International Journal of Epidemiology, 2014, 1895–1920 doi: 10.1093/ije/dyu212 Advance Access Publication Date: 3 November 2014 Original article
Original article
Social capital, mortality, cardiovascular events and cancer: a systematic review of prospective studies Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Minkyoung Choi,1† Marco Mesa-Frias,1,2†** Eveline Nu¨esch,1 James Hargreaves,2 David Prieto-Merino,1 Ann Bowling,3 G Davey Smith,5 Shah Ebrahim,1,4 Caroline E Dale1* and Juan P Casas1,6 1
Department of Non-communicable Disease Epidemiology, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK, 2Department of Social and Environmental Health Research, Faculty of Public Health and Policy, London School of Hygiene and Tropical Medicine, London, UK, 3Faculty of Health Sciences, University of Southampton, Southampton, UK, 4South Asia Network for Chronic Disease, Public Health Foundation of India, New Delhi, India, 5 MRC Centre for Causal Analyses in Translational Epidemiology, School of Social and Community Medicine, University of Bristol, Bristol, UK and 6Department of Epidemiology and Public Health, University College London, London, UK *Corresponding author. E-mail:
[email protected] †
These two authors contributed equally to this work. **At the time of the study Marco Mesa-Frias was at the London School of Hygiene and Tropical Medicine. Accepted 25 September 2014
Abstract Background: Social capital is considered to be an important determinant of life expectancy and cardiovascular health. Evidence on the association between social capital and all-cause mortality, cardiovascular disease (CVD) and cancer was systematically reviewed. Methods: Prospective studies examining the association of social capital with these outcomes were systematically sought in Medline, Embase and PsycInfo, all from inception to 8 October 2012. We categorized the findings from studies according toseven dimensions of social capital, including social participation, social network, civic participation, social support, trust, norm of reciprocity and sense of community, and pooled the estimates across studies to obtain summary relative risks of the health outcomes for each social capital dimension. We excluded studies focusing on children, refugees or immigrants and studies conducted in the former Soviet Union. Results: Fourteen prospective studies were identified. The pooled estimates showed no association between most social capital dimensions and all-cause mortality, CVD or cancer. Limited evidence was found for association of increased mortality with social participation and civic participation when comparing the most extreme risk comparisons. C The Author 2014; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association V
1895
1896
International Journal of Epidemiology, 2014, Vol. 43, No. 6
Conclusions: Evidence to support an association between social capital and health outcomes is limited. Lack of consensus on measurements for social capital hinders the comparability of studies and weakens the evidence base. Key words: Social capital, systematic review, mortality, cardiovascular disease, cancer
Key Messages • Meta-analysis of the association between social capital and mortality is hindered by heterogeneity of measurement
between studies. New prospective studies adequately powered to detect more realistic effect estimates are needed. • Future studies should seek to give greater consideration to instruments of social capital and appropriate modelling
approaches. In particular, future studies should evaluate pathways and mechanisms by which social capital might affect health and should adjust for a broader range of confounders.
Individual determinants of health such as health-related behaviours and biological factors (e.g. blood pressure and cholesterol) are well-known risk factors of mortality and major chronic diseases.1–12 As a determinant of these risk factors,13,14 individual socioeconomic position is also considered an important risk factor.15,16 Successful modification of some of these factors such as smoking, blood pressure and cholesterol, as well as widespread access to medical interventions for acute treatment have contributed to a reduction by almost half in risk of premature mortality occurring in middle-aged adults.17 For the past decade, however, there has been increasing interest in concepts beyond individual determinants in epidemiological studies, one of which is social capital.18 It still remains arguable who first described the notion of social capital. Pierre Bourdieu developed the concept in its present sense.19 He defined social capital as an aggregate of resources available from a social network.20 Coleman considered social capital as various entities which form some aspect of social structure and promote certain actions of individuals for achievement of certain ends.21 Putnam, based on field studies in Italy in the 1970s, defined social capital as features of social organization such as trust, norms and networks that facilitate coordinated actions.22 He also promoted the importance of social capital, claiming that ‘if you belong to no group but decide to join one, you cut your risk of dying over the next year in half. If you smoke and belong to no group, it’s a toss-up statistically whether you should stop smoking or start joining’.23 These and other claims have been critically reviewed.24 It is important to distinguish between the concepts of social capital and of social network. As suggested above, theories of social capital refer to the
community infrastructure which influences the flow of resources to strengthen or weaken social networks. On the other hand, social networks focus on concepts about the structure of relationships and their nature—measured at the ‘subjective’ individual level. Discordance in definition and units of analysis of social capital has been one of the main subjects of critiques in this area.25 Measurement of social capital varies between area-level ‘objective’ data and individual-level data, which tend to be more subjective. In addition, pathways and mechanisms of how social capital affects individual health are yet to be reliably tested and established. Nevertheless, a considerable number of studies can be found in the literature of the association between social capital and a wide range of health-related outcomes, including not only lifestyle factors,26–36 but also perceived health status37–39 and clinical events.40–44 However, most of these studies employed cross-sectional and ecological study designs which are subject to important biases. As a consequence of still ongoing debates on what social capital represents and how best to measure it, a large variety of instruments assessing multiple dimensions of social capital have been used, making it difficult to interpret association of social capital with health outcomes. We conducted a systematic review of prospective studies to evaluate the effect of social capital on three specific health outcomes: all-cause mortality, cardiovascular disease (CVD) and cancer.
Methods Literature search The following databases were used for searching articles published up to 8 October 2012: Embase (from 1980),
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Introduction
International Journal of Epidemiology, 2014, Vol. 43, No. 6
Data extraction From each article meeting the inclusion criteria, the following data were extracted: sample size, range and average of age, duration of follow-up, instruments used for measurement of social capital, health outcomes, number of events, units of measurements, analysis model, level of adjustment and results. Social capital was measured across studies using different items obtained from questionnaires or secondary data sources. Some reported items consisted of several questions grouped into a single index. To manage inconsistencies in the measurements of social capital across studies, social capital items were categorized into seven ‘dimensions’: social participation, social network, civic participation, social support, trust, norm of reciprocity and sense of community. These dimensions were obtained by adapting and modifying the social capital dimensions identified by the UK National Office for Statistics (ONS).45,46 We subdivided the ONS ‘social network and support’ dimension into two separate dimensions following the suggestion of Ziersch47 and of Hunter et al.48 We also subdivided the ‘reciprocity and trust’ dimension into ‘norm of reciprocity’ and ‘trust’, following the work with Putnam49 and Torche and Valenzuela.50 We replaced the ONS ‘social capital’ dimension of ‘views about the area’ with ‘sense of community’, to expand the concept to non-geographical
boundaries including perceived belongingness and integration.51 All relevant dimensions were identified for each item, and we allowed an item to belong to more than one social capital dimension. Disagreements between the two reviewers (M.C. and M.F.) were resolved by consensus through discussion, and external expert input (J.P.C.) was sought for any items where disagreement remained. In addition, information about pathways, mediation effects and confounders used in the analysis of the identified studies was extracted as follows: (i) whether studies conducted pathways analysis; (ii) whether studies assessed mediation effects; (iii) whether studies reported potential mechanisms in the discussion section; (iv) whether cited studies were used to report potential mechanisms; and (v) whether confounders were used in the analysis of the studies.
Data analysis We calculated the proportion of studies assessing each social capital dimension and the average number of items used for each dimension. We also calculated the proportion of studies that assessed a social capital dimension at a particular spatial or geographical unit. In a meta-analysis, a pooled estimate of the measure of effect of social capital on each health outcome was conducted per dimension using the most adjusted measures of effect. The pooled estimates were derived by the following two steps. We combined all estimates for a common dimension within each study to obtain a single measure of effect per dimension by study. These derived measures of effect from each study were then pooled for each type of risk comparison (e.g. quartiles, binary, continuous) across studies. All estimates were combined under the assumption of interchangeability between odds ratio, risk ratio and hazard ratio using random effects models to derive pooled relative risks with 95% confidence interval (CI). Heterogeneity between studies was quantified using the tau squared statistic. We assessed the potential influence of small-study effects using funnel plots of log relative risks on the x-axis against their standard errors on the y-axis.52 Funnel plots were enhanced by contours to distinguish publication bias from other causes of asymmetry.53
Results Selection of studies A non-specific search on PubMed using the keyword ‘social capital’ yielded 1297 articles of which 14.3% were commentaries, reviews or editorials. The notion of social capital has steeply increased in popularity since the year
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Medline (from 1946) and PsycInfo (from 1806). The search strategy is shown in Supplementary Data (Appendix, available at IJE online). Only articles published in English were included. We identified studies satisfying the following inclusion criteria: (i) original article (commentaries, reviews and editorials were excluded); (ii) prospective study design; (iii) social capital as a covariate; and (iv) studies including at least one of the following outcomes: all-cause mortality, fatal or non-fatal cardiovascular disease (CVD) and cancer. We excluded studies focusing only on health outcomes in children as we wanted to focus on the potential role of social capital in common non-communicable diseases of adulthood. We also excluded articles focusing on selective social groups such as refugees and immigrants. Studies conducted in the former Soviet Union were also excluded as it was hypothesized that the political history in that region may influence the construct of social capital in ways not directly comparable with other regions. Titles and abstracts were screened first, and then the full text was assessed. The selection process was independently conducted by M.C. and M.F. and disagreements were resolved by consensus through discussion with external expert input J.P.C. sought where any disagreement remained.
