IJE Advance Access originally published online on September 9, 2008
International Journal of Epidemiology 2008 37(6):1384-1392; doi:10.1093/ije/dyn181
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The diversity in associations between community social capital and health per health outcome, population group and location studied
1 Centre for Prevention and Health Services Research, National Institute for Public Health and the Environment, Bilthoven, The Netherlands.
2 Department of Public Health, Erasmus Medical Centre, Rotterdam, The Netherlands.
3 Statistics Netherlands, Voorburg, The Netherlands.
* Corresponding author. Department of Public Health, Erasmus Medical Centre, Rotterdam, PO Box 2040, 3000 CA Rotterdam, The Netherlands. E-mail: c.vanhooijdonk.1{at}erasmusmc.nl
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
Background Literature on the effect of community social capital on health is inconsistent and could be related to differences in social capital measures, health outcomes, population groups and locations studied. Therefore this study examines the diversity in associations between community social capital and health by investigating different diseases, populations groups and locations.
Methods Mortality records and individual data on sex, age, marital status, ethnic origin and place of residence were available for 6 years (1995–2000). Neighbourhood data, i.e. community social capital, socio-economic level and urbanicity, were linked through postcode information. Community social capital was indicated by measures of community interaction, belongingness, satisfaction and involvement. Variations in all-cause and cause-specific mortality across low and high social capital neighbourhoods were estimated through Poisson regression. In addition, analyses were stratified according to population group and to urbanization level.
Results In the total population, community social capital was not related to all-cause mortality (RR = 1.00; CI: 0.99–1.01). However, residents of high social capital neighbourhoods had lower mortality risks for cancer [especially lung cancer (RR = 0.92; CI: 0.89–0.96)] and for suicide (RR = 0.90; CI: 0.83–0.98). Slightly lower mortality risks were also found for men (RR = 0.98; CI: 0.97–0.99), married individuals (RR = 0.96; CI: 0.94–0.97) and for residents living in socially strong neighbourhoods located in large cities (RR = 0.95; CI: 0.91–0.99).
Conclusions The association between community social capital and health differs per health outcome, study population and location studied. This underlines the need to take such diversity into account when aiming to conceptualize the relation between community social capital and health.
Keywords Community social capital, mortality, population group, urbanization, neighbourhood, The Netherlands
Accepted 7 August 2008
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
A positive association has been reported between community social capital and health,1–11 whereas others have found no such association,1,6,7,12–17 This contradiction might be related to the different definitions used to embody the concept of social capital.18–20 The definition most often used in public health is that of Putnam, in which social capital is defined as those features of social structures (such as levels of interpersonal trust and norms of reciprocity and mutual aid), which act as resources for individuals and facilitate collective action.21 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.19
Another explanation for the contrasting results regarding social capital might be related to differences in health outcome, study population or the setting of the performed studies. First, living in a strong social community might affect only certain health outcomes or diseases.19 One of the pathways between community social capital and health specifies health-related behaviour as a mediator. Since unhealthy behaviour is related to certain diseases, one would expect stronger associations with these diseases. Second, living in a high social capital neighbourhood may be more beneficial for certain groups of people.22,23 Differences in the amount of time spent in the neighbourhood, or in the appreciation of living in a strong social community, might elicit a series of health effects.24,25 Finally, the location of the study might also affect the association between community social capital and health.19 For example, higher levels of social capital are observed in rural than in urban areas.26,27 Since the average level of social capital is higher in rural areas, the relation between social capital and health might be less strong. Whereas the concept of social capital and its pathways to health are strongly debated, less attention is paid to the possible diversity in associations between community social capital and health.
Therefore, the present study focuses on the diversity in associations between community social capital and health with regard to different diseases, population groups and locations. First, we determined the relationship between community social capital and all-cause and cause-specific mortality using (nation-wide) data on individuals and data on socio-economic level, urbanicity and community social capital of neighbourhoods. To embody the concept of community social capital, several dimensions were combined e.g. measures of community interaction, belongingness, satisfaction and involvement. Then we examined the association between community social capital and mortality among different population groups and at different levels of urbanization.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
Data
Data were available at the individual level (including information on demographics and mortality) and on the neighbourhood level (including information on community social capital, socio-economic level and level of urbanization). Individual and neighbourhood data were linked by postcode information.
