I have an article and I need some to paraphrase the article in just 4 parts: method, result, discussion, and conclusion
social_capital_and_health__a_review_of_prospective_multilevel_studies.pdf
Unformatted Attachment Preview
J Epidemiol 2012;22(3):179-187
doi:10.2188/jea.JE20110128
Review Article
Social Capital and Health:
A Review of Prospective Multilevel Studies
Hiroshi Murayama1, Yoshinori Fujiwara1, and Ichiro Kawachi2
1
Research Team for Social Participation and Community Health, Tokyo Metropolitan Institute of Gerontology, Tokyo, Japan
Department of Society, Human Development, and Health, Harvard School of Public Health, Boston, USA
2
Received October 18, 2011; accepted January 26, 2012; released online March 17, 2012
ABSTRACT
Background: This article presents an overview of the concept of social capital, reviews prospective multilevel
analytic studies of the association between social capital and health, and discusses intervention strategies that enhance
social capital.
Methods: We conducted a systematic search of published peer-reviewed literature on the PubMed database and
categorized studies according to health outcome.
Results: We identified 13 articles that satisfied the inclusion criteria for the review. In general, both individual social
capital and area/workplace social capital had positive effects on health outcomes, regardless of study design, setting,
follow-up period, or type of health outcome. Prospective studies that used a multilevel approach were mainly
conducted in Western countries. Although we identified some cross-sectional multilevel studies that were conducted
in Asian countries, including Japan, no prospective studies have been conducted in Asia.
Conclusions: Prospective evidence from multilevel analytic studies of the effect of social capital on health is very
limited at present. If epidemiologic findings on the association between social capital and health are to be put to
practical use, we must gather additional evidence and explore the feasibility of interventions that build social capital
as a means of promoting health.
Key words: health; literature review; multilevel analysis; prospective study; social capital
and they facilitate certain actions of individuals who are
within the structure”.2 In addition, according to Putnam, social
capital refers to “features of social organization, such as trust,
norms and networks, that can improve the efficacy of society
by facilitating coordinated actions”.3 The existing literature
highlights 2 distinct concepts of social capital.4 The first
is that social capital represents the resources available to
members of tightly knit communities. This interpretation
could be described as the “social cohesion” definition. Social
cohesion tends to emphasize social capital as a group attribute
and analyze it as a contextual influence on individual health.
In contrast, the “network” theory of social capital defines
the concept in terms of resources that are embedded within
an individual’s social networks, that is, it is regarded as a
property of individuals.5 To date, the most common approach
to defining social capital in research on population health
has been the social cohesion perspective, ie, social capital
conceptualized as an attribute of a collective (eg,
neighborhoods, workplaces, schools). Social capital can be
broken down into a number of forms and dimensions. A
INTRODUCTION
The effectiveness and efficiency of community-based health
promotion programs vary depending on their context and
location, even when the programs have a similar design.
Such variation may be due to differences in the background
characteristics of the settings in which the interventions
are conducted. One such characteristic is “social capital,” a
concept that has been used in recent years to explain health
disparities. Social capital might provide a theoretical basis for
assessing the impact that community-based health promotion
programs have on the broader health and life of a community.1
In this article, we provide an overview of the concept of
social capital, discuss previous empirical research, and
identify intervention strategies that enhance social capital.
Definition of social capital
Coleman defined social capital as “not a single entity, but
a variety of different entities having two characteristics in
common: They all consist of some aspect of social structure,
Address for correspondence. Hiroshi Murayama, PhD, RN, PHN, Research Team for Social Participation and Community Health, Tokyo Metropolitan Institute of
Gerontology, 35-2 Sakae-cho, Itabashi-ku, Tokyo 173-0015, Japan (e-mail: [email protected]).