1897
1898
International Journal of Epidemiology, 2014, Vol. 43, No. 6
2000 (Figure 1), corresponding to the time when Putnam’s book Bowling Alone was published. The specific search strategy from our systematic review (see Appendix, available as Supplementary data at IJE online) yielded a total of 308 articles of which 65% were original articles. A large proportion of the original articles were excluded mainly due to outcomes not under study in this paper (e.g. mental health, self-rated health, infant and child mortality, healthrelated behaviours, obesity/overweight, suicide, etc.) and study design (non-prospective studies) during the first step of the screening process. Four studies were found to originate from the former Soviet Union and would also have been excluded on the basis of inappropriate outcome and study design. Out of 44 full-text articles assessed, 14 articles were included in the systematic review (Figure 2). The other 30 full-text articles were excluded for the following reasons: non-prospective study design (11), no specific outcome (12), no relevant exposure (5) and duplicates (2).
Study characteristics The characteristics of included studies are shown in Table 1. The majority of studies were conducted in European countries (seven studies), followed by the USA (three studies). No studies in low-middle income countries were identified. All except one study54 (from which the number of participants was derived) reported the number of participants, which ranged between 6789 and 11 037 640 with a mean of 1 278 366. The majority of studies reported results for both men and women. The follow-up periods ranged from 2 to 35 years with a mean 11-year follow-up. Outcomes assessed in each study were as follows: nine studies focused on all-cause mortality, nine studies focused on fatal or non-fatal cardiovascular
events and three studies focused on cancer mortality. Only three studies out of 1454–56 assessed all the three outcomes.
Social capital items, dimensions and level of measurements All questionnaire items and data used for measurement of social capital in each study are shown in Table 2. None of the studies evaluated all seven social capital dimensions. The most frequently assessed social capital dimensions across studies were social participation (75%) and civic participation (67%), whereas the least-assessed dimension was sense of community (33%). Studies adopted different levels of measurement of social capital as the unit of analysis. The levels of measurement were categorised into two broad types, individual-level and area-level. Five studies57–61 measured social capital at individual level using primary data only, and another five studies54,56,62–64 measured social capital at area level using secondary data. Two studies used mixed measures.55,65 Among the seven studies that measured social capital at the area-level, only four studies used hierarchical statistical modelling.54,63–65 Studies assessed social capital dimensions at different levels of exposure and using different geographical units. Trust was analysed mainly at the individual level (66.6%), whereas civic participation was usually analysed as an area-level exposure (75%). The geographical units used at the area-level exposure varied considerably in size and included: neighbourhood and census areas, small-area market statistics (SAMS), electoral wards, municipalities, ZIP code areas and counties. There was considerable variability between individual instruments of measurement of each social capital dimension across studies with the exception of two similar questions used for trust: ‘Generally, you can trust other
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Figure 1. Total number of articles obtained based on a search (1979–2013) using the term ‘social capital’ (including reviews and editorials).
International Journal of Epidemiology, 2014, Vol. 43, No. 6
1899
people?’57 and ‘Generally speaking, would you say that most people can be trusted?’59 Social participation and civic participation were measured mainly using questionnaire responses assessing involvement in group activities which differed due to the variable inclusion of a wide range of activities. Social network was assessed by measuring frequency or range of informal social interaction, and these questions varied according to the types of interaction and relationship. A greater variety of instruments for measurement was found for the social support, norm of reciprocity and sense of community dimensions (Table 2).
Pathways Pathways or underlying mechanisms were discussed in 10 studies50–53,55,59–63 from which social capital was
hypothesized to influence mortality and health outcomes. Pathways and mechanisms were theorized into four broad themes: (i) access to health care and local services; (ii) health risk behaviours and socioeconomic position (SEP); (iii) psychosocial processes (e.g. sense of place); and (iv) dissemination of health-related information (see Table 3). Mediator effects were explicitly assessed, through multivariate regression models, in only three studies54,64,65 using the following variables: individual SES factors (employment status, car access, occupational social class and housing tenure); health-related behaviours (smoking, alcohol, diet and exercise); and personal medical history and treatment. Only one study65 found healthrelated behaviours and individual SES factors to be potential mediators between social capital and all-cause mortality.
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Figure 2. Flow diagram of inclusion of studies.
Sweden
Sweden
USA
Islam et al., 2008
Scheffler et al., 2008
Sweden
Chaix et al., 2008
New Zealand
Blakely et al., 2006 Sundquist et al, 2006
Finland
2805679
Sweden
Ali et al., 2006
Hyyppa et al., 2007
1275090
USA
Wen et al., 2005
34752
95026
7791
7217
13322
12672
7578
UK
Mohan et al, 2005
Sample size
Country
Participants of Statistic Sweden’s Survey of Living Conditions Members of Kaiser Permanente, the nation’s health care delivery system, of Northern California (35 counties) who were hospitalized for acute coronary syndrome between 1998 and 2002
Respondents to the 1999/2000 Health Survey in Scania (HSS)
Adult population in 1683 neighbourhood units (census data) All Swedish men and women aged 45–74 years excluding individuals who were hospitalized for CHD between 1992 and 1997 The Mini-Finland Health Survey participants in 1978/80
Respondents to the 1985 English Health and Lifestyle Survey (HALS) nested within 396 electoral wards and 198 Parliamentary constituencies Medicare beneficiaries living in Chicago (51 ZIP code areas), who were hospitalized for one of 13 serious diseases (AMI, CHF, stroke, CNS /colorectal / head or neck / liver or biliary / lung / pancreatic / urinary cancer, leukaemia, lymphoma, hip fracture) Respondents to the public-health survey in Scania born between 1919 and 1981
Study subjects
30–85
20–84 (48 [18.02])
45–81 (60.5 [10.1])
30–99
45–74
25–74
18–80
67þ (78.6 [7.2])
18þ
Range of (average [SD]) age (years)
1998
1980
2000
1978/80
1998
1996
1999/2000
1993
1984/85
Baseline (year)
Cox proportional hazard models
Multilevel Poisson regression Multilevel logistic regression
Cox proportional hazard models
3
3
24
5
21
3
Multilevel randomintercept Weibull survival models Cox proportional hazard models Multilevel Cox proportional hazards models
Cox proportional hazard models
6
2
Multilevel logistic modelling
Analysis model
16
Follow-up (years)
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Author, year
Table 1. General characteristics of prospective studies associating social capital with mortality or cardiovascular outcomes
(Continued)
All-cause and cardiovascular mortality Acute myocardial infarction (AMI) mortality All-cause and causespecific mortality Fatal and non-fatal recurrence of acute coronary syndrome
Fatal & non-fatal first time acute myocardial infarction (MI) All-cause and causespecific mortality Fatal/non-fatal coronary heart disease (CHD)
All-cause mortality
All-cause mortality
Primary outcome
1900 International Journal of Epidemiology, 2014, Vol. 43, No. 6
Netherlands
Japan
USA
Finland
USA
van Hooijdonk et al., 2008
Aida et al., 2011
Clark et al., 2011
Oksanen et al., 2011
Veenstra and Patterson, 2012
6157
28043
5789
13310
11037640
Sample size
3507 neighbourhoods in The Netherlands (neighbourhoods with a small sample or missing data on neighbourhood socioeconomic level were excluded) Participants of the Aichi Gerontological Evaluation Study (AGES) Project (community-dwelling elderly people living in 6 municipalities) Participants of the Chicago Health and Aging Project (CHAP), who were drawn from residents aged 65 in 82 census block groups in south Chicago between 1993 and 1996 Participants of the Finnish Public Sector Study (all employees working in the service of 10 towns and 21 hospitals) Participants of the Alameda County Health and Ways of Living Study who were 21 years of age or older in 1965
Study subjects
21þ
20–66 (45.9 [8.2])
65þ (74.72 [6.99])
1965
2005
1996
2003
1995
all ages
65þ
Baseline (year)
Range of (average [SD]) age (years)
Cox proportional hazard models
Cox proportional hazard models
11
5
Discrete-time hazard models
Cox proportional hazard models
5
35
Poisson regression models
Analysis model
6
Follow-up (years)
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
NR, not reported; CHF, Coronary Heart Failure; CNS, Central Nervous System.
Country
Author, year
Table 1. Continued
All-cause mortality
All-cause mortality
Stroke mortality and incidence (fatal and non-fatal)
All-cause mortality
All-cause and causespecific mortality
Primary outcome
International Journal of Epidemiology, 2014, Vol. 43, No. 6 1901
1. ‘Do you feel you are part of the community?’ 2. ‘Do you have people that you can rely upon no matter what happens?’ 3. ‘How often do you feel lonely?’
Mohan et al., 2005
Sense of community Social support Social support
Social participation
Social participation
Individual-level questionnaire item / binary Individual-level questionnaire item / binary Individual-level-questionnaire item / binary
Area-level: electoral warda / ordinal: quartile of % engaged in voluntary activity Area-level: electoral warda / ordinal: quartile of % who are ‘core’ volunteers
Social participation
Civic participation
Civic participation
Area-level: electoral warda / ordinal: quartile of % engaged in altruistic activity
Area-level: electoral warda / ordinal: quartile of % engaged in political activity Area-level: electoral warda / ordinal: quartile of % who voted in last election Area-level: electoral warda / ordinal: quartile of % who think local friends are important Area-level: electoral warda / ordinal: quartile of % who belong to the neighbourhood Area-level: electoral warda / ordinal: quartile of % who would work to improve the neighbourhood Area-level: electoral warda /ordinal: quartile of % who talk to neighbourhood Area-level: electoral warda / ordinal: quartile of % who frequently meet local people
Social networks
Social networks
Sense of community
Sense of community
Sense of community
Social participation
Area-level: electoral warda / ordinal: quartile of % engaged in social activity
• ‘often’ or ‘always’ • ‘never’ or ‘only sometimes’
Dimension of social capital
Level of measurements / variable units
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
14. (Item from BHPS) ‘Do you frequently meet people locally?’
