testosterone, there was a highly significant association: decreasing levels wtth mcreasing age. For cortisol, there was a similar decrease with age although this did not reach conventional statistical significance (p =0·06). With regard to the work environment psychological demands tended to decrease with increasing age although this tendency did not reach statistical significance (p = 0·09). Decision latitude and social support did not change wit~ increasing age. Joy tended to decrease with increasing age (p = 0·09), but the ~the_r emottonal states did not change with age. The sleep disturbance index improved sigruficantly with increasing age as did the index 'Do you feel rested in the morning?'

3.2. Differences between groups

Table 3 shows the results of age-adjusted analyses of covariance for the different factors. ~nly those factors which differed significantly (p = 0·05 or less) and those which tended to d~ffer between the groups (p = 0·1 0) are presented in the table. Those factors which did not ~Iffer between the groups, namely cholesterol, GT, Hb A 1 C, sadness, sleep disturbance and Do you feel rested in the morning?' have not been included in the table.

56 T. Theorell et al.

Table 2. Results of the regression analyses of age and other variables (n = 128-147).

Psychological demands at work (5- 20) Skill utilization (4-16) Authority over decisions (2-8) 'Positive factors' at work (1-4) Sadness(%) Joy(%) Worry(%) Anger(%) Sleep disturbance (o-12) Feel rested (1- 5) Smoking (0-4) Cholesterol (mmoljl) Gamma glutamyl transferase (log mmolfl) Hb A 1 C (%) Cortisol (log moljl) Prolactin (log ng/1) Testosterone (mmoljl) Mean systolic blood pressure at work (mm Hg) Mean diastolic blood pressure at work (mm Hg) Max. systolic blood pressure at work (mm Hg) Max. diastolic blood pressure at work (mm Hg) Difference systolic blood pressure (mm Hg) work- leisure Difference diastolic blood pressure (mm Hg) work- leisure

Regr. coeff. ± SEM p

-0·028±0·016 0·010±0·013 0·007 ± 0·012

0·0060 ± 0·0040 -0·013±0·039 -0·394±0·232 -0·071 ±0·100 - 0·033 ± 0·030 -0·037 ±0·015

0·017 ± 0·007 0·014±0·010 0·046 ± 0·008

0·0067 ± 0·0020 0·0096 ± 0·0042

-0·002±0·001 - 0·003 ± 0·003 -0·198±0·067

0·220±0·102 0-400 ± 0·068 0·214±0·107 0· 395 ± 0·076

-0·013±0·044 0·009 ± 0·032

0·09 0·42 0·56 0·14 0·75 0·09 0-48 0·27 0·02 0·03 0·14 0·001 0·001 0·02 0·06 0·22 0·001 0·03 0·001 0·05 0·001 0·77 0·77

A few clear patterns are visible in the biomedical data. Physicians and air traffic controllers showed relatively low blood pressure means whereas freight handlers and symphony musicians had relatively high levels. The differences in self-measured blood pressure levels cannot be easily translated into conventional blood pressure measurements. However, it has been pointed out that blood pressure measured during work activities has greater relevance to cardiovascular disorders than conventional measurements (Devereux et al. 1983). The observed differences between the groups do correspond to real differences in the risk of developing left ventricular hypertrophy, for instance. The difference between systolic blood pressure at work and during leisure also tended to differ significantly between the groups (p = 0·07). Aircraft mechanics demonstrated the most pronounced differences, possibly indicating a high arousal level at work.

With regard to cortisol and testosterone blood levels, waiters showed relatively low levels. This could be interpreted as evidence of premature ageing since both of these hormones in this study, as well as in other studies (Harenstam 1989), tend to decrease with mcreasing age.

Prolactin, which did not show any significant association with age in this study, differed markedly between the groups. Physicians had relatively low levels. Prolactin and cortisol should both show marked circadian variation. Therefore an additional analysis of covariance was conducted for the first blood sampling occasion using blood sampling time as a covariate. The difference between the groups remained significant and the rank order between them unchanged. The same analysis was conducted using duration of wakefulness (time from waking up until blood sampling) as a covariate, and again the difference between the groups remained significant and the rank order between the groups unchanged.

