Professional Documents
Culture Documents
These include:
References This article cites 28 articles, 18 of which can be accessed free at:
http://jech.bmj.com/cgi/content/full/55/12/895#BIBL
Email alerting Receive free email alerts when new articles cite this article - sign up in the box at the
service top right corner of the article
Topic collections Articles on similar topics can be found in the following collections
Notes
www.jech.com
Downloaded from jech.bmj.com on 24 March 2008
one; those with higher incomes are able to pur- a number of diVerent income sources includ-
chase better food, better housing, live in safer ing, in some cases, money paid on their behalf,
environments and have better access to health rather than directly to them (for example,
care. The second emphasises behavioural or housing benefit, which in Britain is paid to
“lifestyle” factors, such as smoking, diet, landlords). Collecting these data requires a
alcohol consumption and appropriate use of large number of questions, preferably supple-
health care, which may vary with cognitive skill mented by examination of relevant documents
and access to information. The third places and this, coupled with the sensitivity of the
more emphasis on psychosocial factors such as topic, results in lower response rates in surveys
empowerment, relative social status and social dealing with income than in other types of sur-
integration, including exposure to stresses that vey.24
may result from low status and low autonomy Education is often regarded as an indicator
in important arenas of life, such as work. of first choice because, as educational attain-
The three most commonly employed indica- ment is normally fixed early in life, problems of
tors of socioeconomic status in contemporary reverse causation are much less serious.30
industrialised societies are income, education, However in Britain, and many other European
and occupation.17 18 Choice of indicator may countries, most of today’s older population left
reflect preference for one theoretical pathway. school at the minimum age with no academic
If materialist explanations for health inequali- qualifications. This means that the extent of
ties are preferred, then income would seem the diVerentiation possible is limited and edu-
most appropriate indicator. Behavioural influ- cational variables may only allow the most
ences might be hypothesised to relate more advantaged to be distinguished from the rest of
closely to education, while occupational char- the population.22
acteristics or measures of relative deprivation, An alternative to individual based measures
is to use couple or household indicators. For
might be chosen by analysts wishing to investi-
married women social class classifications
gate psychosocial links between socioeconomic
based either on husband’s occupation or on the
status and health. However, as all these indica-
occupation of the highest status household
tors are interrelated and none of them capture
member (the “dominance” method) tend to be
in entirety domains identified as important in
associated with greater diVerentiation in mor-
the theoretical literature, such an approach tality and health than measures based on own
may be over simplistic. In practice all three of occupation.26 31 However, for older women,
these indicators, together with others consid- fewer of whom are currently married, Arber
ered below, are widely used in analyses of and Ginn found that both methods gave similar
health inequalities.19–22 In many studies these results.7
measures, which are highly correlated, are used Other household based indicators, such as
almost interchangeably although this may access to cars and housing tenure, are also fre-
hamper eVorts to understand how social posi- quently used32 33 and have been advocated as a
tion aVects health.3 23 One thing all have in sensible way of classifying groups, such as mar-
common is that their use for studying older ried women and older people, not themselves
populations is problematic. in the labour market.34 Ownership of resources
In Britain fewer than 10% of men and 5% of is an indication of wealth, which has been
women aged 65 and over are working and even shown to be associated with health35 and may
among 60–64 year old men who are technically also facilitate social participation. Availability
of “working age” those in employment are now of a car may make shopping for food, access to
a minority.24 This clearly makes use of classifi- health care services and social activities easier.
cations based on occupation problematic, not Housing tenure may be associated with type of
least because hypothesised links between work neighbourhood and to some extent with hous-
environments and health are not directly appli- ing quality although much of the poorest hous-
cable (although work conditions in midlife may ing lived in by older people is owner occupied
be hypothesised to influence health in later or privately rented.36 Certain amenities (such
life). Additional complications arise because as central heating) may reduce the risk of
reasons for leaving work early may be health exposure to damp or cold.37 In general these
related and poor health may be associated with variables are treated as indicators of socioeco-
downward social mobility towards the end of nomic status and as all are associated with the
working life.13 25 Moreover the employment three dimensions of socioeconomic status con-
histories of men and women diVer substan- sidered above (occupational social class, educa-
tially, both in terms of labour market involve- tion and income, as well as with wealth)
ment and types of occupation.7 26 indirect associations are clearly applicable.