13. (Item from BHPS) ‘Do you talk regularly to neighbours?’
11. (Item from BHPS) ‘Do you feel that you belong to the neighbourhood?’ 12. (Item from BHPS) ‘Are you willing to work with others to improve the neighbourhood?’
10. (Item from BHPS) ‘Do you think that local friends are important?’
4. Voluntary activity: (item from GHS) ‘participating in any voluntary activity over last year (non-political or trade-union)’ 5. Voluntary activity, ‘core’: (item from GHS) ‘participating in voluntary activity for 11 days or more over past year (non-political or trade-union)’ 6. Social activity: (item from BHPS) ‘active in 2 or more social activities (including parents’ association, tenants’ group, religious group, voluntary group, other community group, social or sports club, women’s institute)’ 7. Altruistic activity: (item from BHPS) ‘active in 2 or more altruistic activities (including tenants’ group, religious group, voluntary group, other community group and women’s institute)’ 8. Political activity: (item from BHPS) ‘active in political party, trades union or an environmental group’ 9. (Item from BHPS) ‘Did you vote in the last general election?’
Social capital measure (items)
Author, year
Table 2. Measure items, levels and dimensions of social capital
(Continued)
1902 International Journal of Epidemiology, 2014, Vol. 43, No. 6
Wen et al., 2005
Author, year
1. Collective efficacy: (items from the PHDCN-CS) (1) ‘People around here are willing to help their neighbours’, (2) ‘This is a close-knit neighbourhood’, (3) ‘People in this neighbourhood can be trusted’, (4) ‘People in this neighbourhood generally don’t get along with each other’, (5) ‘People in this neighbourhood do not share the same values’, (6) ‘You can count on adults in this neighbourhood to watch out that children are safe and don’t get in trouble’ and (7) ‘People in their neighbourhood would intervene if a fight broke out in front of their house’. 2. Social network density: (items from the PHDCN-CS) (1) ‘How often do you and people in this neighbourhood have parties or other get-togethers where other people in the neighbourhood are invited?’ (2) ‘How often do you and other people in this neighbourhood visit in each other’s homes or on the street?’ (3) ‘How many relatives or in-laws do you have in the neighbourhood?’ and (4) ‘How many friends do you have in the neighbourhood?’ 3. Social support: (items from the PHDCN-CS) (1) ‘How often do you and other people in the neighbourhood ask each other advice about personal things such as child-rearing or job openings?’ (2) ‘How often do you and people in your neighbourhood do favours for each other?’ (3) ‘When a neighbour is not at home, how often do you and other neighbours watch over their property?’ (4) ‘If I were sick I could count on my neighbours to shop for groceries for me?’ 4. Local organizations: (items from the PHDCN-CS) the number of organizations and programmes in the neighbourhood (e.g. a community newspaper, block group or tenant association, crime prevention programme, alcohol/drug treatment programme, mental health centre or family health service) Area-level: ZIP code areasa / continuous: one single index
Area-level: ZIP code areasa / continuous: one single index
Area-level: ZIP code areasa / continuous: one single index
Civic participation
Social support
(Continued)
Social network and social participation
Social network, social and civic participation, trust, and norm of reciprocity
Civic participation
Sense of community
Area-level: electoral warda / ordinal: quartile of % who feel the local area is friend Area-level: electoral wardb / ordinal: quartile of standardised blood donor-ship ratio Area-level: ZIP code areasa / continuous: one single index
15. (Item from SEH) ‘Do you feel that the local area is friendly?’
16. Age- and sex standardized blood donation rate
Dimension of social capital
Level of measurements / variable units
Social capital measure (items)
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Table 2. Continued
International Journal of Epidemiology, 2014, Vol. 43, No. 6 1903
Sundquist et al., 2006
Blakely et al., 2006
Ali et al., 2006
Author, year
Trust
Individual-level questionnaire item / binary
Volunteerism participation in the following unpaid voluntary activities in the past 4 weeks; 1. unpaid training, coaching, teaching, giving advice or counselling, helping at school, etc. 2. attending committee meeting, or organization, administration, policy work, etc., unpaid, for group, church, or marae (Maori tribal meeting place) 3. unpaid fund-raising work, selling, etc. for group, church, or marae 4. other unpaid work for a group, church, or marae Voting rate in the 1998 local government elections
3. Combination of social participation and trust
Social participation
Individual-level questionnaire item / binary
Civic and social participation
Civic participation
Area-level: census area unita / ordinal (quintiles)
Area-level: small-area market statistics (SAMS)b / ordinal:
• high
• low • medium
• highest
• medium • medium-high
• lowest • medium-low
• low SP & high trust • low SP & low trust
• high SP & high trust • high SP & low trust
Social participation and trust
Individual-level questionnaire item / categorical
• low: ‘do not agree at all’ or ‘do not agree’ • high: ‘agree’ or ‘completely agree’
• low: 3 • high: >3
Civic participation
Area-level: ZIP code areasa / continuous: one single index
5. Voluntary associations: (items from the PHDCN-CS) civic involvement in (1) local religious organizations; (2) neighbourhood watch programmes; (3) block group, tenant associations or community councils; (4) business or civic groups; (5) ethnic or nationality clubs; and (6) neighbourhood ward groups or local political organizations 1. Social participation: ‘In the previous 12 months, how many of the following activities have you been involved in: study circle / course at workplace, other study circle / course, union meeting, meeting of other organizations, theatre/cinema, arts exhibitions, church, sports event, letter to the editor of a newspaper/journal, demonstration, nightclub / entertainment, large gathering of relatives, and private party?’ 2. Trust: ‘Generally, you can trust other people’
Dimension of social capital
Level of measurements / variable units
Social capital measure (items)
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Table 2. Continued
(Continued)
1904 International Journal of Epidemiology, 2014, Vol. 43, No. 6
1. Leisure social participation: ‘frequency of engagement in clubs and voluntary societies, cultural and sports attendance (visiting theatre, cinema, concerts, art exhibitions, sports events, or similar), congregational activity (service attendance and/or other congregational events), outdoor and productive activity (hiking, hunting, fishing, gardening or similar), hobby activity (drama, singing, photography, painting, collecting, handicraft or similar), studying and cultural interests (reading books, listening to recorded music)’ 2. Interpersonal trust: ‘Are you satisfied with the trustful relationships with your family relationships / close friends?’ 3. Residential stability: ‘migration from another municipality to the current home municipality; 1 year or longer in the current home municipality / less than 1 year’ 1. Perception of social cohesion: (item from HSS) ‘Do you feel there is a strong cohesion in your neighbourhood?’ (authors’ definition of social cohesion: social relationships characterized by extended and interconnected networks of neighbours, a deep and shared feeling of attachment to the neighbourhood, mutually supportive relationships, and capacities to intervene collectively on behalf of the common good) 2. Social participation: participation in study circle / course; union meeting; meeting of other organizatons; theatre/cinema; art exhibition; church; sports event; letter to the editor of a newspaper/journal; demonstration; nightclub/entertainment; large gathering of relatives; and private party 1. Election participation: the average turn-out for all municipal elections (1982, 1985, 1991 and 1994), % 2. Crime rates: the average number of crimes per 1000 population for all municipalities (1980, 1986 and 2000) 3. Educational qualification: years of schooling
Hyyppa et al., 2007
Islam et al., 2008
Chaix et al., 2008
Social capital measure (items)
N/A
Social networks, reciprocity and sense of community
Individual-level questionnaire item / not reported
Area-level: neighbourhood/ ordinal quartile
Trust
N/A
Individual-level questionnaire item / categorical
Civic participation
Area-level: municipalb / continuous
Area-level: municipalb / continuous
• low: < 4 activities • high: 4
(Continued)
Trust
Individual-level questionnaire item/ not reported
Social participation
Social participation
Individual-level questionnaire item / not reported
Individual-level/Binary:
Dimension of social capital
Level of measurements / variable units
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Author, year
Table 2. Continued
International Journal of Epidemiology, 2014, Vol. 43, No. 6 1905
van Hooijdonk et al., 2008
Scheffler et al., 2008
Author, year
Petris Social Capital Index (PSCI): the number of individuals per 1000 population employed in voluntary organizations Social capital score: a single index derived from 13 social capital items from Housing Demand Survey (WBO) of 1998 including; 1. I talk a lot to my nextdoor neighbours 2. I talk a lot to neighbours other than my nextdoor neighbours 3. In this neighbourhood people treat each other with respect 4. People hardly know each other in this neighbourhood 5. I feel attached to my neighbourhood 6. I feel at home in my neighbourhood 7. I feel at ease with the people in my neighbourhood 8. I live in a neighbourhood with a low level of solidarity 9. I’m satisfied with the population composition of this neighbourhood 10. I’m satisfied with my living environment 11. The buildings in this neighbourhood are attractive 12. It is unpleasant to live in this neighbourhood 13. To what extent are you involved with the liveability of your neighbourhood?