With regard to factors that may possibly cause premature ageing, namely smoking, a poor work environment and adverse emotional states, waiters did report high psychological

A psychosocial and biomedical study of contrasting occupations

~

~ ~N~~N::! "' "5

00\--0\f'") -o--o-666666 +I +I +I +I +I +I 1'-'-0'-0-.:t-.:too --oooll'l c<>c<>c<>c<>c<>c\l

"<t 0 0\ r-- ll'l r-- J;1 .,......N.,.....NNN ""

~ 0 :: "' Ll)~~"'::t-VV ~

666666 ~ +I +I +I +I +I +I .:: N.,......OI'-'-00\

~:±:±~;js:J ~ "' ..::

.;:o

..:: "' "5

'-C

~

f'"lf'"lf'"lf'"l0'-0 ..a..a..;.,..;.,oo..;.,

+I +I +I +I +I +I 1'-1'-.,......"<ti'-N ..;.,..;.,..:....oo..:....oo "<tf'"lf'"lNN.,......

f'"lf'"lf'"lf'"l"<tf'"l 666666 t +I +I +I +I +I +I ..:: N.,.....f'"lf'"l"<tl'- ~ c<>c<>c\lc\l..:....o--'1""""'4,.......; '1""""'4,.......;,.......;

1'-""<tf'"lNf'"li'-NNNc<>c\Jc\J

+1+1 +I +I +1+1 NN'-01'-0N ...t-c<>c\loo..r,c\J -- .......

f'"lf'"lf'"l"<tf'"lf'"l 666666 +I +I +I +I +I +I -.:t00-0\1'--.:t ..a..;.,..r,...,t....,t.c<>

57

Tabl

e 3

(Con

tinu

ed).

R

esul

ts

of

the

anal

yses

o

f va

rian

ce

com

pari

ng

occu

patio

nal

grou

ps.

All

m

eans

ha

ve

been

ad

just

ed fo

r ag

e (M

± SE

M).

M

SE

M

n

% a

nger

dur

ing

an o

rdin

ary

da

yt

Phy

sici

ans

3·7

2±

0·8

1

19

Wai

ters

3·

31 ±

1·0

2

12

Air

craf

t m

echa

nics

2

·89

±0

·72

24

S

ymph

ony

mus

icia

ns

1·6

8±

0·6

8

27

Fre

ight

han

dler

s 1·

52

±0

·67

28

A

ir t

raff

ic c

ontr

olle

rs

0·7

6±

0·8

0

20

Cig

aret

te s

mok

ing

inde

x (0

--4)

W

aite

rs

1·3

8±

0·3

0

15

Fre

ight

han

dler

s 1

·07

±0

·22

27

P

hysi

cian

s 0

·76

±0

·27

19

S

ymph

ony

mus

icia

ns

0·7

2±

0·2

2

28

Air

tra

ffic

con

trol

lers

0·

51 ±

0·2

6

15

Air

craf

t m

echa

nics

0

·26

±0

·23

25

Mea

n sy

stol

ic b

lood

pre

ssur

e at

wor

k (m

m H

g)

Fre

ight

han

dler

s 1

32

·0±

2·2

27

S

ymph

ony

mus

icia

ns

130·

1 ±

2·2

26

A

ircr

aft

mec

hani

cs

13

0·0

±2

·3

24

Wai

ters

1

26

·7±

3·3

12

A

ir t

raff

ic c

ontr

olle

rs

126·

1 ±

2·6

20

P

hysi

cian

s 1

22

·6±

2·6

19

t F=

2·1

1;

df=

5/1

24

; p

=0

·07.

tF

=2

·32

; d

f=5

/14

3;p

=0

·05

. §

F=

1·9

6;

df=

5/1

22

; p

=0

·09

.

Tabl

e 3

(Con

tinu

ed).

R

esul

ts

of

the

anal

yses

o

f va

rian

ce

com

pari

ng

occu

patio

nal

grou

ps.

All

m

eans

ha

ve

been

ad

just

ed fo

r ag

e (M

± SE

M).