Individual data
Mortality records and demographic data for the years 1995 through 2000 were provided by Statistics Netherlands (CBS) and linked by personal identification number. During this 6 year period, persons could enter (through birth or immigration) or leave the study (by death or emigration). All persons who died during the study period were registered, irrespective of whether the death occurred in The Netherlands or abroad.
The primary cause of death was classified according to the ICD-9 (1995) and ICD-10 (1996–2000). The following causes of death were distinguished: (i) all-cause mortality, (ii) cancer mortality, in particular lung cancer and breast cancer, (iii) cardiovascular disease mortality, in particular ischaemic and cerebrovascular diseases and (iv) external causes, in particular suicide and road traffic accidents.
Demographic information on sex, age, marital status (never married, married, divorced and widowed) and ethnic origin (Dutch, Turkish, Moroccan, Antillean, Surinamese and other) of each individual was available. Five year age categories were used, except for the age groups 0- to 1-year old and 95 years and older. The standard definition of CBS was used to define ethnic origin,28 i.e. a person was considered to be of non-Western origin if at least one parent or the person in question was born in a non-Western country or continent, i.e. Turkey, Africa, Latin America or Asia. In families from mixed origin the country of birth of the mother prevailed.
Neighbourhood data
Neighbourhoods were chosen as the geographical unit of analyses because neighbourhoods are small and the boundaries are based on topography or socio-economic similarities of residents. In 1995, The Netherlands consisted of 10 381 neighbourhoods with on average 1486 residents and an area surface of 3.4 km2. All neighbourhood data were provided by CBS.
Community social capital data originated from the Housing Demand Survey (WBO) of 1998. This national survey was carried out by The Netherlands Ministry of Housing, Spatial Planning and Environment in a two-step procedure. First, one or more municipalities were selected out of each COROP area (a large geographical division in The Netherlands). Then, persons aged 18 years and older living in the selected municipalities were drawn at random. The total response rate across all municipalities was 51%. The data were collected by means of telephone interviews, face-to-face interviews or internet questionnaires. Information was available for 117 569 individuals, i.e. 0.76% of the Dutch population. We identified 13 items, which represented dimensions that could be linked to the definition of social capital as stated by Putnam21 (Table 1). These items represented themes such as community interaction, belongingness, satisfaction and involvement. The 13 items were used to create a single component to indicate community social capital through an unrotated principal component analysis with correlation matrix. This single component had an eigenvalue of 6.95 and explained 54% of the variation between the items. All 13 items were moderate to highly correlated with the calculated component (Table 1). In addition, a reliability test was performed to examine whether the 13 items reflected one single dimension, i.e. community social capital. A Cronbach's alpha of 0.60 was found, indicating an acceptable reliability of the calculated social capital measure. Subsequently, neighbourhood social capital was calculated as the mean score of all interviewed neighbourhood residents. The neighbourhood social capital score had a mean of 0 and a standard deviation of 0.43.
|
Socio-economic status (SES) is an important confounder in the relationship between community social capital and mortality; therefore, appropriate adjustments should be made during analyses. In The Netherlands, there is no national registry for the individual SES. Therefore, the neighbourhood socio-economic level, which can be used as a proxy for individual SES,29 was used. Neighbourhood socio-economic level was indicated by the percentage of neighbourhood inhabitants with a low income; this is an income below the 40% level of the national income distribution (<(
12 025). More detailed measures on neighbourhood income were available, but the percentage of low-income inhabitants was selected because this explained most of the geographical variation in mortality across neighbourhoods in The Netherlands (14%). We selected only one variable to indicate the socio-economic level of the neighbourhood, rather than any combination of the six available income variables, since such a combination did not substantially improve the explanation of the geographical variation in health outcomes (20%). No information was available on the mean educational or occupational level of the area residents. It has been suggested that education or occupation may be less appropriate measures than income to indicate area socio-economic level.30–32 The neighbourhoods were divided into three levels of urbanization. The highest level included neighbourhoods located in the four largest cities in The Netherlands (Amsterdam, Rotterdam, The Hague and Utrecht); this category included 148 neighbourhoods with 1 235 960 residents. Neighbourhoods included in the medium category were located in one of the smaller cities (including Almelo, Arnhem, Breda, Deventer, Eindhoven, Enschede, Groningen, Helmond, Hengelo, ‘s-Hertogenbosch, Leeuwarden, Maastricht, Nijmegen, Tilburg and Zwolle); this level included 421 neighbourhoods with 1 359 500 residents. The remaining neighbourhoods were categorized in the lowest urbanization level and included 2938 neighbourhoods with 8 442 180 residents.