Income is strongly associated with employ- Household-based measures obviously have the
ment27 28 and so for the same reasons problem- advantage of applying to nearly everybody
atic of as an indicator of socioeconomic status (except those in institutions) but may be influ-
in older age groups. Additionally, as those with enced by factors other than socioeconomic sta-
serious health or disability problems are tus, including health, in the older population.
eligible for various types of financial assistance, Lack of access to a car, for example, may be a
reverse causation problems are compounded.29 consequence of widowhood or some health
Apart from these factors that complicate impairment rather than low income. Deterio-
interpretation, collecting comprehensive and rating health may also prompt support related
accurate data on income is diYcult especially moves to the households of relatives or special-
for groups, such as older people, who may have ist housing and associated changes in tenure
www.jech.com
Downloaded from jech.bmj.com on 24 March 2008
type and household resources38 39 that reflect older population, those responsible for collect-
socioeconomic status of the relatives rather ing data for health needs assessments of popu-
than the older person. This problem is more lations and health care planners.
marked in studies of the older old as in the age
groups considered in this study the proportions
Methods
who move to live with relatives is very small.
The data we use come from the Retirement
Related to, but not equivalent to, systems of
and Retirement Plans Survey that was carried
socioeconomic classification are the issues of
out in 1988/9, with a follow up survey
identifying poverty and deprivation. Primary
conducted in 1994. The survey was commis-
poverty—that is, lack of suYcient food and
sioned by the Department of Social Security
basic shelter—now aVects only a small minor-
and carried out by the OYce for National Sta-
ity in developed countries such as Britain.
tistics and has been described in detail
Deprivation, defined by Townsend40 as the
elsewhere.29 46 In brief, at baseline a question-
inability to participate in normal activities
naire was administered to a nationally rep-
because of a lack of material resources, is more
resentative sample of 3543 adults aged 55–69
widespread. Measuring deprivation requires a
identified through a prior sift of a random
conceptual framework that takes account of the
sample of addresses. In this sift 2717 house-
circumstances of the whole population and
holds including one of more adults in the target
defines some within it as lacking resources
range were identified, of which 88% agreed to
considered normative. Townsend used a check-
take part in the survey. In 1994, surviving
list approach to ask about resources commonly
respondents were traced and 70% (2243) were
regarded as necessities. A similar approach has
re-interviewed using essentially the same ques-
been used in more recent studies, in some cases
tionnaire. A re-weighting procedure was ap-
with additional distinctions made between
plied to the data to reduce bias arising from
“basic” or “luxury” items.23 41 Explanations for
diVerential response to follow up (see survey
associations between deprivation and health
report for full description29). We have previ-
rest on psychosocial theories about health, as
ously used these data to analyse diVerentials in
well as on possible direct links between owner-
health and disability status.10 47 Our aim in
ship of certain resources, such as adequate
these earlier analyses was to analyse the eVect
heating and cooked meals, and health out-
of demographic and socioeconomic factors,
comes.
both current and past, on health and disability
One other important approach is the use of
and changes over time in disability. Here we are
ecological data. The literature suggests that liv-
concerned with illuminating the use of com-
ing in a deprived area has direct eVects on
mon indicators and identifying which may be
health status42 43; area classifications are also
most useful in studies of the health of the older
often used to draw inferences about the socio-
population.
economic status of people living in these areas.
Three measures based on individual charac-
The possible “ecological fallacy” arising from
teristics were used in the analysis, these were:
this approach is widely recognised.44 This may
be particularly true for older people as often
the socioeconomic indicators included in area Occupationally defined social class
based classifications, such as unemployment The dataset includes occupational histories
rates, are not directly applicable to them. from which lifetime social class measures may
Moreover localities that are now deprived may be derived. However, as this level of detail is
not have been so in the past when the older not available in most surveys, we used infor-
people resident in them set up home. This may mation on current or last occupation to derive
explain why in some areas with high excess social class, based on the OPCS Classification
mortality among younger people, mortality of Occupations48 and the registrar general’s six
among elderly people is below the national category social class grouping, as this was the
average.45 The social characteristics of neigh- classification current when the data were
bourhoods, particularly “social capital” indi- collected. Married women’s social class was
cated by the extent of community participa- assigned on the basis of their own last occupa-
tion, feelings of trust and mutual support and, tion.