M

SE

M

n

Mea

n di

asto

lic b

lood

pre

ssur

e at

wor

k (m

m H

g)t

F

reig

ht h

andl

ers

85

·8±

1·6

26

S

ym

ph

on

y m

usic

ians

85

·1 ±

1·5

27

A

ircr

aft

mec

hani

cs

82·8

± 1

·6

24

Air

tra

ffic

con

trol

lers

82

·5 ±

2·3

12

W

aite

rs

82·2

±1

·8

20

Phy

sici

ans

78· 7

±1

·8

19

Max

imal

sys

tolic

blo

od p

ress

ure

at w

ork

(mm

Hg

)t

Fre

ight

han

dler

s 9

1·9

±1

·6

27

Sy

mp

ho

ny

mus

icia

ns

90

·8±

1·6

26

A

ir t

raff

ic c

ontr

olle

rs

88

·8±

1·9

20

A

ircr

aft

mec

hani

cs

88·7

± 1

·7

24

Wai

ters

87

·6±

2-4

12

P

hysi

cian

s 8

5·2

±1

·9

20

Diff

eren

ce

betw

een

syst

olic

blo

od p

ress

ure

at w

ork

and

duri

ng

leis

ure

(mm

Hg)

§ A

ircr

aft

mec

hani

cs

Wai

ters

S

ym

ph

on

y m

usic

ians

F

reig

ht h

andl

ers

Air

tra

ffic

con

trol

lers

P

hysi

cian

s

t F=

2·2

0;

df=

5/1

22

; p

=0

·06

. t

F =

2·6

6; d

f = 5

{123

; p

= 0

·03.

§

F=

2·0

7;

df=

5{12

2; p

=0

·07

.

6·9

2±

1·0

3 4

·65

± 1

-46

4·5

7±

0·9

9 4

·13

±0

·98

3

·95

±1

·15

2

·05

±1

·16

24

12

26

27

20

19

Tabl

e 3

(Con

tinu

ed).

R

esul

ts o

f the

ana

lyse

s o

f var

ianc

e co

mpa

ring

occ

upat

iona

l gro

ups.

All

mea

ns h

ave

been

ad

just

ed fo

r ag

e (M

± SE

M).

Mor

ning

pla

sma

cort

isol

log

(m

mol

fl)t

P

hysi

cian

s F

reig

ht h

andl

ers

Air

tra

ffic

con

trol

lers

A

ircr

aft

mec

hani

cs

Sym

phon

y m

usic

ians

W

aite

rs

Mor

ning

pla

sma

prol

actin

log

(ng

fl)t

W

aite

rs

Air

craf

t m

echa

nics

F

reig

ht h

andl

ers

Air

tra

ffic

con

trol

lers

S

ymph

ony

mus

icia

ns

Phy

sici

ans

Mor

ning

pla

sma

test

oste

rone

(m

mol

fl)§

M

SE

M

2·7

4±

0·0

2

2·71

±0

·02

2

·67

±0

·02

2

·64

±0

·02

2·

61 ±

0·0

2

2·5

3±

0·0

2

0·81

±0

·07

0·

81 ±

0·0

6

0·6

5±

0·0

6

0·5

3±

0·0

7

0·3

7±

0·0

6

0·2

5±

0·0

7

Fre

ight

han

dler

s 28

·0 ±

1· 3

A

ir t

raff

ic c

ontr

olle

rs

27·8

± 1

·4

Sym

phon

y m

usic

ians

26

·8 ±

1·2

A

ircr

aft

mec

hani

cs

24· 3

± 1

· 7

Phy

sici

ans

22·8

± 2

·0

Wai

ters

2

2·2

±1

·5

t F=

10·

06;

df=

5/1

44

; p

=0

·00

1.

tF=

11

·64

; d

f=5

/14

4;

p=

0·0

01

. §

F=

2·8

0;

df=

5/1

44

; p

=0

·02

.

Ant

ilog

, m

ean

lim

its±

SE

M

52

5-5

75

4

90

-53

7

44

7-4

90

4

17

-45

7

38

9-4

27

3

39

-35

5

5·5

-7·6

5

·6-7

·4

3·9

-5·1

2

·9-4

·0

2·0

-2·7

1

·5-2

·1

n 20

30

20

30

30

20

20

30

30

20

30

20

30

20

30

30

20

20

(Jl

00

~

~

"' 0 ~ ~ f;:..

:A

'"1:3 ~

.....

;:s-

o v.

0 ..... ~

!:>

;:s

!::>... ~

c· :; "' ::.: §..

v. E' ~

~

8 ~ .. !:>

~

~- 0 ~

~ ~ c·

::s v.