Information on neighbourhood socio-economic level and level of urbanization for the year 1995 was used because mortality records were available from this year onwards.
Neighbourhoods with less than five observations for the 1998 WBO survey (6863 neighbourhoods) were excluded from analyses because social capital scores based on such a small sample are less reliable. An additional 11 neighbourhoods were excluded because of missing data on neighbourhood socio-economic level. Analyses were performed on the remaining 3507 neighbourhoods (33.78% of all neighbourhoods in The Netherlands) with 11 037 640 residents (71.56% of all Dutch citizens) of which 91 656 were interviewed (77.96% of all interviewed individuals in the WBO survey). Mostly rural neighbourhoods with few residents were excluded from our analyses. Rural areas generally have a lower mortality rate33–37 and a higher level of social capital,26,27 thus excluding these areas might have led to underestimation of the association between community social capital and mortality. The 3507 neighbourhoods were ranked according to their community social capital score and divided into five categories with 20% of the study population in each category.
Analyses
All-cause mortality
Poisson regression models were used to calculate relative all-cause mortality risks for each social capital category, with the lowest social capital category as reference. These analyses were adjusted for population composition by including the logarithm of the expected number of deaths in a neighbourhood as offset variable. The logarithm of the expected number of deaths is the usual offset, because the outcome, i.e. the number of observed deaths (dependent variable), is logged in a Poisson model. The expected number of deaths in a neighbourhood was the sum of the expected number of deaths of each of the 336 population groups based on sex, age, marital status and ethnic origin (Dutch, Turkish, Moroccan, Antillean, Surinamese and other) in one neighbourhood. The expected number of deaths per population group was calculated by multiplying the mortality risk of each population group, based on Poisson regression of the total Dutch population, by the number of people in each population group. Additionally, the regression models to estimate relative mortality risks per social capital category were adjusted for neighbourhood socio-economic and urbanization level.
Cause-specific mortality, population group and urbanization level
Similar to all-cause mortality, Poisson regression analyses were performed to estimate the association between community social capital and specific causes of death and community social capital and all-cause mortality among the population groups and the urbanization levels. The regression analyses were stratified according to the population groups (male or female; non-married or married; Western or non-Western ethnic origin) and urbanization levels (high, medium and low).
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
Neighbourhoods with a high level of community social capital had a higher percentage of residents who were married, aged 45 years and older and of Western origin (Table 2), and fewer residents with a low income and of lower age. There were also fewer deaths in this type of neighbourhood.
|
After adjusting for population composition, lower mortality risks were found in high social capital neighbourhoods (Table 3). When socio-economic level and urbanicity of the neighbourhood were taken into account, mortality risks in neighbourhoods in the highest social capital category were similar to those found in the lowest social capital category. After adjustment for population composition, neighbourhood socio-economic level and urbanicity, the slightly lower mortality risks remained for the intermediate social capital category.
|
Although no association was found between community social capital and all-cause mortality, social capital was associated with both lower and higher mortality risks for specific causes of death (Table 4). The lowest mortality risks for cancer and lung cancer were observed among residents living in neighbourhoods with the highest social capital levels. A pattern similar to that for all-cause mortality was found for death due to external causes and suicide. Residents living in strong social neighbourhoods had a higher risk for mortality from cerebrovascular heart disease. No associations were found between community social capital and mortality caused by breast cancer, cardiovascular diseases, ischaemic heart disease and road traffic accidents.
|
Table 5 presents data on all-cause mortality risks per social capital category for different population groups and urbanization levels. Slightly lower mortality risks were found across the social capital categories for residents who are male, married and of Western origin; only in the highest social capital category, residents of Western origin showed no increased risk. Somewhat higher mortality risks were observed in the highest social capital neighbourhoods (but not in the three intermediate categories) for women, non-married residents and residents of non-Western origin.