7. Number of children in the household
4. Age 5. Income 6. Cohabitation status
Social capital measure (items)
• High
• III • IV
(Continued)
Social network, sense of community, and civic participation
Area-level: municipala / categorical: the mean of individual scores ¼ community score ->5 categories • low • II
Social participation
Area-level: countyb / continuous
• two • more than three
• no children • one
N/A
N/A N/A N/A
Individual-level questionnaire item / continuous Individual-level questionnaire item / continuous Individual-level questionnaire item / binary: live alone or not Individual-level questionnaire item / categorical:
• secondary education • university education
• pre-secondary education • short education
Dimension of social capital
Level of measurements / variable units
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Table 2. Continued
1906 International Journal of Epidemiology, 2014, Vol. 43, No. 6
1. Cognitive social capital-(1): ‘Generally speaking, would you say that most people can be trusted?’
Aida et al., 2011
• yes (high social capital) • no (low social capital)
Individual-level/binary:
• yes (high social capital) • no (low social capital)
Individual-level/binary:
• yes (high social capital) • no (low social capital)
Individual-level/ binary:
• yes (high social capital) • no (low social capital)
Individual-level/ binary:
• yes (high social capital) • no (low social capital)
Individual-level/ binary:
• yes (high social capital) • no (low social capital)
Individual-level/ binary:
• yes (high social capital) • no (low social capital)
Individual-level/ binary:
• no (low social capital)
• yes (high social capital) • depends
Individual-level/categorical
• no (low social capital)
• yes (high social capital) • depends
Individual-level/categorical
• no (low social capital)
• yes (high social capital) • depends
Individual-level/categorical:
Level of measurements / variable units
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
10. Structural social capital -(2): industry group participation
9. Structural social capital-(1): political group participation
8. Cognitive social capital -(8): ‘Do you look after someone when he/she is sick and stays in bed for a few days?’
7. Cognitive social capital -(7): ‘Do you listen to someone’s concerns and complaints?’
6. Cognitive social capital -(6): ‘Do you have someone who acknowledges your existence and value?’
5. Cognitive social capital -(5): ‘Do you have someone who looks after you when you are sick and stay in bed for a few days?’
4. Cognitive social capital -(4): ‘Do you have someone who listens to your concerns and complaints?’
3. Cognitive social capital -(3): ‘Would you say that most of the time people try to be helpful?’
2. Cognitive social capital -(2): ‘Do you think most people would try to take advantage of you if they got a chance?’
Social capital measure (items)
Author, year
Table 2. Continued
Civic participation
Civic participation
Norm of reciprocity
Norm of reciprocity
Social support
Social support
Social support
Trustc
Trust
Trust
Dimension of social capital
(Continued)
International Journal of Epidemiology, 2014, Vol. 43, No. 6 1907
Clark et al., 2011
Author, year
Social cohesion score: a single score derived from individual’s response to 6 items of social cohesion including: 1. How often do you see neighbours and friends talking outside in the yard or on the street? 2. How often do you see neighbours taking care of each other such as doing yard work or watching children? 3. How often do you see neighbours watching out for each other such as calling if they see a problem? 4. How many neighbours do you know by name? 5. How many neighbours do you have a friendly talk with at least once a week?
17. Structural social capital -(9): ‘How often do you meet your friend?’
16. Structural social capital -(8): a vocation group participation
15. Structural social capital -(7): neighbourhood group participation
14. Structural social capital -(6): sports group participation
13. Structural social capital -(5): religious group participation
12. Structural social capital -(4): citizen group participation
Individual-level/ binary:
11. Structural social capital -(3): volunteer group participation
(Continued)
Social network, social support and norm of reciprocity
Area-level: census block / continuous and ordinal quartile An average neighbourhood cohesion score (per unit increase means higher cohesion score)
• rarely • having no friends
• once or more/month • several times/year
Social network
Civic participation
Civic participation
Social participation
Social participation
Civic participation
Social participation
Dimension of social capital
Individual-level/ categorical:
• yes (high social capital) • no (low social capital)
Individual-level/ binary:
• yes (high social capital) • no (low social capital)
Individual-level/binary:
• yes (high social capital) • no (low social capital)
Individual-level/binary:
• yes (high social capital) • no (low social capital)
Individual-level/ binary:
• yes (high social capital) • no (low social capital)
Individual-level/ binary:
• yes (high social capital) • no (low social capital)
Level of measurements / variable units
Social capital measure (items)
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Table 2. Continued
1908 International Journal of Epidemiology, 2014, Vol. 43, No. 6
6. Belonging to a group concerned with community betterment, charity, or service
4. Belonging to a labour union, commercial group, or a professional association 5. Belonging to a group concerned with children
3. Belonging to a social or recreational group
2. Frequency of church attendance
6. How many neighbours could you call on for assistance in doing something around your home or yard or to ‘borrow a cup of sugar’ or some other small favour? Workplace social capital score: a single score derived from individual’s/co-worker’s responses to8 items of workplace social capital including; 1. ‘Our supervisor treats us with kindness and consideration.’ 2. ‘Our supervisor shows concern for our rights as an employee.’ 3. ‘We have a ‘we are together’ attitude.’ 4. ‘People keep each other informed about work-related issues in the work unit.’ 5. ‘People feel understood and accepted by each other.’ 6. ‘Do members of the work unit build on each other’s ideas in order to achieve the best possible outcome?’ 7. ‘People in the work unit cooperate in order to help develop and apply new ideas.’ 8. ‘We can trust our supervisor.’ 1. The number of close friends
Social capital measure (items)
Civic participation
Civic participation
Civic participation
Social participation
Social participation
Social network
Social support, sense of community, norm of reciprocity, and trust
An average neighbourhood cohesion score (per unit increase means higher cohesion score)
Individual-level/ binary: having 3 or more friends or not Individual-level/ binary regularly attend or not Individual-level/binary yes or no Individual-level/binary yes or no Individual-level /binary yes or no Individual-level/ binary yes or no
Dimension of social capital
Level of measurements / variable units
b
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Area-level social capital variable derived from aggregates of individual data of other studies (secondary data). Area-level social capital variable derived from secondary data measured at area level. c Considered the possibility of grouped questions, it was decided to categorize the item into ‘trust’ dimension as the authors did. SEH, Survey of English Housing; PHDCN, Project on Human Development in Chicago Neighborhood; WBO, Dutch Housing Demand Survey; BHPS, British Health Panel Survey; GHS, General Household Survey; SP, Social Participation; BMI, Body Mass Index; SBP, Systolic Blood Pressure.
a
Veenstra and Patterson, 2012
Oksanen et al., 2011
Author, year
Table 2. Continued
International Journal of Epidemiology, 2014, Vol. 43, No. 6 1909
‘Social capital has been hypothesized to be associated with health by four causal pathways including psychosocial mechanisms, the norms and values influencing health related behaviours, access to health care, and crime, particularly violent crime.’
‘Potential pathways from social capital to health include its influence on behaviour (e.g. by more rapid diffusion of health information), facilitation of access to services and amenities (e.g. more socially cohesive neighbourhoods may be more successful at securing services), or psychosocial processes (e.g. sense of place).’
‘There are some possible mechanisms behind our finding that low linking social capital is associated with CHD. Increased electoral participation could be the result of a mobilization of less wealthy people who, when mobilized, vote for politicians and political parties more prone to support the welfare state or, in the case of the USA, different welfare programmes.’ ‘Participation in elections is central to achieve empowerment. In other
Ali et al., 2006
Blakely et al., 2006
Sundquist et al., 2006
‘(1) the severity of the AMI: greater in low-cohesion neighbourhoods (2) acute triggers of coronary plaque rupture: exist in residential environments with inharmonious social interactions. Residents of low social cohesion neighbourhood may not benefit from a supportive
mal power gradient.’ ‘Moreover, it gives individuals not only a feeling of political empowerment, but also a real ability to participate effectively in relationships across gradients of power and authority and, subsequently, more control over their lives. If empowerment is a mediator between linking social capital and CHD, it is possible that stress due to feelings of powerlessness or lack of control may be the biological link between low neighbourhood linking social capital and CHD.’
individuals the opportunity to interact across an institutional or for-
of people with a high level of horizontal and vertical trust gives
Cited 27 times
Cited 63 times
Cited 46 times
Cited 22 times
Cited 90 times
Citation count Lindstrom M, Merlo J, Ostergren P.-O. Social capital and sense of insecurity in the neighbourhood: A population-based analysis in Malmo, Sweden. Soc Sci Med 2003;6:111–1120
Kawachi I, Kennedy BP, Lochner K et al Social capital, income inequality, and mortality. Am J Public Health 1997;87;1491–98 Kawachi I, Berkman L. Social cohesion, social capital, and health. In: Berkman L, Kawachi I (eds). Social Epidemiology. New York: Oxford University Press, 2000 1. Blakely TA, Kennedy BP, Kawachi I. Socioeconomic inequality in voting participation and self rated health. Am J Public Health 2001;,91:99–104 2. Szreter, S, Woolcock M. Health by association? Social capital, social theory, and the political economy of public health. Int J Epidemiol 2004,33:650–67
1. Baylin A, Hernandez-Diaz S, Kabagambe EK et al. Transient exposure to coffee as a trigger of a first nonfatal myocardial
Health-related behaviours (smoking, alcohol, diet, exercise), social class and housing tenure, level of deprivation
Leisure-time physical activity, daily smoking and BMI, education and economic stress
Household income, education, car access, and employment status, ethnicity, and marital status at individual level
Country of birth, education, marital status, and housing tenure
Marital status, previously diagnosed diabetes, hypertension, other heart disease, self-rated health, education, self-reported
(Continued)
Reference citation on mechanism
Adjustments
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Chaix et al., 2008
‘The effect of social capital was also lessened when health-related behaviours or individual material circumstances (social class and housing tenure) were included in age- and sex-adjusted models, suggesting that these factors may be part of a possible mediating pathway between social capital and health. It is, however, possible that mortality is simply not sensitive enough, as an indicator of individual health, to detect real and valid impacts of social capital.’