less positively crime and vandalism, are also
increasingly recognised as potentially impor- Income
tant influences on health.46 However, few Information on all sources of income was
surveys of people include relevant indicators.4 6 collected. The measure we use here is net
In this paper we examine diVerentiation in a “benefit unit” income (benefit units are
sample of 55–75 year olds using seven roughly equivalent to families and in the case of
indicators of socioeconomic status and analyse this survey essentially distinguish married from
associations between these variables, singly and unmarried respondents). In the case of cou-
in combination, with an indicator of health sta- ples, individual income was set at 0.8 of the
tus. Our aim is to see which measure or pair of income of the couple, in line with recognised
measures best meet the criteria discussed practice.49 In the analyses shown here we have
above (grounded in theory, easily collected, diVerentiated the sample by income quartile.
sensitive enough to allow identification of gra-
dients and identification of the most disadvan- Educational qualifications
taged and not an outcome of health status). Information on a wide range of qualifications,
Our broader purpose is to provide information including vocational qualifications gained at
useful to analysts of health diVerentials in the work, was collected. Here we distinguish
www.jech.com
Downloaded from jech.bmj.com on 24 March 2008
www.jech.com
Downloaded from jech.bmj.com on 24 March 2008
16
14 Townsend
deprivation
indicators
12
10
Odds ratio
Household
8
resources
Income lacked
Educational quartile
6 Social Car
class attainment Housing access
tenure
4
0
IIInm
IIIm
IV & V
O level
Trade
None
2nd
3rd
Lowest
1
2
3 or more
1
2
3 or more
Private
LA/HA
No access
Figure 1 Odds ratios (95% confidence intervals) of fair or not good health by each of the socioeconomic status variables,
men 1988/9 (aged 55–69).
6 Townsend
deprivation
indicators
5 Income
quartile Household
Educational
resources
attainment
Social lacked
Odds ratio
4
class Housing
Car
tenure
access
3
0
IIInm
IIIm
IV & V
O level
Trade
None
2nd
3rd
Lowest
1
2
3 or more
1
2
3 or more
Private
LA/HA
No access
Figure 2 Odds ratios (95% confidence intervals) of fair or not good health by each of the socioeconomic status variables,
women 1988/9 (aged 55–69).
www.jech.com
Downloaded from jech.bmj.com on 24 March 2008
Household
resources
6
lacked
Car
access
5 Townsend
Educational deprivation
attainment indicators
Social
Income Housing
class
4 tenure
Odds ratio
quartile
0
IIInm
IIIm
IV & V
O level
Trade
None
2nd
3rd
Lowest
1
2
3 or more
1
2
3 or more
Private
LA/HA
No access
Figure 3 Odds ratios (95% confidence intervals) of fair or not good health by each of the socioeconomic status variables,
men 1994 (aged 59–75).
variables investigated, controlling for age (sin- by income quartile, household resources lacked
gle years). The graphs are shown separately for and Townsend indicator and, for women, also
men and women and baseline and follow up by educational qualification and social class
surveys. All the variables show significant asso- (although as can be seen by examining the
ciations with the outcome measure. Where the confidence intervals diVerences between sub-
data are structured in an ordered ranking, the groups were generally not significant at the 5%
results for 1988 generally show gradients in the level). In the older, and smaller, 1994 sample
expected direction in odds of fair/poor health income quartile and the Townsend indicator
7
Townsend
Household
deprivation
6 resources
indicators
lacked Housing
Social tenure
Odds ratio
5 class Income
Educational quartile
4 attainment
3 Car
access
2
0
IIInm
IIIm
IV & V
O level
Trade
None
2nd
3rd
Lowest
1
2
3 or more
1
2
3 or more
Private
LA/HA
No access
Figure 4 Odds ratios (95% confidence intervals) of fair or not good health by each of the socioeconomic status variables,
women 1994 (aged 59–75).