(Jl

\0

60 T. The ore 11 et al.

demands and poor social support, a relatively low level of joy and more cigarette smoking

compared to the other groups. Physicians in general reported relatively high levels of negative emotions but also

relatively high levels of joy. Furthermore, they reported a favourable ratio between demands and decision latitude in contrast to waiters, who reported a poor ratio.

Aircraft mechanics reported a relatively high level of intellectual discretion.

4. Discussion This study concerned samples of men from six different occupational groups. For various

' reasons, the samples studied might not be truly representative of all working men in those occupations, and therefore the results must be interpreted with caution. The samples studied were relatively small and in some of them the attrition rate was relatively high. Among 'blue-collar workers' (freight handlers, waiters and aircraft mechanics), the attrition rate was higher than among 'white-collar workers' (physicians, air traffic controllers and symphony

musicians): 45% versus 39% . A Swedish epidemiological study has shown that alcohol problems are more common

among non-participants than among participants in screening studies (Theorell et al. 1975) . Therefore, in the present study, it is possible that problems related to alcohol abuse might have been underestimated. However, the results indicated that serum GT levels did not differ between the groups. Therefore it is not likely that differences in alcohol abuse could explain other observed differences between groups. At the same time, the relatively large non-participation rates might have influenced the results in the direction of being 'too healthy'. This might be of particular importance in the groups with the largest non

participation rates, namely waiters and physicians. Due to the nature of the study- a cross-sectional study of relatively small samples- it is

impossible to speculate about causal factors. All that can be said is that there are biomedical differences between groups and also that those groups also describe their situations in markedly different ways. One group of particular interest is the group of waiters who reported the most unfavourable work environment with high psychological demands, low decision latitude and poor support. According to epidemiological observations, this combination is associated with premature development of cardiovascular symptoms (Norwegian Central Statistical Bureau 1976). In agreement with this observation, waiters reported smoking more cigarettes than the other participants and also showed little joy and lower cortisol and testosterone levels, the latter possibly being part of an endocrine profile associated with premature ageing. Due to their high level of cigarette smoking and relatively high levels of cholesterol and blood pressure, the coronary risk for waiters was

high, as expected. Another group with low decision latitude- freight handlers - also showed very high

blood pressure levels, which may have been at least partly due to physical factors such as heavy lifting and noise. Relatively high testosterone levels were also found. It is possible that these are also associated with the physical nature of the freight handler's job since physical activity may be associated with elevated testosterone levels (Adlercreutz et al. 1982). Leisure activities may further stimulate testosterone production.

Physicians, as expected, were shown to have a low coronary risk. This was expected since the ratio between demands and decision latitude was favourable. On the other hand, observations indicated that this caring job was emotionally demanding, arousing both negative and positive emotions. The observation that the physicians had relatively high mean plasma cortisol levels in the morning is difficult to interpret. Cardiac surgeons have also been observed to have high plasma cortisol levels (Payne et al. 1984). This is consistent


with other observations regarding a high suicide rate among physicians. Since the blood tests ':"ere pe~formed before work started, the levels are assumed to reflect the degree of uneasmess or JOY that the physician feels before work starts. It is possible that the emotional upset that is often caused by decision-making related to the suffering of others may induce uneasiness among clinically active physicians before they start their work. The high levels may also reflect an aroused mood in general (Fibiger et al. 1984).

It . is intere_sting to note that the physicians had low prolactin levels. Laboratory expenments with rats under ecological stress have shown that those rats who show evidence of being defeated have high prolactin levels (Henry et al. 1986). Although results from animal experiments are not necessarily relevant to human beings these results may have relevanc.e to.thts observation: physicians may feel that although they have an emotionally demandmg JOb they may still be able to control the situation and to influence decisions regarding their own job.

Air t~affic c~ntrollers described a relatively strenuous work situation. Despite this they showed httle ev~dence of ill health and few negative emotions. This may be surprising since a North Amencan study of air traffic controllers has pointed up pronounced health problems in this group (Rose et al. 1978). However, it should be noted that the intake of Swedish air traffic controllers in the 1980s is different from that of the Americans in the 1970s. Furthermore, great attention is being paid to work organization in the Swedish setting. Recent studies of air traffic controllers in the USA and outside North America (Maxwell et al. 1983, Booze and Simcox 1985) have been unable to verify the initial North American findings.