Mohan et al., 2005
words, living in a neighbourhood with well-functioning networks
Mechanism Explained
Article
Table 3. Pathways and underlying mechanism suggested in studies reviewed
1910 International Journal of Epidemiology, 2014, Vol. 43, No. 6
Islam et al., 2008
Article
‘At least three different pathways have been suggested by which community-level social capital is theorized to affect health. First, community-level social capital may increase the availability of information on behaviours that influence cardiovascular disease risk. Second, higher levels of community-level social capital, such as a higher density of voluntary organizations, may lower the effort required to politically organize which may result in more health resources being brought into a community. This suggests that geographical variation in medical resources and medical treatment is likely to be correlated with the geographical variation in community-level social capital. Third, higher levels of community-level social capital, such as a higher density of voluntary organizations, may make social support, which is associated with improved cardiovascular outcomes in many studies, more accessible.’
environment that would help them cope with intense worries that could trigger AMI. (3) the time of access to emergency care after AMI: neighbourhood cohesion, as a facilitator of the circulation of information between residents, may increase individuals’ recognition of AMI symptoms and awareness of the appropriate action to undertake. Residents of strongly cohesive neighbourhoods may have more opportunities to receive social support from their neighbours, which may be critical to obtaining post-AMI emergency care in due time. (4) compliance or not with post-AMI medical recommendation: relies on the hypothesis of a greater amount of social support available to the residents of socially cohesive neighbourhoods. Neighbourhoodbased social support in its instrumental, informational, appraisal and emotional forms may help people adhere to medical recommendations and prescribed treatments.’
Mechanism Explained
Cited 23 times
Citation count
1. Baylin A, Hernandez-Diaz S, Kabagambe EK et al. Transient exposure to coffee as a trigger of a first nonfatal myocardial infarction. Epidemiology 2006;17:506–11 2. Hallqvist J, Moller J, Ahlbom A et al. Does heavy physical exertion trigger myocardial infarction? A case-crossover analysis nested in a populationbased case referent study. Am J Epidemiol 2000;151:459–67 3. Moller J, Hallqvist J, Diderichsen F et al. Do episodes of anger trigger myocardial infarction? A casecrossover analysis in the Stockholm Heart Epidemiology Program (SHEEP). Psychosom Med 1999;61:842–49
Income, cohabitation status, number of children in the household, education
(Continued)
infarction. Epidemiology 2006;17:506–11 2. Hallqvist J, Moller J, Ahlbom A et al. Does heavy physical exertion trigger myocardial infarction? A case-crossover analysis nested in a populationbased case referentstudy. Am J Epidemiol 2000;151:459–67 3. Moller J, Hallqvist J, Diderichsen F et al. Do episodes of anger trigger myocardial infarction? A casecrossover analysis in the Stockholm Heart Epidemiology Program (SHEEP). Psychosom Med 1999;61:842–49
Reference citation on mechanism
financial strain, income, social participation þneighbourhood safety þneighbourhood income þpopulation density þ% from low-income countries þresidential stability þdistance to the hospital
Adjustments
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Table 3. Continued International Journal of Epidemiology, 2014, Vol. 43, No. 6 1911
‘There are several plausible pathways linking social capital to health. First, social capital may affect individual health by influencing health-
van Hooijdonk et al., 2008
Clark et al., 2011
‘The potential mechanisms through which social capital may contribute to the risk of mortality have not been fully elucidated, but at least 4 plausible pathways have been suggested: health risk behaviours, dissemination of health-related information, access to local services and amenities, and psychosocial processes.’
sion may be related to stroke mortality but not incidence.’
not stroke incidence, offering a credible explanation for why cohe-
tective relationship between social support and stroke mortality but
health, and recent longitudinal research has found a significant pro-
mechanisms linking neighbourhood conditions and individual
‘Social support and access to services have been theorized as potential
members.’
plementation of policies which ensure the security of all its
egalitarian patterns of political participation that result in the im-
effects of life events on mental health. Fourth, the communities with higher social capital produce more
distress. Social networks and social support can buffer the negative
health. Third, there are associations between social capital and psychological
portation, clinics and community health centres could improve
to local services and amenities. Good access to service such as trans-
health-related behaviours. Second, higher social capital may promote health by increasing access
health information and by exerting social control over deviant
Cited 4 times
Cited 5 times
Cited 21 times
Citation count
Berkman LF, Glass T. Social inte-
Personal medical history (CHD, hypertension, diabetes, depression, stroke, heart failure, peripheral vascular disease) and CVD treatment and medication use, median income
SBP, physical activity, smoking status, chronic conditions, history of stroke, and neighbourhood-level socioeconomic status (% on public assistance, % of households earning $25,000 per year, % with a college degree or higher and % of owneroccupied dwellings valued at $200,000)
Kawachi I, Berkman L. Social
BMI, self-rated health, current illness, smoking history, alcohol consumption, exercise, equivalent income and educational attainment
Kawachi I, Berkman L. Social cohesion, social capital, and health. In: Berkman L, Kawachi I, eds. Social Epidemiology. New York, NY: Oxford University Press, 2000
Press; 2000:174–190.
I. New York: Oxford University
Edited by: Berkman L, Kawachi
health. In: Social epidemiology.
cohesion, social capital, and
Kawachi I, Berkman LF. Social cohesion, social capital and health. In: Berkman LF, Kawachi I (eds). Social Epidemiology. New York, NY: Oxford University Press, 2000
Marital status and ethnicity þneighbourhood SES (% of neighbourhood inhabitants with a low income) þ urbanicity
University Press, 2000
New York, NY: Oxford
I (eds). Social Epidemiology.
health. In Berkman LF, Kawachi
gration, social networks, and
Reference citation on mechanism
Adjustments
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Aida et al., 2011
‘The way in which community social capital might be related to health is still debated. However, at least three plausible pathways are suggested: (i) through health-related behaviours, (ii) through access to services and amenities and (iii) through psychosocial processes.’
Scheffler et al., 2008
related behaviours through promotion of more rapid diffusion of
Mechanism Explained
Article
Table 3. Continued
1912 International Journal of Epidemiology, 2014, Vol. 43, No. 6
International Journal of Epidemiology, 2014, Vol. 43, No. 6
Associations of social capital with health outcomes
Discussion In this systematic review, we identified 14 published studies that examined the association between social capital and all-cause mortality, CVD and cancer using prospective data.66–79 The pooled results of the individual studies for each dimension showed only limited evidence of association. Other reviews of prospective multilevel studies of social capital have also found very limited evidence.80,81 In this paper we identify more studies related to the outcomes of social capital and allcause mortality and CVD and, where possible, provide synthesized risk estimates by social capital dimension. Estimates in the most extreme risk comparisons suggested that indicators of social participation and civic participation were associated with some beneficial effects on all-cause mortality and cardiovascular mortality, respectively. In the studies identified, findings for other dimensions did not support the other aspects of social capital as an effective predictor of mortality, CVD or cancer.
Despite the growing number of publications on social capital found in PubMed (Figure 1), few studies have investigated the effects of social capital on mortality, CVD and cancer in a prospective manner. Social capital has only recently increased in popularity in public health, and this could be one reason for the small number of prospective studies. Most studies in this review used data collected before the notion of social capital began to acquire its current popularity. As a result, authors tended to use data obtained for other purposes rather than using de novo variables designed for measuring social capital. The diversity of instruments for social capital measurement found across the studies may reflect the limitation of accessible data. This limitation could generate the impression that some dimensions are more related to the health outcomes than others. Our pooled estimates are dominated by the predominance of the two dimensions, civic participation and social participation. As such, it may be too early to emphasize promoting civic and social participation as a means of preventing premature mortality or CVD events. The limitations in the way the complex phenomenon of social capital has been operationalized in these observational studies might explain the relative weak associations. In addition, the relatively low number of incident events in the majority of studies identified is likely to limit power for metaanalysis. Evidence of the precise mechanisms occurring between social capital and health is needed. Association studies to date have not tested causal mechanisms or underlying pathways in which social capital may affect health. Most studies examined association and then referred to hypothetical pathways or mechanisms derived from ecological studies or reviews that do not actually provide evidence of any mechanism. Despite the limited evidence on pathways and mechanisms, there is a willingness by governmental organizations to adopt social capital as an effective way of improving health.82–84 The effects of social capital may differ in different cultural contexts59 and may be a consequence of other wider contextual political, social or economic processes.85 In this paper we found that the majority of the studies reviewed were conducted in Europe and North America. Therefore, the effect of social capital on the health outcomes for other populations and contexts is not known.