|
In urban areas, residents living in socially strong neighbourhoods had a lower mortality risk than those in socially weak neighbourhoods. The opposite was found for residents of neighbourhoods in the medium urbanization level, with slightly elevated mortality risks among residents of high social capital neighbourhoods. In the lowest urbanization level, no higher or lower mortality risks were found for residents living in strong or weak social neighbourhoods.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
Key findings
This study found a great diversity in associations between social capital and health when different diseases, population groups and locations were taken into consideration. However, most of the observed associations between community social capital and mortality were weak, with relative risks ranging from 0.92 to 1.09 for the different causes of death, from 0.96 to 1.05 for the different demographic groups and from 0.95 to 1.04 for the different urbanization levels. No association was found between community social capital and all-cause mortality when taking the total Dutch population into consideration. However, living in a strong social neighbourhood was related to a lower mortality risk for cancer (especially lung cancer), and for external causes (especially suicide). Men, married individuals and residents of urban areas living in a socially strong neighbourhood had a slightly lower mortality risk than those living in a socially weak neighbourhood.
Evaluation of data and methods
Some methodological limitations of the present study need to be addressed.
First, the clustering of individuals within neighbourhoods and the spatial clustering of neighbourhoods was disregarded. Residents living in the same neighbourhood tend to be more comparable in terms of demographic and socio-economic characteristics as opposed to residents of other neighbourhoods. More appropriate statistical methods (such as multilevel analysis) would take this clustering of individuals into account, leading to better estimates of contextual influences.38,39 In addition, geographically adjoining neighbourhoods are comparable in terms of contextual variables, which might also influence associations found in the present study.40–42 Unfortunately, we were unable to adjust for clustering between individuals and neighbourhoods because the data set was too large to perform such analyses within the CBS infrastructure. Indirect standardized mortality rates were used to account for clustering between individuals within neighbourhoods. The spatial correlation between adjoining neighbourhoods in community social capital seemed minimal when neighbourhoods were mapped across The Netherlands (map not shown). However, disregarding the clustering of individuals within neighbourhoods and the spatial clustering between neighbourhoods could have led to overestimation of the significance levels of the observed associations; however, we believe that such an overestimation will be minimal.
Second, the concept of social capital remains debatable and various definitions are used to embody this concept.18–20 We used the definition of Putnam21 and created an index containing several social capital items; however, some aspects of community social capital were not included because data were not available. Nevertheless, we believe that including more items would not have substantially changed the geographical variation in community social capital.
Third, community social capital was based on the perceptions of a sample of neighbourhood residents; this sample was sometimes as small as five persons (neighbourhoods with fewer respondents were excluded from our analysis). This minimum number might be too low and too unreliable to indicate neighbourhood social capital. Therefore, we evaluated to what extent our results would differ if a higher cut-off, limit, i.e. 20 respondents per neighbourhood, had been used to include the neighbourhoods. Had we used the later cut-off point, the final study sample would have included 1002 neighbourhoods with 4 481 960 residents. Most of the studied associations did not reach statistical significance, but did give an indication that supported our main findings. For example, slightly lower all-cause mortality risks (RR = 0.98; CI: 0.96–0.99) and lower risks for ischaemic heart disease mortality (RR = 0.95; CI: 0.92–0.99) were found among residents living in the highest social capital category. Comparable associations were observed for the different population groups, except that women, non-married residents and residents of non-Western origin showed no elevated mortality risks in the highest social capital category. No associations were found between community social capital and mortality in any of the three urbanization levels.
Finally, the variation in our community social capital measure was small (Table 1) and might be responsible for the weak association found between community social capital and mortality. Other studies reported a greater variation in their social capital measure1,5,6,16 but comparison is difficult because different constructs are used to embody community social capital.18–20
Comparison to previous studies
In the present study, no association was found between community social capital and all-cause mortality. Although some earlier studies also reported no association,12,14,17 others found a lower mortality risk in strong social communities.3,7 Although the association between community social capital and mortality proved to be weak, it might be stronger for less extreme health outcomes, such as self-perceived health. Most studies,2,4–6,8,10 but not all,6,15,16 found that living in a strong social community was related to better self-perceived health.
