death before the age it
is normally expected.
Usually set at deaths
under the age of 75.
Plate 2.2 The health of these children will be affected both by factors within their
immediate environment, and as a result of their environment being relatively deprived
when compared to financially more secure children.
Source: Nathan Benn/Corbis
THE IMPACT OF POVERTY ON HEALTH 45
Ferrie, J.E., Shipley, M.J., Stansfeld, S.A. (2003). Future uncertainty and socioeconomic inequalities in
health: the Whitehall II study. Social Science and Medicine, 57: 637–46.
The Whitehall studies may appear an odd way of addressing the relationship between personal
behaviour, social inequalities and health. The first study, Whitehall I, involved a crosssectional
study of people working in the British civil service headquarters in Whitehall.
Whitehall II was a longitudinal study examining the long-term relationship between these
factors. Among other things, the studies have examined the relationship between the type
of job people held, their behaviour and their health. Its ‘take home’ message has been that
after taking account of individual behaviour, the grade an individual holds within the civil
service is strongly predictive of how long they will live – the higher the grade, the better the
But what does a study of London-based British civil servants have to tell us? The strength
of the Whitehall study is the similarity of working and living conditions of the entire cohort
of people studied. Other studies have examined the relationship between disease rates and
socio-economic status in much more diverse populations. While this has some advantages, it
also means that comparisons between people from different socio-economic groups inevitably
involves comparisons of people in different types of job, as well as working and living environments.
Accordingly, differences in mortality between these various groups may be the
result of differential exposure to pollutants, accidents, and so on. In the civil service, no such
differences exist – from cleaner to chief civil servant, all people work in the same environment.
Similarly, virtually all the workers live in London and thus share the same levels of air pollution
and other environmental factors. These similarities mean that any differences in health
or mortality across the social spectrum cannot be accounted for by differences in working or
The Whitehall studies were seminal in identifying the risk of ill-health and premature mortality
associated with low socio-economic status. They have now begun to explore some of the
mechanisms through which these differences may arise. The paper by Ferrie et al. examined
the relationship between job and financial insecurity and both mental and physical health.
Participants in the Whitehall II study were all London-based civil servants. At baseline, the
study population comprised 6,895 men and 3,413 women of all grades in the civil service.
Baseline data were collected between 1985 and 1988. Follow-up measures were taken at
various times since then. The present study reported the fifth follow-up assessment (phase 5),
which occurred between 1997 and 1999. Participants received a clinical examination measuring
blood pressure, cholesterol, weight, and so on. In addition, they completed a questionnaire
that included lifestyle, work characteristics, social support, life events and chronic
difficulties. Of particular relevance to this report were the following measures:
n Health rating: ‘In general would you say your health is: excellent, very good, good, fair, poor?’
n General health questionnaire (Goldberg 1997): measuring depression and anxiety;
n Job security: ‘How secure do you feel in your present job?’
n Financial security: ‘Thinking of the next ten years, how financially secure do you feel?’
46 CHAPTER 2 • HEALTH INEQUALITIES
Minority status and health
A second set of factors that discriminate between people in society is whether
or not they occupy majority or minority status within the general population.
This is usually considered in terms of the ethnic or cultural background of the
individual, but it may also reflect other differences such as those associated
with sexuality and religion.
Perhaps the most obvious minority in any population are people who differ
from the majority in terms of skin colour, religion and so on: often considered
By the fifth phase of the study, 7,270 of the original participants had completed the full
follow-up assessment. A further 560 participants completed a shorter assessment process by
telephone, bringing the total number of people with relevant data to 7,630: 76 percent of
the original sample. This is a good response rate, particularly as non-completers included
people who had died over the intervening years or who had left the civil service and could
not be found. The key results of the study were:
n Participants in higher-ranked jobs reported better physical and mental health, less longstanding
illness, lower levels of obesity and lower blood pressure levels. However, they did
not vary in their cholesterol levels.
n A similar association between previous job rank and measures of health was found among
people unemployed at the time of phase 5 assessment.
n People in lower job grades reported higher levels of job insecurity than those in the higher
n People in lower grades, or who were unemployed at the time of the assessment and who had
been in lower grades, reported more financial insecurity than those in the higher grades.
n Job insecurity was only marginally associated with any of the measures of health.
n By contrast, financial insecurity was associated with poorer self-rated health, long-term
illness (in men but not women) and levels of depression among both employed and
This is an interesting paper in many ways, but it also leaves a number of questions. An important
finding was the association between job grade and health. People in higher grades did better
on both objective and subjective measures of health. What was also evident was that these
inequalities in health continue even when people leave the civil service and become unemployed.
Also of interest was that financial security but not job security was associated with health.
It seems that we are less concerned about whether we can keep a particular job than whether
we are likely to continue to earn sufficient money to maintain an established lifestyle.
MINORITY STATUS AND HEALTH 47
under the rubric of ethnic minorities. Nazroo (1998) pointed out that ethnicity
encompasses a variety of issues: language, religion, experience of races
and migration, culture, ancestry, and forms of identity. Each of these may
individually or together contribute to differences between the health of different
ethnic groups. He therefore warned about considering all people in all
ethnic minority groups as one single entity and thereby failing to recognise
the reality of their differing lives.
These cautions are perhaps reflected in findings that in the UK, while rates
of ill-health and premature mortality among people from ethnic minorities
are generally higher than those of the white population, people from the
Caribbean experience better health (Wild and McKeigue 1997). Levels of
disease also differ across ethnic groups. Rates of heart disease among British
men from the Indian subcontinent, for example, are 36 percent higher
than the national average. Among young people, these differences are even
greater: young Indian men are nearly three times more likely to develop
heart disease than their white counterparts. The Afro-Caribbean population
has particularly high rates of hypertension and strokes, while levels
of diabetes are high among Asians. By contrast, rates of lung cancer are
relatively low in people of Caribbean or West African origin (Balarajan and
In searching for explanations for the relatively poor health of people from
ethnic minorities, a number of issues have to be borne in mind. Perhaps the
most important is that a disproportionate number of people from ethnic
minorities also occupy low socio-economic groups. Before suggesting that
being in an ethnic minority alone influences health, the effects of these socioeconomic
factors need to be taken into account in any comparison between
ethnic minorities and majority populations. This can be done by comparing
disease rates between people in ethnic minorities and people from the majority
population matched for income or other markers of SES, or by statistically
partialling out the effects of SES in comparisons between majority and
minority populations. Once these are done, any differences in mortality
between the two groups lessen. Haan and Kaplan (1985), for example, found
significantly higher rates of disease and premature mortality between
American black and white populations (as large as a 30 percent difference),
which disappeared after partialling out the effects of SES. Other studies (e.g.
Sorlie, Backlund and Keller 1995) have found a reduction, but not negation,
of health differentials after partialling out the effects of SES.
Socio-economic status certainly exerts an influence within ethnic minorities:
just as for the majority population, people in the higher socio-economic
groups generally live longer and have higher health ratings throughout the
life course than those with fewer economic resources (Harding and Maxwell
1997; Davey-Smith et al. 1996). However, again highlighting the dangers
of considering people in different ethnic minorities as one single group, there
are some exceptions to this rule. In the UK, for example, there appears to be
no SES-related differential risk for CHD among men born in the Caribbean
or West or South Africa (e.g. Harding and Maxwell 1997). Despite these
cautionary notes, there is a general consensus that ethnicity does impact on
health, although different factors may affect different ethnic groups. A number
of explanations for these differences have been proposed. These mirror,
to some extent, those associated with SES: differences in health-related
behaviour, stress and access to healthcare.