Background
It is well established that poorer people tend to experience more
illness and to die at a younger age than their more affluent counterparts.
The problem of such health variations is difficult to solve, as
its causes are unclear. Traditional explanations, for example, unhealthy
behaviours (like smoking, lack of exercise or poor diet) only explain
part of the differences. Different levels of psychological stress
across social classes has been suggested as the missing explanation.
Several investigators have studied this question but methodological
problems have prevented the reaching of any firm conclusions. For
example, there may be common reporting tendencies such that people
who report higher levels of stress also tend to report higher levels
of morbidity, with this common reporting tendency underlying the
apparent relationship between stress and health. It is also possible
that ill-health itself can generate increased perceptions (and therefore
reporting) of stress, with the direction of causation being from
health to stress, rather than from stress to health. Furthermore,
rather than stress "explaining" socio-economic differentials
in health the fact that both stress and health are related to socio-economic
position could generate an apparent, but non-causal, association
between the two. Finally, stress may be related to future health
through an influence of stress on health-related behaviours (such
as smoking or alcohol consumption). This study aims to advance our
understanding of these issues.
Aims and
Objectives
This study will investigate:
- whether reporting
of stress is related to a general reporting tendency relating
to health;
- whether there
are cross-sectional associations between stress and health-related
behaviours;
- whether there
are cross-sectional associations between stress and measures of
health status;
- whether changes
in stress between two time points are related to changes in health
related behaviours or health status;
- whether stress
predicts mortality and hospital admission over a 20 year follow-up
period;
- whether any
associations between stress and health over this follow-up period
account for the observed socio-economic differentials in health.
Study Design
The study is a prospective investigation of around 6,000 Scottish
men and 1,200 Scottish women examined at their work places in the
early 1970s and followed-up for hospital admissions and mortality
since then. Extensive data on health-related behaviours, socio-economic
position, biological risk factors and morbidity were collected,
together with reports of the experience of stress. Hospital admission
data, cancer registrations and mortality follow-up have been collected
on this cohort over the subsequent 20 years. We will relate socio-economic
position to health outcomes and investigate the extent to which
stress accounts for these differentials.
Policy Implications
Stress - including work stress - is widely perceived to be an important
influence on health status. In particular, studies of lay epidemiology
repeatedly find stress to be among the major health concerns of
large sections of the population. However, robust evidence on the
degree to which stress influences physical health is lacking. The
methodological limitations of many previous studies feed into this.
If stress, including work stress, can be shown to be an important
influence on health and an important contributor to inequalities
in health then the need to make organisational changes which would
reduce such stress would be emphasised.
Project Summary
Psychological stress has been suggested as a factor contributing
to ill health and to health inequalities. However, researching the
links between stress and health presents a number of complex methodological
issues. For example, causation may run from health to stress and,
where both health and stress are measured through self-report, both
measurements may be inflated by a common reporting tendency. Further,
an apparent relation between health and stress may be confounded
by their association with socio-economic position; the effects of
stress on health may also be best understood through consideration
of cumulative or changing stress over time.
The project
used longitudinal data from a large cohort study of working men
and women to investigate the contribution of psychological stress
to socio-economic differentials in morbidity and mortality, in analyses
which took account of the role of physiological risk factors and
health behaviours. The West of Scotland Collaborative Study recruited
6000 men and 1000 women from 27 workplaces in 1970-3.
Key findings
- Among men,
higher socio-economic groups registered higher perceived stress
scores. Among women, there was no clear association between perceived
stress and occupational class. There was also an association between
greater stress and better health among men. For example, the relative
hazard of all cause mortality was reduced in medium and high stress,
as compared to low stress, participants. Most causes of mortality
showed this pattern, as did change in stress score (i.e. reduced
risk of mortality among men with increasing or stable stress scores
as compared with those with decreasing stress scores). These associations
reflected the fact that both higher stress and better health were
related to social advantage.
- Perceived
stress was positively associated with physiological risk factors
for coronary heart disease, including high plasma cholesterol
(higher reported stress: higher plasma cholesterol). However,
it was negatively related to high diastolic blood pressure, body
mass index (BMI) and low forced expiratory volume (FEV1).
With the exception of BMI, correction for socio-economic position
removed the associations between stress and physiological risk
factors, indicating that they were largely due to confounding
by SES. This suggests that the effects of perceived stress on
physiological risk are, to an extent, a function of the particular
patterning of stress and physiological risk across different social
groups.
- Perceived
stress was associated with an adverse profile of health-related
behaviours in the expected direction. Greater stress was associated
with greater number of cigarettes and greater alcohol consumption
among both men and women and with less recreational exercise for
men only. These associations were largely independent of confounding.
This suggests that the association between high stress and risk-related
behaviour is unlikely to be a function of the coincidence of high
stress and high risk in particular social classes.
- Reported
stress showed a strong direct association with reporting tendency
(tendency to over-report somatic symptoms in the absence of somatic
pathology) in men.
- Analyses
of the associations between job satisfaction, perceived stress
and mortality found an association between lack of job satisfaction
and greater perceived stress. But there was little evidence to
suggest that lack of job satisfaction was associated with cardiovascular
risk factors and no evidence to suggest that it was associated
with cardiovascular mortality. With respect to policy, these findings
suggest that interventions to reduce stress and improve the psycho-social
environment may have only modest effects on health inequalities.
- Analyses
examined the cumulative effect of socio-economic position on risk
factors (diastolic blood pressure, plasma cholesterol concentration,
BMI and FEV1) and mortality among working women. It
demonstrated that cardiovascular risk factors were related to
both past and present socio-economic position and that lifecourse
socio-economic experience was a more potent predictor of all cause
mortality and mortality from cardiovascular disease than other
measures of socio-economic position. This suggests that there
is great potential for reducing health inequalities by improving
material and social lifetime circumstances which contribute to
risk factors for mortality.
|