Very few studies have examined the relation between community social capital and health for certain population groups. One study found that living in religiously affiliated communities in Israel decreased mortality risks and that these risks were lower for men than for women.3 A study examining the influence of community social capital on survival after hospitalization among the elderly found no evidence for any interactions between social capital and sex or between social capital and ethnicity.11 A study performed in the Helsinki metropolitan area found that community social capital was related to lower mortality rates among the younger age groups but not among the elderly.7
In the present study, the association between community social capital and mortality varied slightly depending on the urbanization level. To our knowledge, only one other study stratified their study sample into various urbanization levels but found no association between community social capital and self-perceived health in rural or in urban areas.26
Explanations for the results
No association was found between community social capital and all-cause mortality for the total Dutch population. An explanation for this might be that, in The Netherlands, community social capital is a less important determinant of health compared with the social environments created in other places (e.g. work, school or sports club). Compared with contacts with friends or relatives, contacts with neighbours are often seen as ‘weaker’ social ties25 and have less impact on health than the ‘stronger’ social ties.43
Living in a high social capital neighbourhood was most clearly related to a lower mortality risk among men and married persons. Different levels of appreciation of living in a strong social community9,26 might explain the observed differences in associations between the married and non-married groups. The higher risk for mortality in strong social neighbourhoods for women and for residents of non-Western origin might be related to processes of social pressure and social control, which might prohibit these groups from making the optimal choice for their own well-being.13
Ethnic heterogeneity within the neighbourhood might also influence the relation between social capital and mortality. Studies on racial residential segregation in the USA reported worse health outcomes among residents of racially segregated areas.44–46 However, compared with the USA, ethnic heterogeneity within Dutch neighbourhoods is less dramatic with only 15% of the neighbourhoods containing >20% migrants. In additional analyses, in which neighbourhoods were stratified according to the percentage of migrants, no substantial differences emerged for the relation between social capital and mortality between neighbourhoods with a low or high percentage of migrants.
Lung cancer mortality risk was decreased among residents living in high social capital neighbourhoods. Informal social control and social pressure from neighbours might prevent smoking and thereby reduce lung cancer rates. In additional analyses, the role of smoking was assessed by examining smoking percentages across the social capital categories, using data from the WBO survey. From the highest to the lowest social capital category the percentages of smokers were 29, 30, 31, 33 and 38% (P-value for trend <0.001), which supported the role of smoking.
Finally, selective migration processes might contribute to the observed associations between community social capital and mortality as observed in the different population groups and locations. Healthier individuals are more likely to move towards the ‘more attractive’ areas such as high social capital neighbourhoods while unhealthier individuals stay behind in the ‘less attractive’ areas such as low social capital neighbourhoods. In The Netherlands only weak associations were found between self-perceived health and migration to deprived or less deprived areas.47 Selective migration processes are also related to personal characteristics such as age and marital status48–50 and the resulting health differences across neighbourhoods might therefore be larger for certain population groups. However, the effect of selective migration processes on the relationship between community social capital and health has not yet been fully investigated.
|
Conclusions |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
This study found that the association between community social capital and health differs according to the cause of death, population group and location studied; however, most of the associations found were weak. The range of health effects found in previous research on social capital might be explained, in part, by the differences in health outcomes, populations or locations studied. Comparisons of different health outcomes, population groups and locations might help to further elucidate the relation between community social capital and health.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
The authors thank Statistics Netherlands (CBS) for the use of their data and Vivian Bos for preparation of the study data.
Conflict of interest: None declared.
KEY MESSAGES
|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
1 Araya R, Dunstan F, Playle R, Thomas H, Palmer S, Lewis G. Perceptions of social capital and the built environment and mental health. Soc Sci Med (2006) 62:3072–83.[CrossRef][Web of Science][Medline]
2 Sundquist K, Yang M. Linking social capital and self-rated health: a multilevel analysis of 11 175 men and women in Sweden. Health Place (2007) 13:324–34.[CrossRef][Web of Science][Medline]
3 Jaffe DH, Eisenbach Z, Neumark YD, Manor O. Does living in a religiously affiliated neighborhood lower mortality? Ann Epidemiol (2005) 15:804–10.[CrossRef][Web of Science][Medline]
4 Kawachi I, Kennedy BP, Glass R. Social capital and self-rated health: a contextual analysis. Am J Public Health (1999) 89:1187–93.
5 Kim D, Kawachi I. A multilevel analysis of key forms of community- and individual-level social capital as predictors of self-rated health in the United States. J Urban Health (2006) 83:813–26.[CrossRef][Web of Science][Medline]
6 Kim D, Subramanian SV, Kawachi I. Bonding versus bridging social capital and their associations with self-rated health: a multilevel analysis of 40 US communities. J Epidemiol Community Health (2006) 60:116–22.