Copyright © 2012 by the Japan Epidemiological Association
179
180
Review of Prospective Multilevel Studies on Social Capital
SOCIAL CAPITAL
Cognitive Social Capital
Structural Social Capital
People’s perceptions of the level of
interpersonal trust, sharing, and
reciprocity
Density of social networks, or
patterns of civic engagement
Bonding Social Capital
Bridging Social Capital*
Relationships within homogeneous
groups (ie, strong ties that connect
family members, neighbors, and
close friends and colleagues)
Links between individuals/groups in
different structural positions of
power (ie, weak ties that link
different ethnic and occupational
backgrounds); can refer to links
above and below
This figure was reproduced from Islam et al.8
* We
regard linking social capital that refers to connections between individuals/groups who interact across
explicit power or authority gradients in society as a special case of bridging social capital.
Figure 1. Conceptual arrangement of social capital
common distinction in research on social capital is
between structural and cognitive dimensions.3 The structural
dimension includes externally observable aspects of social
organization and is characterized by behavioral manifestations
of network connections or civic engagement. The cognitive
dimension reflects subjective attitudes such as trust in
others and norms of reciprocity. An additional distinction
has been drawn between bonding and bridging social capital.3
Bonding social capital refers to trusting and cooperative
relations within homogeneous groups, that is, the strong ties
between members of a network who are similar in terms of
sociodemographic or social characteristics (eg, age, ethnicity,
social class). Bridging social capital describes relations
between individuals who are dissimilar with respect to
social identity and power.4,6,7 Figure 1 shows the conceptual
arrangement of social capital.
Social capital and health
Kawachi and Berkman identified at least 8 fields of
social inquiry that have examined the links between social
capital and diverse outcomes, including: (1) families and
youth behavior problems, (2) schooling and education, (3)
community life, (4) work and organizations, (5) democracy
and governance, (6) economic development, (7) criminology,
and (8) public health.9
There is considerable evidence of an association between
social capital and various indicators of health. Kawachi et al
J Epidemiol 2012;22(3):179-187
reported ecologic associations between social capital and
mortality in 39 US states.10 Their research provided strong
ecologic evidence of a relationship between state-level social
mistrust and mortality rates and between state per capita group
membership and mortality rates. In addition, individual-level
evidence has also been presented in numerous studies
(summarized in Kawachi et al11).
Although both ecologic and individual-level studies of
social capital have yielded useful insights, a proper
examination of social capital as a collective (and contextual)
influence on health requires multilevel analysis.12 In social
cohesion theory, social capital is a contextual concept.
Macinko and Starfield identified 4 analytic levels in the
association between social capital and health13: the macro
level (countries, states, regions, and local municipalities),
meso level (neighborhoods and blocks), micro level
(social networks and social participants), and individual
psychological level (trust and norm). Some researchers
have studied the relationship between social capital and
health at the macro and meso levels (ie, ecologic studies),
while others have done so at the micro and psychological
levels (individual-level studies). To examine the influence
of the contextual effect of social capital on individual
health outcomes over and above the individual effect, a
multilevel approach needs to be adopted in studies of social
capital and health. In addition, a multilevel approach enables
detailed examination of cross-level interactions, such as
Murayama H, et al.
those between community social cohesion and individual
characteristics.
A multilevel framework simultaneously examines groups
(eg, area, neighborhood) and the individuals nested within
them and offers a comprehensive framework for understanding the ways in which places affect people (contextual
effect) or, alternatively, how people can affect the groups or
places to which they belong (compositional effect). Variability
can be examined at both the group level and the individual
level, and the role of group-level and individual-level constructs can be investigated to explain variation in outcomes
among individuals and groups. Adopting a multilevel
framework implies that variations in health outcomes are
determined by both individual risk and protective factors,
as well as by community risk and resilience factors. Thus,
interventions to mitigate adverse health outcomes can be
offered at both the individual and community level.12
Multilevel analysis can be used for 2 purposes: (1) to
examine between-group and within-group variability in
outcomes and the degree to which between-group variability
is accounted for by group-level and individual-level variables
and (2) to estimate associations between group characteristics
and individual-level outcomes after adjustment for individuallevel confounders.14
Kawachi and Berkman discussed the mechanisms by which
social capital exerts a contextual effect on individual health.
They identified 4 plausible pathways: diffusion of knowledge
on health promotion, maintenance of healthy behavioral
norms through informal social control, promotion of access
to local services and amenities, and psychological processes
that provide affective support and mutual respect.9
The contextual unit used has varied across studies.