A way forward More theoretical studies are required to provide a concordant methodological base, on which more reliable data on social capital can be collected. Analysis of pathways and underlying mechanism is needed to increase the validity of associations and infer effective ways of modification
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Risk ratios were described using different types of risk comparison groups such as: the contrast of extreme quartiles, quintiles, highest vs lowest median categories, continuous and binary groups. All studies adjusted for age, sex and at least one individual socioeconomic position (SEP) variable (e.g. education, income, occupation or housing tenure). Only four out of 14 studies54–56,62 controlled for different area-level characteristics (e.g. levels of urbanization, neighbourhood SEP and area deprivation) as contextual factors. The summary risk ratios (RRs) for each dimension and outcome are reported in Table 4. A total of two associations with a nominal P-value lower than 0.05 were found out of 21 summary RRs that evaluated seven dimensions and three outcomes. In the most extreme comparison groups, low social participation and civic participation were suggested to be associated with all-cause mortality 1.32 (95% CI: 1.17–1.49) and CVD mortality 1.23 (95% CI: 1.14–1.33), respectively. Social support, social network, norm of reciprocity and sense of community had only marginal effects on all-cause mortality. For trust, between the highest and lowest risk comparison groups, relative risks of all-cause mortality and CVD mortality were 0.96 (0.77–1.19) and 0.80 (0.52–1.24), respectively. The asymmetry in the contour-enhanced funnel plots for all outcomes was investigated and did not show any evidence of small-study effects (Figure 3).
1913
1370116
2012 Islam et al., 2008 and Blakely et al, 2005
Cancer mortality Continuous
2805679
Wen et al., 2005
Mohan et al., 2005
Continuous
Social network All-cause mortality Quartile
Social support All-cause mortality Binary
Aida et al., 2011 and Mohan et al., 2005
7578
12672
20888
41969
et al., 2008 Scheffler et al., 2008 and Hyyppa et al., 2007
Continuous
21113
Ali et al., 2006 and Chaix
CVD mortality Binary
7217
Hyyppa et al., 2007
7578 6157
95026
95026
Continuous
Mohan et al., 2005 Veenstra and Patterson et al., 2012
Islam et al., 2008
Continuous
Social participation All-cause mortality Quartile Binary
Sundquist et al., 2006
Islam et al., 2008
CVD mortality Ordinal
NA
1.68 (0.88–3.19) 0.131
0.95 (0.92–0.99) NA
1.32 (1.17–1.49) NA 0.88 (0.77–1.00) NA
0.97(0.93–1.02)
NR
3211
1214
0.92 (0.71–1.19) NA
1.02 (0.73–1.43) NA
1.08 (0.84–1.39) 0.018
10392 0.95 (0.89–1.01) 0.001
153
2476
NR 3168
9000
NA
individual SEPa
Age, sex and
1
1
1
Q4 (ref)vs Q1
Per 1 unit score for social support
Often/always vs never/only sometimes (ref); low vs high SC (ref)
Per 1 unit score for social participation; and per 1 unit score of Petris social capital index
High (ref) vs low
Per 1 unit score for social participation
1
1
2
2
2
1
Q4 (ref) vs Q1 1 Regular vs not regular (ref); 1 yes vs no (ref)
Per 1 unit score for civic participation
Per 1 unit score for civic participation
C3 (ref) vs C1
2
0
0
0
1
1
1
0 0
1
1
1
1
0 0
status
0 0
0
0 0 1
2 2
1
1
0
0 0
0
0 0 1
1 1
1
0
1
0
0
0
1
0
1
0 0
0
0
0
0
1
0
1
1
0
0 0
1
1
0
2
(Continued)
1
0
0
0 0
0 0
1 0
0 0
healtha characteristicsa
Biological Mental Area-level
behavioursa factorsa
Cohabitation Health-
No. studies adjusted by group of confounders
Q4 (ref) vs Q1 1 Regular vs not regular (ref); 1
Risk comparison
yes vs no (ref) 0.0000.000 Per 1 unit score for civic participation
0.99 (0.94–1.04) NA
44800 1.23(1.14–1.33) 9000
t2
1.11 (0.97–1.27) NA 0.98 (0.91–1.06) NA
12211 0.99(0.95–1.04)
NR 3168
b
Relative
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Continuous
7578 6157
Mohan et al., 2005 Veenstra and Patterson,
Civic participation All-cause mortality Quartile Binary
No.
participants events risk
No.
Author, year
Dimension
Table 4. Relative risk of social capital dimension effects on all-cause mortality, CVD and cancer mortality
1914 International Journal of Epidemiology, 2014, Vol. 43, No. 6
Ali et al., 2006
Hyyppa et al., 2007 and Islam et al., 2008
0.80 (0.52–1.24) NA
1214
NR
7578
9000
NA
1.03 (0.97–1.17) NA
1.13(0.93–1.38)
1.00 (0.96–1.05) NA
11476 0.97 (0.90–1.04) 0.002
153
13310
95026
102243
13322
0.96 (0.77–1.19) NA
1.16 (0.50–2.69) NA
11476 1.01 (0.97–1.06) 0%
1214
1214
t2
Q4 (ref) vs Q1
Low vs high (ref)
Per 1 unit score for trust
Per 1 unit score for trust
High (ref) vs low
Per 1 unit score for trust
C3 (ref) vs C1
C4 (ref) vs C1
Risk comparison
1
1
1
2
1
2
1
1
0
1
1
2
0
2
1
1
1 1
0
1
0
1 1
0
0
1
1 1
0 1
1
0
0
0
0
1
0
1
0
0
0
0
1
1
0
1
0
0
Age, sex and Cohabitation HealthBiological Mental Area-level individual SEPa status behavioursa factorsa healtha characteristicsa
No. studies adjusted by group of confounders
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
a Adjustments: Individual Socioeconomic Position (SEP), including income, education, occupation and housing tenure; health related behaviours, including smoking, alcohol use, physical activity; biological factors, including diagnosed chronic disease, cholesterol, blood pressure; mental health status was adjusted only for women in all studies listed; area-level characteristics, including area-level SEP, urbanization, population density, arealevel safety, health maintenance organizations penetration. NR, not reported; NA, not applicable; Q4, highest quartile; Q1, lowest quartile; C4, highest category out of 4 categories; C3, highest category out of 3 categories; C1, lowest category; ref, reference group of comparison; CVD, cardiovascular disease. b Relative risk:wherever possible, relative risks are summary relative risks from meta-analysis of studies with common social capital dimensions and outcome. Where multiple studies with the same measures did not exist, relative risks from single studies are presented. Note: Some studies were not pooled because it was not possible to separate all dimensions as they reported different dimensions (e.g. social network, civic participation and sense of comunity) into one aggregate measure of effect (vanHooijdonk et al 2008, Oksanes et al 2011, Clark et al 2011). SC, Social Capital.
All-cause mortality Binary Aida et al., 2011 Sense of community All-cause mortality Quartile Mohan et al., 2005
Cancer mortality Continuous Islam et al., 2008 Norm of reciprocity
Continuous
CVD mortality Binary
102243
13310
Islam et al., 2008 and Hyyppa et al., 2007
Aida et al., 2011
Continuous
13310
Aida et al., 2011
Categorical Trust All-cause mortality Categorical
No. No. Relative participants events riskb
Author, year
Dimension
Table 4. Continued
International Journal of Epidemiology, 2014, Vol. 43, No. 6 1915
1916
of social capital at the community level. Even though theoretical limitations in measuring social capital are ongoing, efforts have been made to improve instruments for measuring social capital.86–89 In order to improve generalizability and comparability of the findings, more studies from more heterogeneous populations and standardized geographical units of measurement are needed. The latter are required not only for social capital, but also for other area-level exposures. This could be addressed by applying area-based methods from other fields in the analysis of health outcomes.90 Whereas we recognize that the mechanisms of social capital are still subject to ongoing debate, in order to evaluate the appropriateness of the modelling in the identified papers we refer to a directed acyclic graph (DAG) based on the mechanisms proposed by Kawachi and Berkman,91 by which social capital may influence health (Figure 4). According to this DAG, it is clear that the modelling approach of the majority of studies exhibits a number of limitations and could be improved. First, the adjustment for confounders, and in particular arealevel confounders, appears incomplete with many studies only adjusting for basic characteristics of the individual. Second, several studies adjust for variables that may be mediators of the association between social capital and health as confounders (for example, health behaviours). Third, none of the studies identified in our review undertook proper mediation analysis. Some studies used regression analysis, but this approach is recognized to have substantial limitations for addressing mediators.92 Fourth, only five studies applied a multilevel approach to the modelling; this has previously been identified as important for social capital to capture the potential multidimensional nature of its effects.93 Fifth, most studies only used baseline variables as confounders where it would be more appropriate to model time-varying confounders to recognize that these may be dynamic over time. We recommend that future studies of social capital and mortality should seek to address these limitations in their study design. Given the inevitable complexities in measuring dimensions of social capital, a complementary strategy to potentially shed light on the importance of social capital mechanisms in influencing health outcomes may lie in the conduct and synthesis of randomized trials of interventions that invoke theories of social capital in their design. For example, interventions might be designed that seek to increase levels of civic and social participation in later life. Such interventions might be studied in randomized trials including measurement of the outcomes and potential mediators identified in this paper.
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Figure 3. Funnel plots of prospective studies assessing social capital with health outcomes.