7 Martikainen P, Kauppinen TM, Valkonen T. Effects of the characteristics of neighbourhoods and the characteristics of people on cause specific mortality: a register based follow-up study of 252 000 men. J Epidemiol Community Health (2003) 57:210–17.
8 Poortinga W. Social capital: an individual or collective resource for health? Soc Sci Med (2006) 62:292–302.[CrossRef][Web of Science][Medline]
9 Skrabski A, Kopp M, Kawachi I. Social capital in a changing society: cross-sectional associations with middle aged female and male mortality rates. J Epidemiol Community Health (2003) 57:114–19.
10 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.[CrossRef][Web of Science][Medline]
11 Wen M, Christakis NA. Neighborhood effects on posthospitalization mortality: a population-based cohort study of the elderly in Chicago. Health Serv Res (2005) 40:1108–27.[CrossRef][Web of Science][Medline]
12 Turrell G, Kavanagh A, Subramanian SV. Area variation in mortality in Tasmania (Australia): the contributions of socioeconomic disadvantage, social capital and geographic remoteness. Health Place (2006) 12:291–305.[CrossRef][Web of Science][Medline]
13 Mohan J, Barnard S, Jones K, Twigg L. Social Capital, Place and Health: Creating, Validating and Applying Small-Area Indicators in the Modeling of Health Outcomes. (2004) Wetherby, Yorkshire: National Institute for Health and Clinical Excellence.
14 Blakely T, Atkinson J, Ivory V, Collings S, Wilton J, Howden-Chapman P. No association of neighbourhood volunteerism with mortality in New Zealand: a national multilevel cohort study. Int J Epidemiol (2006) 35:981–89.
15 Carpiano RM. Neighborhood social capital and adult health: An empirical test of a Bourdieu-based model. Health Place (2007) 13:639–55.[CrossRef][Web of Science][Medline]
16 Kavanagh AM, Turrell G, Subramanian SV. Does area-based social capital matter for the health of Australians? A multilevel analysis of self-rated health in Tasmania. Int J Epidemiol (2006) 35:607–13.
17 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.[CrossRef][Web of Science][Medline]
18 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.[CrossRef][Medline]
19 Kawachi I, Berkman LF. Social cohesion, social capital and health. In: Social Epidemiology.—Berkman LF, Kawachi I, eds. (2000) New York: Oxford University Press. 174–90.
20 Macinko J, Starfield B. The utility of social capital in research on health determinants. Milbank Q (2001) 79:387–427.[CrossRef][Web of Science][Medline]
21 Putnam RD. Making Democracy Work. (1993) New Jersey: Princeton University Press.
22 Macintyre S, Ellaway A. Neighbourhoods and health: an overview. In: Neighbourhoods and Health.—Kawachi I, Berkman LF, eds. (2003) New York: Oxford University Press.
23 Robert S. Socio-economic position and health: the independent contribution of community socio-economic context. Annu Rev Sociol (1999) 25:489–516.[CrossRef][Web of Science]
24 Kawachi I, Berkman LF. Social ties and mental health. J Urban Health (2001) 78:458–67.[Web of Science][Medline]
25 Wellman B, Wortley S. Different strokes from different folks; community ties and social support. Am J Sociol (1990) 96:558–88.[CrossRef]
26 Greiner KA, Li C, Kawachi I, Hunt DC, Ahluwalia JS. The relationships of social participation and community ratings to health and health behaviors in areas with high and low population density. Soc Sci Med (2004) 59:2303–12.[CrossRef][Web of Science][Medline]
27 Onyx J, Bullen P. Measuring social capital in five communities. J Appl Behav Sci (2000) 36:23–42.
28 Keij-Deerenberg IM. Numbers of foreigners according to several definitions (Aantallen allochtonen volgens verschillende definities). Maandstatistiek Bevolking (2000) 5:14–17.
29 Bos V, Kunst AE, Mackenbach JP. Examining socio-economic mortality differences based on information at the small geographical scale (Nationale gegevens over social-economische sterfteverschillen op basis van informatie over kleine geografische eenheden). (2000) Rotterdam: Department of Public Health, Erasmus University, Rotterdam. 37–50.
30 Duncan GJ, Daly MC, McDonough P, Williams DR. Optimal indicators of socioeconomic status for health research. Am J Public Health (2002) 92:1151–57.