Previous studies have adopted widely varying spatial scales
as their contextual unit of analysis, ranging from whole
countries,15 states or prefectures within countries,10,16–18
local municipalities,19–21 postal code areas,22–24 small-area
neighborhoods (eg, enumeration districts, administrative
districts),25–27 companies,28,29 functional work units within
workplaces,30–32 and schools.33 However, these definitions of
contextual level suggest an important problem in multilevel
analysis. Failure to identify the correct entity at the contextual
level can result in a lack of association even when a contextual
level association is actually present. For example, the effects
of social capital on crime control were reported to vary
depending on changes in the geographic range of a
neighborhood.34 Thus, specifying the spatial scale for social
capital requires sound theory.
To explain the mechanism by which social capital
influences health, it is essential to establish a causal
relationship between social capital and health. Impediments
to causal inference include the possibility of reverse causality,
ie, good health may be a determinant of social capital rather
than the reverse.35 Identifying a causal relationship between
social capital and health would contribute to the development
181
of intervention strategies. Prospective data analysis is an
established method to improve causal inference (versus crosssectional studies). Of course, specifying the correct temporal
sequence between exposure (social capital) and outcome
(health) is only the first step. Additional obstacles to causal
inference need to be addressed, including confounding by
omitted variables at both the individual and group levels.
During the course of our systematic review, we identified
several prospective studies that examined the influence of
social capital on health outcomes, including studies on
mortality (including suicide),36–38 self-rated health,39,40 and
depression.41–44 In general, these studies show a protective
effect of social capital on adverse health outcomes. However,
many of these prospective studies only examined individuallevel associations between perceptions of social capital and
health outcomes.
As mentioned above, a multilevel approach is an effective
tool when using prospective data to accumulate robust
evidence of an association between social capital and health.
We reviewed prospective multilevel analytic studies
to investigate the association between social capital and
health.
METHODS
We used the PubMed database to conduct a systematic search
of peer-reviewed studies published up to 31 August 2011. The
following keywords were used in the search: [“social capital”
OR “social cohesion” OR “collective efficacy”], [“health”],
[“multilevel” OR “contextual effect”], and [“prospective” OR
“longitudinal” OR “cohort study”]. The keywords were
combined in the searches. We mainly included studies that
examined the direct contextual association between social
capital and health. When the searches were completed, we first
reviewed the title, keywords, and abstracts. If this initial
review suggested that the study was relevant, we then
reviewed the full text of the article for final selection.
Articles published in languages other than English were
excluded.
RESULTS
These search strategies identified 13 articles suitable for
review. The Table shows the sources and characteristics of
the 13 reviewed articles, including study country and setting,
year of survey, study subjects, measure of social capital,
outcome variables, analytic strategy, and main findings. Most
of the articles were from northern Europe (Finland and
Sweden: 8 articles). The study setting was divided into
2 types: community (9 articles) and workplace (4 articles). We
found several definitions of the analytic unit of contextual
effect: neighborhood was defined by, for example, ZIP
code area, electoral ward, administrative area,23,24,45–49
municipality50 or state,51 and functional work unit.30–32,52
J Epidemiol 2012;22(3):179-187
Community
Community
Community
New
Zealand
Finland
Sweden
Blomgren
et al46
Lofors and Sweden
Sundquist47
United
Kingdom
Blakely
et al45
J Epidemiol 2012;22(3):179-187
Mohan
et al48
Sundquist
et al49
Community
Community
Baseline: 1997
Follow-up: until 1999
Baseline: 1984–1985
Follow-up: until 2001
Baseline: 1997
Follow-up: until 1999
Baseline: 1990
Follow-up: until 1999
Baseline: 1996
Follow-up: until 1999
Baseline: 1993
(individual-level),
1994–1995 (contextual
social environment), and
1999 (contextual
socioeconomic status)
Follow-up: until 1999
Community
United
States
Wen
et al24
Baseline: 1998–1999
Follow-up: 2003
Year of survey
Community
United
Kingdom
Snelgrove
et al23
Setting
Country
Author
Self-rated health
Outcome
Unpaid voluntary activities outside the
respondent’s home over 4 weeks (4 items)
at neighborhood level and regional level
(aggregated). (No individual-level social
capital variable was used.)