www.jech.com
Downloaded from jech.bmj.com on 24 March 2008
Table 2 Results from logistic regression analysis of risk of fair or not good health for pairs of socioeconomic variables
1988/9 1994
also seem to be associated with health status in be taken into account. In the 1994 female sam-
a graded fashion (although confidence inter- ple there was little to choose between adding
vals are wide). However, among men associa- housing tenure or the Townsend indicator, we
tions with household resources are not as have chosen the Townsend indicator for inclu-
expected. sion in the summary of “best” indicators shown
To examine further the best indicators of in table 3, partly because it is easier to interpret
social status, individual level factors were and because it shows gradients with outcome,
paired with household level ones and the which tenure does not. Table 3 also shows
whether the addition of other socioeconomic
eVects on model fit were investigated. Table 2
variables significantly improved the model fur-
gives the results of this analysis. The individual
ther, in three cases the addition of housing ten-
level variable that resulted in the largest reduc- ure and another variables produced further
tion in deviance varied. In the 1988 male sam- significant reductions in deviance.
ple the education variable performed best in
this regard, in 1994 it was social class. For the
Discussion
women in 1988 income quartile seemed best,
The issue of how best to measure socioeco-
in 1994 either income or education. All
nomic status is an important one for both
pairings with a household based measure researchers and policy makers and the replace-
produced improvements in model fit, as ment of the registrar general’s Social Class
indicated by statistically significant reductions schema with the new UK National Statistics
in deviance. (Changes in deviance follow an Socioeconomic Classification53 makes such a
approximate ÷2 distribution and can be tested consideration timely. Previous work on this
using a ÷2 test.54) In general, car aVordability topic has largely focused on how to measure
gave the least reduction in deviance when women’s status.7 26 34 Davey Smith and others
paired with an individual level measure, while also investigated whether social class (based on
the Townsend deprivation and household occupation) or education was more strongly
resources indicators appeared best, although associated with diVerentials in smoking and in
diVerences in degrees of freedom also need to mortality using data from the West of Scotland
workplace study, a follow up study of men
Table 3 Variables (pair) included in best fit models
recruited from a range of workplaces 1970–
Men Women 73.13 They concluded that in this study social
class performed better. We are not aware of
SES measure 1988/9 1994 1988/9 1994
other studies using a wide range of indicators,
Individual level variables as here, and focusing on an older sample of
Social class x both men and women.
Income quartile x In the introductory section of this paper we
Educational qualifications x x
Household level variables identified various criteria against which the
Townsend indicator x x x utility of socioeconomic status indicators in
Household resources x studies of health inequalities could be judged.
Housing tenure
Car access These included having a theoretical basis, ease
Does model improve with the the No Yes Yes Yes of collection, suYcient sensitivity to allow
addition of further variables? identification of gradients and identification of
If so, which variables? tenure tenure tenure
education consumer durables the disadvantaged, and not an outcome of
2
r 0.1028 0.0875 0.0706 0.1012 health. None of the variables we have consid-
÷2 170 93.8 127.9 129.9
Df 7 9 12 12
ered meet all these criteria on their own. From
our results, income quartile seems attractive as
www.jech.com
Downloaded from jech.bmj.com on 24 March 2008
www.jech.com
Downloaded from jech.bmj.com on 24 March 2008
“reverse causation” problems. The combina- 13 Davey Smith G, Hart C, Hole D, et al. Education and occu-
pational social class: which is the more important indicator
tion that would seem most promising would be of mortality risk? J Epidemiol Community Health 1998;52:
either occupational social class or education 153–60.
14 Marmot M, Theorell T. Social class and cardiovascular
paired with the Townsend deprivation indica- disease: the contribution of work. Int J Health Serv
tor or perhaps a combination household 1988;18:659–74.
15 Blaxter M. Fifty years on—inequalities in health. In: Murphy
resource/Townsend deprivation indicator. We M, Hobcraft J, eds. Population research in Britain. London:
would suggest that using a combination of this Population Investigation Committee, 1991:69–94.
16 Wilkinson R. Unhealthy societies: the aZictions of inequality.
type in studies of health inequalities and diVer- London: Routledge, 1996.
entials in use of health care would improve our 17 Liberatos P, Link B, Kelsey J. The measurement of social
class in epidemiology. Epidemiol Rev 1988;10:87–121.
knowledge of the extent, possible causes and 18 Preston SH, Taubman P. Socioeconomic diVerences in
consequences of social inequalities in the older adult mortality and health. In: Martin LG, Preston SH,
eds. Demography of aging. Washington DC: National Acad-
population. emy Press, 1994:279–318.