International Journal of Epidemiology, 2014, Vol. 43, No. 6
International Journal of Epidemiology, 2014, Vol. 43, No. 6
1917
Limitations
Conclusion
This systematic review aimed to demonstrate the nature of the concept of social capital and the multiple forms of usage it has had in epidemiology. Using specific keywords, we identified studies on social capital, and the categorization we imposed on the data was to make it more tractable for presentation. Although the contour-enhanced funnel plots for all outcomes did not reveal any evidence of smallstudy effects, these are possible since this systematic review did not include articles published in other languages (19) potential. Review of the abstracts of these papers, however, showed they had inappropriate study designs and outcomes, thus their exclusion would not have introduced publication bias. Another source of bias may lie in the process of our classification of social capital items according to social capital dimensions. Since the pooling of results was conducted on the basis of each dimension, if there had been any errors in the categorization, the validity and reliability of our findings would have been affected. We tried to reduce this bias by having two researchers independently assign social capital dimensions to each social capital item, and resolved disagreements by consensus. Finally, owing to the heterogeneity of social capital measurements made, it was not possible to provide many pooled estimates.
Several different measures of social capital have been used in prospective studies assessing health outcomes. Despite the popularity of the concept of social capital in epidemiology and public health, there is no consensus on how to measure social capital. This hinders comparison and pooling of results and limits conclusions about the health effects of social capital. In the face of this limitation, we did not find strong evidence of social capital affecting all-cause mortality, CVD events or cancer, although there was limited evidence for some specific dimensions of social capital.
Supplementary Data Supplementary data are available at IJE online.
Funding This study was supported by grants from the British Heart Foundation (EPNCCD08) and the Department of Health (009/ 0049).
Acknowledgements The British Women Heart & Health Study (BWHHS) is supported by grants from the British Heart Foundation (EPNCCD08) and the
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
Figure 4. Suggested causal DAG to illustrate mechanisms by which social capital may influence non-communicable disease and mortality. Adapted from Kawachi and Berkman91.
1918 Department of Health (009/0049). The funders had no role in the study design, the collection, analysis or interpretation of data, the writing of the report or the decision to submit the paper for publication. M.C., M.F. and J.P. contributed to the study design. M.C. and M.F. led the drafting. All authors contributed to the analyses, interpretation and final manuscript drafting.
International Journal of Epidemiology, 2014, Vol. 43, No. 6
15.
16.
Conflict of interest: None declared.
17.
1. Allender S, Balakrishnan R, Scarborough P, Webster P, Rayner M. The burden of smoking-related ill health in the UK. Tob Control 2009;18:262–67. 2. Prasad DS, Kabir Z, Dash AK, Das BC. Smoking and cardiovascular health: a review of the epidemiology, pathogenesis, prevention and control of tobacco. Indian J Med Sci 2009; 63: 520–33. 3. Katanoda K, Marugame T, Saika K, et al. Population attributable fraction of mortality associated with tobacco smoking in Japan: a pooled analysis of three large-scale cohort studies. J Epidemiol 2008;18:251–64. 4. Rehm J, Baliunas D, Borges GL et al. The relation between different dimensions of alcohol consumption and burden of disease: an overview. Addiction 2010;105:817–43. 5. Balakrishnan R, Allender S, Scarborough P, Webster P, Rayner M. The burden of alcohol-related ill health in the United Kingdom. J Public Health 2009;31:366–73. 6. Adami PE, Negro A, Lala N, Martelletti P. The role of physical activity in the prevention and treatment of chronic diseases. Clin Ter 2010;161:537–41. 7. Warburton DER, Nicol CW, Bredin SSD. Health benefits of physical activity: the evidence. CMAJ 2006;174:801–09. 8. Samitz G, Egger M, Zwahlen M. Domains of physical activity and all-cause mortality: systematic review and doseresponse meta-analysis of cohort studies. Int J Epidemiol 2011;40:1382–400. 9. Virdis A, Bruno RM, Neves MF, Bernini G, Taddei S, Ghiadoni L. Hypertension in the elderly: an evidence-based review. Curr Pharm Des 2011;17:3020–31. 10. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R. Agespecific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies. Lancet 2002;360:1903–13. 11. Muldoon MF, Manuck SB, Matthews KA. Lowering cholesterol concentrations and mortality: a quantitative review of primary prevention trials. BMJ (Clinical research ed.) 1990;301:309–14. 12. Haapanen-Niemi N, Vuori I, Pasanen M. Public health burden of coronary heart disease risk factors among middle-aged and elderly men. Prev Med 1999;28:343–48. 13. Schumann B, Kluttig A, Tiller D, Werdan K, Haerting J, Greiser KH. Association of childhood and adult socioeconomic indicators with cardiovascular risk factors and its modification by age: the CARLA Study 2002–2006. BMC Public Health 2011;11:289. 14. Ramsay SE, Whincup PH, Hardoon SL, Lennon LT, Morris RW, Wannamethee SG. Social class differences in secular trends in established coronary risk factors over 20 years: a cohort study
18. 19. 20. 21. 22.
23. 24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
References
of British men from 1978–80 to 1998–2000. PloS One 2011;6:e19742. Manrique-Garcia E, Sidorchuk A, Hallqvist J, Moradi T. Socioeconomic position and incidence of acute myocardial infarction: a meta-analysis. J Epidemiol Community Health 2011;65:301–09. Pollitt RA, Rose KM, Kaufman JS. Evaluating the evidence for models of life course socioeconomic factors and cardiovascular outcomes: a systematic review. BMC Public Health 2005;5:7. Leon DA. Trends in European life expectancy: a salutary view. Int J Epidemiol 2011;40:271–77. Kawachi I, Subramanian SV, Kim D. Social Capital and Health. New York, NY: Springer, 2007. Macinko J, Starfield B. The utility of social capital in research on health determinants. Milbank Q 2001;79:387–427. Richardson JG. Handbook of Theory and Research for the Sociology of Education. New York, NY: Greenwood, 1986. Coleman JS. Foundations of Social Theory Cambridge, MA: Harvard University Press, 1990. Putnam RD, Leonardi R, Nanetti R. Making Democracy Work: Civic Tradition in Modern Italy Princeton, NJ: Princeton University Press, 1993. Putnam RD. Bowling Alone: The Collapse and Revival of American Community. New York, NY: Simon and Schuster, 2000. Muntaner C, Lynch J. Social capital, class gender and race conflict, and population health: an essay review of Bowling Alone’s implications for social epidemiology. Int J Epidemiol 2002;31:261–67. Islam MK, Merlo J, Kawachi I, Lindstrom M, Gerdtham UG. Social capital and health: Does egalitarianism matter? A literature review. Int J Equity Health 2006;5:3. Giordano GN, Lindstrom M. The impact of social capital on changes in smoking behaviour: a longitudinal cohort study. Eur J Public Health 2011;21:347–54. Johnson CM, Sharkey JR, Dean WR. Eating behaviors and social capital are associated with fruit and vegetable intake among rural adults. J Hunger Environ Nutr 2010;5:302–15. Kouvonen A, Oksanen T, Vahtera J et al. Work-place social capital and smoking cessation: the Finnish Public Sector Study. Addiction (Abingdon, England) 2008;103:1857–65. Lindstrom M. Social capital, the miniaturization of community and high alcohol consumption: a population-based study. Alcohol Alcohol 2005;40:556–62. Lindstrom M. Social capital, political trust and purchase of illegal liquor: a population-based study in southern Sweden. Health Policy 2008;86:266–75. Lindstrom M. Social capital, political trust and daily smoking and smoking cessation: a population-based study in southern Sweden. Public Health 2009;123:496–501. Weitzman ER, Chen YY. Risk modifying effect of social capital on measures of heavy alcohol consumption, alcohol abuse, harms, and secondhand effects: national survey findings. J Epidemiol Community Health 2005;59:303–09. Ball K, Cleland VJ, Timperio AF, Salmon J, Giles-Corti B, Crawford DA. Love thy neighbour? Associations of social capital and crime with physical activity amongst women. Social Sci Med 2010;71:807–14.