Mortality (all-cause,
cardiovascular
disease, cancer,
unintentional injury,
and suicide)
Collective efficacy (7 items; ie, mutual
All-cause mortality
help, social trust), social network density
(4 items), social support (4 items), local
organizations and voluntary associations (the
numbers of these organizations in the area)
at the ZIP code area level (aggregated).
(No individual-level social capital variable
was used.)
Social trust and civic participation
at individual level and area level
(aggregated).
Social capital
Community-dwelling residents
aged 45–74 (n = 2 805 679)
in 9667 small administrative area
units (neighborhood areas)
Community-dwelling adults
(n = 7578) in 9667 small
administrative area units (ie, electoral
wards and neighborhood areas)
Entire Swedish population aged
25–64 (n = 4 516 787) in 9120
neighborhood units
First hospitalization
for psychosis or
depression
The proportion of people in the neighborhood
who voted in the 1998 local government
elections as neighborhood-level linking social
capital. (No individual-level social capital
variable was used.)
Key findings
Low family cohesion and high voter turnout in the
region were associated with alcohol-related
mortality, and the independent effects of these
remained after adjustments for individual
sociodemographic characteristics and area-level
characteristics (proportion of unemployment,
median household income, Gini coefficient, etc).
Multilevel
Poisson
regression
analysis
Continued on next page:
Low linking social capital was associated with
hospitalization for CHD in both men and women,
after adjustment for sociodemographic
characteristics.
Multilevel logistic Lower proportions of engagement in activities in the
neighborhood area were associated with mortality,
regression
but the others did not produce conclusive contextual
analysis
associations with mortality, after adjustment for age,
sex, and health-related behaviors.
Multilevel logistic Low voter participation in neighborhoods was
associated with hospitalization for psychosis in both
regression
men and women, but not with hospitalization for
analysis
depression, after adjustment for individual
sociodemographic characteristics and
neighborhood-level deprivation.
There was no significant association of
neighborhood- or regional-level social capital with
any cause of death after adjustment for
sociodemographic characteristics.
Contextual collective efficacy had a protective effect
on mortality, whereas community social network
density was detrimental. Social support, local
organizations, and voluntary associations did not
affect mortality after adjustments for
sociodemographic characteristics and health status
at the baseline.
Multilevel
Poisson
regression
analysis
Multilevel Cox
proportional
hazards model
Multilevel logistic High individual and area social trust were inversely
regression
associated with poor self-rated health, but civic
analysis
participation was not associated with individual or
area levels after adjustment for sociodemographic
characteristics and health-related behaviors.
Analysis
First hospitalization Multilevel logistic
for a fatal or nonfatal regression
coronary heart
analysis
disease (CHD) event
Engagement in activities (5 items), voting in All-cause mortality
the last election (1 item), sense of community
(5 items), social network (2 items): the
proportions of these in the area were used as
real indicators of social capital.
(No individual-level social capital variable
was used.)
Mean voting participation at neighborhood
unit-level. (No individual-level social capital
variable was used.)
Community-dwelling males aged
Family cohesion (proportion of persons living Alcohol-related
25–64 (n = 6 516 066) in 84 functional alone, of persons divorced by 1993 who were mortality
regions
married in 1990, and of 1-parent families
from all families with children) and civic
participation (voting turnout) in the regions.
(No individual-level social capital variable
was used.)
Community-dwelling residents
aged 25–74 (4.75 million person
years) in 1683 census area units
(neighborhood area) in 73 regions
Patients newly diagnosed in 1993
with 1 of 13 serious illnesses
(eg, acute myocardial infarction,
congestive heart failure, central
nervous system) (n = 12 672)
in 51 ZIP code areas in Chicago
Community-dwelling residents
(n = 3075) in 250 postcode sectors
Study subjects
Table. Prospective multilevel analytic studies of the association between social capital and health
182
Review of Prospective Multilevel Studies on Social Capital
Community
Workplace
Workpla …
Purchase answer to see full
attachment