19 Davey Smith G, Hart C, Blane D, et al. Lifetime
socioeconomic position and mortality: prospective obser-
Funding: the research reported here was funded by the Economic vational study. BMJ 1997;314:547–52.
and Social Research Council (UK) as part of its Health Variations 20 Cavelaars A, Kunst A, Geurts J, et al. DiVerences in self
Programme, grant reference number L128 25 1040. reported morbidity by educational level: a comparison of
Conflicts of interest: none. 11 Western European countries. J Epidemiol Community
Health 1998;52:219–27.
21 Kennedy B, Kawachi I, Glass R, et al. Income distribution,
socioeconomic status, and self rated health in the United
Appendix A—Household resources States: multilevel analysis. BMJ 1998;317:917–21.
“Can you look at the things listed on this card and tell me 22 Martelin T. Mortality by indicators of socioeconomic status
which you have and which you do not have in this household?” among the Finnish elderly. Soc Sci Med 1994;38:1257–78.
% in national sample50 considering item necessary: 23 Marmot M, Bobak M. International comparators and pov-
A colour television 56 erty and health in Europe. BMJ 2000; 321:1124–8.
24 Department of Social Security. Family Resources Survey,
A refrigerator 89 Great Britain 1996–97. London: Corporate Document
A deep freeze or fridge freezer 68 Services, 1998.
A washing machine 76 25 Moore D, Hayward M. Occupational careers and mortality
A tumble dryer 20 of elderly men. Demography 1990;27:31–53.
A telephone 71 26 Sacker A, Firth D, Fitzpatrick R, et al. Comparing health
A dishwasher 7 inequality in men and women: prospective study of
A video recorder 19 mortality 1986–96. BMJ 2000;320:1303–7.
Central heating — 27 Stronks K, van de Mheen H, van den Bos J, et al. The inter-
A car or van 38 relationship between income, health and employment
“Do you not have (ITEM) because you do not want or need status. Int J Epidemiol 1997;26:592–600.
28 Mackenbach J, Kunst A, Cavelaars E, et al. Socioeconomic
one, or because you can’t aVord it?” inequalities in morbidity and mortality in western Europe.
Appendix B—Deprivation indicators Lancet 1997;349:1655–9.
“Can you look at the things listed on this card and tell me 29 Disney R, Grundy E, Johnson P. The dynamics of retirement:
which you have and which you do not have?” % in national analysis of the retirement surveys. Department of Social
sample50 considering item necessary Security Research Report, no 72. London: The Stationery
A cooked meal every day — OYce, 1997.
Meat or fish ever other day 79 30 Kitigawa EM, Hauser PM. DiVerential mortality in the United
A roast meat joint (or its equivalent) once a 56 States: a study in socioeconomic epidemiology. Cambridge,
MA: Harvard University Press, 1973.
week 31 Vagero D. Health inequalities in women and men. BMJ
A warm winter coat 85 2000;320:1286–7.
Two pairs of all weather shoes 64 32 Filakti H, Fox J. DiVerences in mortality by housing tenure
New, not second hand clothes when you need 48 and by car access from the OPCS longitudinal study. Popul
them Trends 1995;81:27–30.
Presents for friends or family once a year 56 33 Fox AJ, Goldblatt PO. Longitudinal study: Socio-demographic
Celebrations on special occasions such as 83 mortality diVerentials. London: HMSO, 1982.
Christmas 34 Moser K, Pugh H, Goldblatt P. Inequalities in women’s
A holiday away from home each year 55 health: developing an alternative approach. BMJ 1988;296:
A holiday abroad every year or so — 1221–4.
35 Menchik PL. Economic status as a determinant of black and
“Do you not have (ITEM) because you do not want it/them, white older men: does poverty kill? Population Studies 1993;
or because you can’t aVord it?” 47:427–36.
36 National Statistics. Poor housing by tenure 1996: Social
Results on perception of adult necessities from JRF Omnibus Trends Dataset. National Statistics statbase, 2000.
Survey50 37 Platt S, Martin C, Hunt S, et al. Damp housing, mould
growth and symptomatic health state. BMJ 1989;298:
1673–8.
1 Department of Health. Reducing health inequalities: an action 38 Grundy E. Moves into supported private households among
report. London: Department of Health, 1999. elderly people in England and Wales. Environment and
2 Acheson D. Independent inquiry into inequalities in health Planning A 1993;25:1467–9.
report. London: The Stationery OYce, 1998. 39 Glaser K, Murphy M, Grundy E. Limiting long-term illness
3 Cm 44445. Opportunity for all- tackling poverty and social and household structure among people aged 45 and over,
exclusion. London: The Stationery OYce, 1999. Great Britain 1991. Ageing and Society 1997;17:3–19.