In general the findings did support results from other studies indicating that low decision latitude is associated with blood pressure elevation (Harenstam 1989, Schnall et al. 1989, Pieper et al. 1989, Theorell et al. 1988, Theorell et al. 1985) in the sense that physicians, a gro~? with. high decision latitude, had low blood pressure whereas two groups with low declSlon latitude- freight handlers and symphony musicians - had high blood pressure. In general terms, the study also supports the hypothesis that a combination of high demands and low decision latitude may have adverse effects on health (Karasek 1979).

The findings also point up the unfavourable psychosocial job situation of waiters as well as their unfavourable biomedical situation. The findings in physicians indicate that our theoretical understanding of the possible role of plasma cortisol in job stress is insufficient.

Acknowledgements Financial support for this study was given by the Swedish Work Environment Fund. Robert A. Karasek, Michael Marmot and James P. Henry have given valuable advice.

References ADLERCREUTZ, H. K., KuOPPASALMI, S., NARVANEN, K., KosuNEN, H. and HEIKKINEN R. 1982 Use of

hypnosis in studies of the effect of stress on cardiovascular function and horm~nes Acta' Medica o Scandinavica, 660 (suppl.), 84-94. ' AKERSTED~, T. and TORSVALL, L. 1977, Medicinska, psykologiska och sociala aspekter pa skiftarbete vid

spwalstalverken i Soderfors. Stress Research Reports 64 (Stress Research Laboratory o Stockholm). ' AKERSTE~T, T. and ToRSVALL, L. 1978, Experimental changes in shift schedules - their effects on well

bemg, Ergonomics, 21, 849-856. ARNETZ, ~: B., HORTE, L. G., HEDBERG, A., THEORELL, T., ALLANDER, E. and MALKER, H. 1987,

Smc1de . patterns among physicians related to other academics as well as to the general populat10n, Acta Psychiatrica Scandinavica, 75, 139-143.

62 T. Theorell et al.

BoHM-AUDORF, U. and SIEGRIST,]. 1983, Occupational morbidity data in myocardial infarction: a case reference study in West Germany, Journal of Occupational Medicine, 25,367-371.

BoozE, C. F. JR. and SIMCOX, L. S. 1985, Blood pressure levels of active pilots compared with those of air traffic controllers, Aviation, Space and Environmental Medicine, 56, 1092-1096.

DEVEREUX, R. B., PICKERING, T. G., HARSHFIELD, G. A., KLEINERT, H. D., DERBY, L., PREGIBON, D ., jASON, M., KLEINER, B., BORER, j. S. and LARAGH,). H. 1983, Left ventricular hypertrophy in patients with hypertension. Importance ofblood pressure response to regularly recurring stress,

Circulation, 68, 47(}--476. DIAGNOSTIC PRODUCTS CORPORATION, 1985 a, Cortisol No Predilution I 125 Double Antibody

Radioimmunoassay (DPC, Los Angeles). DIAGNOSTIC PRODUCTS CORPORATION, 1985 b, Double Antibody Prolactin-NHS (DPC, Los Angeles) . DIAGNOSTIC PRODUCTS CORPORATION, 1985 c, Double Antibody Testosterone (DPC, Los Angeles). FIBIGER, W., SrNBER, G., MILLER, A. J., ARMSTRONG, S. and DATAR, M. 1984, Cortisol and

catecholamines changes as functions of time of day and self-reported mood, Neuroscience and

Biobehavioural Sciences, 8, 523-530. HA.RENSTAM, A. 1989, Prison personnel- working conditions, stress and health. A study of2000 prison

employees in Sweden. Doctoral thesis, Karolinska Institute, Stockholm. HENRY,]. P., STEPHENS, P. M. and ELY, D. L. 1986, Editorial review. Psychosocial hypertension and

the defence and defeat reactions, Journal of Hypertension, 4, 687-697. JEPPSON, J.-0. and ENGLUND, H. 1983, Determination of HbA 1C Using a Monodispet~se Exchanger

(Pharmacia Fine Chemicals). KARASEK, R. A. 1979, Job demands, job decision latitude and mental strain: implications for job

redesign . Administrative Science Quarterly, 24, 285-307. KARASEK, R. A., THEORELL, T., SCHNALL, P., PIEPER, C. and MICHELA,j. 1988,Job characteristics in

relation to the prevalence of myocardia! infarction in the US HES and HANES, American

Journal of Public Health, 78, 910-918. MAXWELL, V. B., CRUMP,). H. and THORP,). 1983, The measurements of risk indicators for coronary

heart disease in air traffic control officers: a screening study in a healthy population, Aviation, Space and Environmental Medicine, 54, 246-249.