31 Geyer S, Peter R. Income, occupational position, qualification and health inequalities—competing risks? (Comparing indicators of social status). J Epidemiol Community Health (2000) 54:299–305.
32 Lantz PM, House JS, Lepkowski JM, Williams DR, Mero RP, Chen J. Socioeconomic factors, health behaviors, and mortality: results from a nationally representative prospective study of US adults. JAMA (1998) 279:1703–708.
33 Barnett E, Strogatz D, Armstrong D, Wing S. Urbanisation and coronary heart disease mortality among African Americans in the US South. J Epidemiol Community Health (1996) 50:252–57.
34 Haynes R. Inequalities in health and health service use: evidence from the General Household Survey. Soc Sci Med (1991) 33:361–68.[CrossRef][Web of Science][Medline]
35 Haynes R, Gale S. Mortality, long-term illness and deprivation in rural and metropolitan wards of England and Wales. Health Place (1999) 5:301–12.[CrossRef][Web of Science][Medline]
36 House JS, Lepkowski JM, Williams DR, Mero RP, Lantz PM, Robert SA, Chen J. Excess mortality among urban residents: how much, for whom, and why? Am J Public Health (2000) 90:1898–904.
37 Kindig DA, Seplaki CL, Libby DL. Death rate variation in US subpopulations. Bull World Health Organ (2002) 80:9–15.[Web of Science][Medline]
38 Duncan C, Jones K, Moon G. Context, composition and heterogeneity: using multilevel models in health research. Soc Sci Med (1998) 46:97–117.[CrossRef][Web of Science][Medline]
39 Macintyre S, Ellaway A. Ecological approaches: rediscovering the role of the physical and social environment. In: Social Epidemiology.—Berkman LF, Kawachi I, eds. (2000) New York: Oxford University Press. 332–48.
40 Lorant V, Thomas I, Deliege D, Tonglet R. Deprivation and mortality: the implications of spatial autocorrelation for health resources allocation. Soc Sci Med (2001) 53:1711–19.[CrossRef][Web of Science][Medline]
41 Richardson S. Statistical methods for geographical correlation studies. In:. In: Geographical and Environmental Epidemiology: Methods for Small-Area Studies.—Elliott P, Cuzick J, English D, Stern R, eds. (2000) New York: Oxford University Press. 181–204.
42 Sridharan S, Tunstall H, Lawder R, Mitchell R. An exploratory spatial data analysis approach to understanding the relationship between deprivation and mortality in Scotland. Soc Sci Med (2007) 65:1942–52.[CrossRef][Web of Science][Medline]
43 Cattell V. Poor people, poor places and poor health: the mediating role of social networks and social capital. Soc Sci Med (2001) 52:1501–16.[CrossRef][Web of Science][Medline]
44 Acevedo-Garcia D, Lochner KA, Osypuk TL, Subramanian SV. Future directions in residential segregation and health research: a multilevel approach. Am J Public Health (2003) 93:215–21.
45 Inagami S, Borrell LN, Wong MD, Fang J, Shapiro MF, Asch SM. Residential segregation and latino, black and white mortality in New York city. J Urban Health (2006) 83:406–20.[CrossRef][Web of Science][Medline]
46 Subramanian SV, Acevedo-Garcia D, Osypuk TL. Racial residential segregation and geographic heterogeneity in black/white disparity in poor self-rated health in the US: a multilevel statistical analysis. Soc Sci Med (2005) 60:1667–79.[CrossRef][Web of Science][Medline]
47 van Lenthe FJ, Martikainen P, Mackenbach JP. Neighbourhood inequalities in health and health-related behaviour: results of selective migration? Health Place (2007) 13:123–37.[CrossRef][Web of Science][Medline]
48 Bentham G. Migration and morbidity: implications for geographical studies of disease. Soc Sci Med (1988) 26:49–54.[CrossRef][Web of Science][Medline]
49 Boyle P, Norman P, Rees P. Does migration exaggerate the relationship between deprivation and limiting long-term illness? A Scottish analysis. Soc Sci Med (2002) 55:21–31.[CrossRef][Web of Science][Medline]
50 Norman P, Boyle P, Rees P. Selective migration, health and deprivation: a longitudinal analysis. Soc Sci Med (2005) 60:2755–71.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. J Kunitz Commentary: Comments on the diversity in associations between community social capital and health Int. J. Epidemiol., December 1, 2008; 37(6): 1393 - 1394. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||