4 McNiece R, Majeed A. Socioeconomic diVerences in 40 Townsend P. Poverty in the United Kingdom: a survey of house-
general practice consultation rates in patients aged 65 and hold resources and standards of living. 1st ed. Middlesex: Pen-
over: prospective cohort study. BMJ 1999;319:26–8 guin Books, 1979.
5 Blane D, Berney L, Davey Smith G, et al. Reconstructing 41 Stronks K, van de Mheen HD, Mackenbach JP. A higher
the life course: health during early old age in a follow-up prevalence of health problems in low income groups: does
study on the Boyd Orr cohort. Public Health 1999;113:117– it reflect relative deprivation? J Epidemiol Community Health
24 1998;52:548–57.
6 Fox A, Goldblatt P, Jones D. Social class mortality 42 Davey Smith G, Hart C, Watt G, et al. Individual social
diVerentials: artefact, selection or life circumstances? J Epi- class, area-based deprivation, cardiovascular disease risk
demiol Community Health 1985;39:1–8. factors, and mortality: the Renfrew and Paisley study. J
7 Arber S, Ginn J. Gender and inequalities in health in later Epidemiol Community Health 1998;52:399–405.
life. Soc Sci Med 1993;36:33–46. 43 Sloggett A, Joshi H. Higher mortality in deprived areas:
8 Marmot M, Shipley M. Do socioeconomic diVerences in community or personal disadvantage? BMJ 1998;309:
mortality persist after retirement? 25 year follow up of civil 1470–4.
servants from the first Whitehall study. BMJ 1996;313: 44 Piantadosi S, Byar DP, Green SB. The ecological fallacy. Am
1177–80. J Epidemiol 1988;27:893–904.
9 Victor C. Inequalities in health in later life. Age Ageing 1989; 45 Jacobson B, Leahy G. Health in Hackney. Annual public
18:387–91. health report on the people of City and Hackney. London: City
10 Grundy E, Glaser K. Socio-demographic diVerences in the and Hackney Health Authority, 1991.
onset and progression of disability in early old age: results 46 Bone M, Gregory J, Gill B, et al. Retirement and retirement
from a British longitudinal study. Age Ageing 2000;29:149– plans. London: HMSO, 1992.
57. 47 Grundy E, Holt G. Adult life experiences and health in early
11 MRC. The Health of the UK’s elderly people. London: Medical old age in Great Britain. Soc Sci Med 2000;51:1061–74.
Research Council, 1994. 48 OPCS. Classification of occupations and coding index. London:
12 Bartley M, Sacker A, Firth D, et al. Understanding social HMSO, 1980.
variation in cardiovascular risk factors: the advantage of 49 McClements L. The economics of social security. London:
theoretically based measures. Soc Sci Med 1999;49:831–45. Heinemann Educational Books, 1978.
www.jech.com
Downloaded from jech.bmj.com on 24 March 2008
50 Joseph Rowntree Foundation. Poverty and social exclusion in 54 Aitken M, Anderson D, Francis B, et al. Statistical Modelling
Britain. Findings 930, York: Joseph Rowntree Foundation, in GLIM. Oxford: Clarendon Press, 1989.
2000. 55 Chandola T. Social class diVerences in mortality using the
51 Blaxter M. Self definition of health status and consultation new UK national statistics socio-economic classification.
rates in primary care. Quarterly Journal of Social AVairs Soc Sci Med 2000;50:641–9.
1985;1:131. 56 OYce for National Statistics. Living in Britain: Results from
52 Idler E, Benyamini Y. Self-rated health and mortality: a the 1998 General Household Survey. London: The Stationery
review of twenty-seven community studies. J Health Soc OYce, 2000.
Behav 1997;38:21–37. 57 Grundy E, Glaser K. Trends in, and transitions to,
53 Martin J, Meltzer H, Elliot D. OPCS surveys of disability in institutional residence among older people in England and
Great Britain. Report 1: The prevalence of disability among Wales, 1971–91. J Epidemiol Community Health 1997;51:
adults. London: HMSO, 1988. 531–40.
www.jech.com
www.jech.com