NoRWEGIAN CENTRAL STA'TISTICAL BuREAU, 1976, Yrke og dodlighet (Statistisk Centralbyra, Oslo) . OLSEN, 0. and SONDERGAARD-KRISTENSEN, T. 1988, Hjerte-karsygdomme og arbejdsmiljo

(Arbejdsmiljofondet, Copenhagen). PAYNE, R. L., RICK,]. T., SMITH, G. H. and CoOPER, R. G. 1984, Multiple indicators of stress in an

active job- cardiothoracic surgery, Journal of Occupational Medicine, 26, 805-808. PIEPER, C., LA CROIX, A. Z. and KARASEK, R. A. 1989, Job strain and cardiovascular risk factors for

male workers in the US HES, HANES I and HANES II, Western Collaborative Group Study and Exercise Heart Study, American Journal of Epidemiology, 129, 483-494.

ROLLASON,J. G ., PINCHERLE, G. and ROBINSON, D. 1972, Serum gamma glutamyl transpeptidase in relation to alcohol consumption, Clinica Chimica Acta, 39, 75-83.

RosE, R. M., HuRST, L. and JENKINS, C. S. 1978, Air Traffic Controller Health Change Study (Boston University School of Medicine), rep. FFA contract no. DOT-FA 73 WA-3211.

SCANDINAVIAN SOCIETY FOR CLINICAL CHEMISTRY AND CLINICAL PHYSIOLOGY (Committee on Enzymes) 1976, Recommended method for the determination of gammaglutamyltranspeptidase in blood, Scandinavian Journal of Clinical Laboratory Investigation, 36,

119-125. ScHNALL, P. L., PIEPER, C., KARASEK, R. A., ScHWARTz,]., ScHLUSSEL, Y., DEVEREUX, R. B., GANAU,

A., ALDERMAN, M., WARREN, K. and PICKERING, T. G. 1989, The relationship between job strain, workplace diastolic blood pressure and left ventricular mass. Results of a case-control

study, JAMA, in press. SJOBERG, L., SvENSSON, E. and PERSSON, L.-0. 1979, The measurement of mood, Scandinavian Journal

of Psychology, 20, 1-18. SWEDISH CENTRAL STASTISTICAL BUREAU, 1981, Dod.ifallsregister 1961-1970 (Statistiska Centralbyran,

PM 1981: 5, Stockholm), p. 5. THEORELL, T., LIND, E. and FwDfmus, B. 1975, The relationship of disturbing life changes and

emotions to the early development of myocardial infarction and other serious illness,

International Journal of Epidemiology, 4, 281-293. THEORELL, T., KNOW, S., SvENSSON, J. and WALLER, D . 1985, Blood pressure variations during a

working day at age 28: effects of different types of work and blood pressure at age 18,Journal of

Human Stress, 11, 36-41.


THEORELL, T., AHLBERG-H~~TEN, G., BERGGREN, T ., PERSKI, A ., SIGALA, F., SvENSSON,). and WALLIN, B .-~- 1987, Arbetsmtljo, levnadsvanor och risk for hjartkarlsjukdomar. Stress Research Reports 195 (Nattonal Institute of rsychosocial Factors and Health, Stockholm).

THEORELL, T., PERSKI, A ., AKERSTEDT, T., SIGALA, F., AHLBERG-HULTEN, G., SvENSSON, J. and ENE~OTH_, P. 1988, Changes in job strain in relation to changes in physiological states: a longttudmal study, Scandinavian journal of Work, Environment and Health, 14, 189-196.

Manuscript received 20 June 1989 Manuscript accepted 10 December 1989

A psychosocial and biomedical comparison between men in ... pc shelter/ill scans/theorell... · that waiters, the group reporting the most unfavourable working conditions, tended

Documents