|
Background
Obesity is increasing rapidly in most industrialised countries and
poses a public health problem for which effective, practical and
long-lasting preventive and therapeutic solutions are lacking. Obesity
has a strongly negative association with socio-economic status (SES)
among women in developed countries: in developing countries, obesity
rates rise in line with increasing socio-economic status. In the
UK, women in social class V have a higher Body Mass Index (BMI)
than women in social class I and run almost twice the risk of obesity.
SES differences may still be increasing. Earlier research has pointed
to the contribution of genetic factors, social mobility, lifestyle
and dieting to the socio-economic gradient in obesity. However,
the contribution of psycho-social processes has been a neglected
area of research, both within obesity research and within the health
inequalities field.
Aims and
Objectives
The aim of the present study is to shed light on the psycho-social
processes which lead to SES differences in obesity. In addition
to testing the hypothesis that lifestyle and dietary restraint differences
underlie the obesity variation, we shall be examining social norms,
attitudes and behaviour in relation to weight control in women in
critical life stages for the development of obesity: adolescence,
pregnancy and mid-life.
Study Design
Two new samples of women will be recruited for study, selected on
the basis that they are at phases of their lives which are important
for the development of obesity: 1200 adolescent girls and 600 women
who have just had a first pregnancy. In addition, data from the
Whitehall II Study (Phase 5) will be used to look at a sample of
older working women (aged 47 - 67), where SES is measured by employment
grade.
In the adolescent
and post-natal samples, information will be collected on several
indicators of socio-economic status. Information on body measurements
will be collected at the time of the interview along with information
on body image and weight satisfaction. Family history of obesity
will be based on participants' re-call. Diet will be assessed with
a food frequency questionnaire. Physical activity and smoking will
be assessed using the same questions as in the Health Survey for
England, for comparative purposes. Stress levels and behavioural
responses to stress will be recorded. Dieting practices will also
be assessed, along with nutritional knowledge and understanding
of causes of obesity. Finally, possible symptoms of eating disorder
will be measured, as there may be some relationship between weight
control practices and eating disorders.
Participants
in the Whitehall II Study are completing the same dietary record
and similar measures of dietary restraint, hunger and emotional
eating. Activity is measured using a novel instrument which is being
validated against heart rate monitor and motion sensor readings.
Body weight and waist/hip ratio have also been measured three times
in this sample. Measures of self-reported health, satisfaction with
life, health locus of control and psycho-social well-being are included.
In each case,
the results will be analysed first by looking at SES differences
in the main dependent variables and then using multivariate analysis
techniques to see whether the weight-related attitudes and behaviours
contribute to SES variation in weight. Path analysis will be used
to test the model underlying the project: that material and cultural
factors influence individual knowledge and motivation, which in
turn affect behaviour.
Policy Implications
The results of this study should provide valuable information on
SES differences in weight control which will inform policy-makers
concerned with reducing population weight and reducing SES differences
in obesity. The focus on practical issues such as knowledge about
weight control and opportunities to adopt healthier practices will
be useful in relation to considering the role of health education
and leisure and environmental provision. This project has a very
specific focus, but the model it develops and tests could be illustrative
of a research approach with wider applicability across other health
behaviours.
Project Summary
Body weight is attracting increasing attention from the public health
community as the prevalence of overweight and obesity rises and
evidence of its role in the development of diseases like coronary
heart disease and diabetes increases. Among women, there is a marked
socio-economic gradient in obesity, with the proportion rising from
14% in social class I to 25% in social class V. The gradient among
children and among men is less pronounced but there is evidence
that it is becoming more like the female gradient over time.
The project's
aim was to increase understanding of the socio-economic variation
in obesity by focusing on women at three critical life stages for
the development of obesity: in adolescent girls, postpartum women
and older women working in the civil service (Whitehall II Study).
It conducted a school-based survey of girls aged 13-15 years (n=1248)
in Wirral and Cheshire. Analyses of national data (for Britain),
collected as part of the Office of National Statistics (ONS) Omnibus
survey in March 1999, were also undertaken to examine the association
between socio-economic status (SES) and attitudes and practices
in relation to weight control. It should be noted that the ONS study
relies on self-reported data, which tend to under-estimate weight
and body mass index (BMI). However, given the focus on SES differences,
the shortcomings of these data are not likely to substantially affect
the findings.
Key findings
- Analyses
of the ONS national data pointed to marked social class differences
in attitudes and practices regarding weight control. Those in
the lowest SES group were significantly less likely to feel overweight
than those in the highest group. For both men and women, those
in higher SES groups were also more likely to monitor their weight,
to feel overweight and to try to lose weight. This difference
is even more striking among women, in view of the lower prevalence
of obesity among those in higher socio-economic groups. These
findings suggest that attitudes and behaviour in relation to weight
control could play a part in the generation of SES differences
in obesity. However, the cross-sectional nature of the study means
that causal hypotheses could not be tested.
- The school-based
study of adolescents found no significant socio-economic gradient
in weight or BMI, or in body size ideal. However, there were important
differences in other aspects of body image. Greater awareness
of the social ideals of thinness was reported by girls with parents
in higher SES groups. Their social environments also provided
more modeling of weight control, in the form of discussion of
issues of weight control in the family, and friends and family
members who were trying to lose weight. Higher SES girls were
also more dissatisfied with their bodies and reported more restrained
eating and healthier weight control.
- A major problem
for studies of socio-economic influences in adolescence is the
difficulty of measuring SES using conventional indicators like
parents' income, education and occupation. As a result, surveys
often have a high proportion of respondents with missing data
for parental SES, which in turn can introduce biases into the
study. Building on a family affluence scale developed by researchers
at Edinburgh University, the project team constructed a Home Affluence
Scale composed of material items (family owning a car, family
owning two cars, the family owning the home, the family owning
a computer and the student not having the option of free school
meals). The new scale was tested in the adolescent survey (1248
girls and 567 boys). There were high completion rates for all
items in the scale, which offered a reliable substitute for occupational
and educational data. Additionally, it prevents exclusion of those
materially less well off adolescents who are most likely to fail
to complete conventional SES items.
- Analysis
of the Whitehall II Study examined whether social class differences
in eating behaviours and cognitions might account for part of
the gradient in overweight and obesity in women. Three psychological
eating behaviour variables were studied: restraint (conscious
control of eating behaviour), disinhibition (an unrestrained approach
towards food) and hunger (feeling too hungry to diet or to control
food intake). Of the three, hunger and disinhibition explained
a moderate amount of the gradient in body size across employment
grades and may be useful concepts for future research on the socio-economic
differentials in obesity and overweight.
|