International Journal of Epidemiology, 2014, Vol. 43, No. 6
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
with mortality in New Zealand: a national multilevel cohort study. Int J Epidemiol 2006;35:981–89. Islam MK, Gerdtham UG, Gullberg B, Lindstrom M, Merlo J. Social capital externalities and mortality in Sweden. Econ Hum Biol 2008;6:19–42. van Hooijdonk C, Droomers M, Deerenberg IM, Mackenbach JP, Kunst AE. The diversity in associations between community social capital and health per health outcome, population group and location studied. Int J Epidemiol 2008;37:1384–92. Ali SM, Merlo J, Rosvall M, Lithman T, Lindstrom M. Social capital, the miniaturisation of community, traditionalism and first time acute myocardial infarction: a prospective cohort study in southern Sweden. Soc Sci Med 2006;63:2204–17. Hyyppa MT, Maki J, Impivaara O, Aromaa A. Individual-level measures of social capital as predictors of all-cause and cardiovascular mortality: a population-based prospective study of men and women in Finland. Eur J Epidemiol 2007;22:589–97. Aida J, Kondo K, Hirai H et al. Assessing the association between all-cause mortality and multiple aspects of individual social capital among the older Japanese. BMC Public Health 2011;11:499. Oksanen T, Kivimaki M, Kawachi I et al. Workplace social capital and all-cause mortality: a prospective cohort study of 28,043 public-sector employees in Finland. Am J Public Health 2011;101:1742–48. Veenstra G, Patterson AC. Capital relations and health: mediating and moderating effects of cultural, economic, and social capitals on mortality in Alameda County, California. Int J health Serv 2012;42:277–91. Wen M, Cagney KA, Christakis NA. Effect of specific aspects of community social environment on the mortality of individuals diagnosed with serious illness. Soc Sci Med 2005;61:1119–34. Sundquist J, Johansson SE, Yang M, Sundquist K. Low linking social capital as a predictor of coronary heart disease in Sweden: a cohort study of 2.8 million people. Soc Sci Med 2006;62:954–63. Scheffler RM, Brown TT, Syme L, Kawachi I, Tolstykh I, Iribarren C. Community-level social capital and recurrence of acute coronary syndrome. Soc Sci Med 2008;66:1603–13. Mohan J, Twigg L, Barnard S, Jones K. Social capital, geography and health: a small-area analysis for England. Soc Sci Med 2005;60:1267–83. Mohan J, Twigg L, Barnard S, Jones K. Social capital, geography and health: a small-area analysis for England. Soc Sci Med 2005;60:1267–83. Wen M, Cagney KA, Christakis NA. Effect of specific aspects of community social environment on the mortality of individuals diagnosed with serious illness. Soc Sci Med 2005;61:1119–34. Ali SM, Merlo J, Rosvall M, Lithman T, Lindstrom M. Social capital, the miniaturisation of community, traditionalism and first time acute myocardial infarction: A prospective cohort study in southern Sweden. Soc Sci Med 2006;63:2204–17. Blakely T, Atkinson J, Ivory V, Collings S, Wilton J, HowdenChapman P. No association of neighbourhood volunteerism with mortality in New Zealand: A national multilevel cohort study. Int J Epidemiol 2006;35:981–8. Sundquist J, Johansson SE, Yang M, Sundquist K. Low linking social capital as a predictor of coronary heart disease in
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
34. Broyles ST, Mowen AJ, Theall KP, Gustat J, Rung AL. Integrating social capital into a park-use and active-living framework. Am J Prev Med 2011;40:522–29. 35. Lindstrom M. Social capital, desire to increase physical activity and leisure-time physical activity: a population-based study. Public Health 2011;125:442–47. 36. Lindstrom M, Moghaddassi M, Merlo J. Social capital and leisure time physical activity: a population based multilevel analysis in Malmo, Sweden. J Epidemiol Community Health 2003;57:23–28. 37. Lynch J, Smith GD, Hillemeier M, Shaw M, Raghunathan T, Kaplan G. Income inequality, the psychosocial environment, and health: comparisons of wealthy nations. Lancet 2001;358:194–200. 38. Kawachi I, Kennedy BP, Glass R. Social capital and self-rated health: a contextual analysis. Am J PublicHhealth 1999;89:1187–93. 39. Poortinga W. Social capital: an individual or collective resource for health? Soc Sci Med 2006;62:292–302. 40. Ahern MM, Hendryx MS. Social capital and risk for chronic illnesses. Chronic Illn 2005;1:183–90. 41. Kirkbride JB, Boydell J, Ploubidis GB et al. Testing the association between the incidence of schizophrenia and social capital in an urban area. Psychol Med 2008;38:1083–94. 42. Kirkbride JB, Morgan C, Fearon P, Dazzan P, Murray RM, Jones PB. Neighbourhood-level effects on psychoses: re-examining the role of context. Psychol Med 2007;37:1413–25. 43. Siahpush M, Singh GK. Social integration and mortality in Australia. Aust N Z J Public Health 1999;23:571–77. 44. Lochner KA, Kawachi I, Brennan RT, Buka SL. Social capital and neighborhood mortality rates in Chicago. Soc Sci Med 2003;56:1797–805. 45. Babb P. Measurement of Social Capital in the UK. London: Office for National Statistics, 2005. 46. Harper R. Measurement of Social Capital in the United Kingdom. London: Office for National Statistics, 2002. 47. Ziersch AM. Health implications of access to social capital: findings from an Australian study. Soc Sci Med 2005;61:2119–31. 48. Hunter BD, Neiger B, West J. The importance of addressing social determinants of health at the local level: the case for social capital. Health Soc Care Community 2011;19:522–30. 49. Putnam RD, Leonardi R, Nanetti RY. Making Democracy Work: Civic TraditionsiIn Modern Italy. Princeton, NJ: Princeton University Press, 1993. 50. Torche F, Valenzuela E. Trust and reciprocity: A theoretical distinction of the sources of social capital. Eur J Soc Theory 2011;14:181–98. 51. McMillan DW, Chavis DM. Sense of community: A definition and theory. J Community Psychol 1986;14:6–23. 52. Sterne JA, Sutton AJ, Ioannidis JP et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ (Clinical research ed.) 2011;343:d4002. 53. Peters JL, Sutton AJ, Jones DR, Abrams KR, Rushton L. Contour-enhanced meta-analysis funnel plots help distinguish publication bias from other causes of asymmetry. J Clin Epidemiol 2008;61:991–96. 54. Blakely T, Atkinson J, Ivory V, Collings S, Wilton J, HowdenChapman P. No association of neighbourhood volunteerism
1919
1920
71.
72.
73.
74.
75.
77.
78.
79.
80.
81. Islam MK, Merlo J, Kawachi I, Lindstrom M, Gerdtham UG. Social capital and health: does egalitarianism matter? A literature review. Int J Equity Health 2006;5:3. 82. Policy Research Initiative. Social Capital as a Public Policy Tool: Project Report. Ottawa, ON: Government of Canada, 2005. 83. Grootaert C, van Bastelaer T. Understanding and Measuring Social Capital: A Synthesis of Findings and Recommendations from the Social Capital Initiative. Washington, DC: Social Capital Initiative, 2001. 84. Lynch J, Due P, Muntaner C, Davey Smith G. Social capital—is it a good investment strategy for public health? J Epidemiol Community Health 2000;54:404–08. 85. Pearce N, Davey Smith G. Is social capital the key to inequalities in health? Am J Public Health 2003;93:122–29. 86. Kawachi I, Kim D, Coutts A, Subramanian SV. Commentary: Reconciling the three accounts of social capital. Int J Epidemiol 2004;33:682–90; discussion 700–04. 87. Harpham T, Grant E, Thomas E. Measuring social capital within health surveys: key issues. Health Policy Plan 2002;17:106–11. 88. Department of Health. Healthy Lives, Healthy People: Our Strategy for Public Health in England. London: DH, 2010. 89. Department of Health. Practical Approaches to Improving the Lives of Disabled and Older People Through Building Stronger Communities. London: Department of Health, 2010. 90. Peters PA, Oliver LN, Carriere GM. Geozones: an area-based method for analysis of health outcomes. Health Rep 2012;23:55–64. 91. Kawachi I, Berkman LF. Social cohesion, social capital, and health. In: Berkman LF, Kawachi I (eds). Social Epidemiology. New York, NY: Oxford University Press,2000. 92. Vanderweele TJ, Vansteelandt S, Robins JM. Effect decomposition in the presence of an exposure-induced mediator-outcome confounder. Epidemiology 2014;25:300–06. 93. Kawachi I, Subramanian SV, Kim D. Social Capital and Health. New York, NY: Springer, 2010.
Downloaded from http://ije.oxfordjournals.org/ at University of Sussex on February 2, 2015
76.
Sweden: A cohort study of 2.8 million people. Soc Sci Med 2006;62:954–63. Hyyppa MT, Maki J, Impivaara O, Aromaa A. Individual-level measures of social capital as predictors of all-cause and cardiovascular mortality: A population-based prospective study of men and women in Finland. Eur J Epidemiol 2007;22:589–97. Chaix B, Lindstrom M, Rosvall M, Merlo J. Neighbourhood social interactions and risk of acute myocardial infarction. J Epidemiol Community Health 2008;62:62–68. Islam MK, Gerdtham UG, Gullberg B, Lindstrom M, Merlo J. Social capital externalities and mortality in Sweden. Econ Hum Biol 2008;6:19–42. Scheffler RM, Brown TT, Syme L, Kawachi I, Tolstykh I, Iribarren C. Community-level social capital and recurrence of acute coronary syndrome. Soc Sci Med 2008;66:1603–13. Van Hooijdonk C, Droomers M, Deerenberg IM, Mackenbach JP, Kunst AE. The diversity in associations between community social capital and health per health outcome, population group and location studied. Int J Epidemiol 2008;37:1384–92. Aida J, Kondo K, Hirai H, et al. Assessing the association between all-cause mortality and multiple aspects of individual social capital among the older Japanese. BMC Public Health 2011;11:499. Clark CJ, Guo H, Lunos S et al. Neighborhood cohesion is associated with reduced risk of stroke mortality. Stroke 2011;42:1212–17. Oksanen T, Kivimaki M, Kawachi I et al. Workplace social capital and all-cause mortality: a prospective cohort study of 28,043 public-sector employees in Finland. Am J Public Health 2011;101:1742–48. Veenstra G, Patterson AC. Capital relations and health: Mediating and moderating effects of cultural, economic, and social capitals on mortality in Alameda County, California. Int J Health Serv 2012;42:277–91. Murayama H, Fujiwara Y, Kawachi I. Social capital and health: a review of prospective multilevel studies. J Epidemiol 2012;22:179–87.
International Journal of Epidemiology, 2014, Vol. 43, No. 6