- Split View
-
Views
-
Cite
Cite
Tony Blakely, June Atkinson, Vivienne Ivory, Sunny Collings, Jenny Wilton, Philippa Howden-Chapman, No association of neighbourhood volunteerism with mortality in New Zealand: a national multilevel cohort study, International Journal of Epidemiology, Volume 35, Issue 4, August 2006, Pages 981–989, https://doi.org/10.1093/ije/dyl088
- Share Icon Share
Abstract
Background The association of social capital with health and mortality is contentious, and empirical findings are inconsistent. This study tests the association of neighbourhood-level volunteerism with mortality.
Methods Cohort study of 1996 New Zealand census respondents aged 25–74 years (4.75 million person years) using multilevel Poisson regression analyses. Neighbourhood (average population 2034) measures included indices of social capital (volunteering activities for all census respondents) and deprivation.
Results Adjusting for just age and marital status, the mortality rate ratios for people living in the quintile of neighbourhoods with the lowest compared with highest volunteerism were 1.16 (95% confidence interval 1.08–1.24) and 1.09 (1.01–1.18), for males and females, respectively. Adjusting for potential individual-level and neighbourhood-level socioeconomic confounders reduced the rate ratios to 0.94 (0.88–1.01) and 0.92 (0.85–1.01), respectively. There was no significant association with any cause of death, including suicide [rate ratios 0.89 (0.64–1.22) and 0.57 (0.31–1.05), respectively]. Restricting the analyses to only those census respondents living at their census night address for five or more years, and therefore ‘exposed’ to that level of volunteerism for a longer period, did not substantially alter findings.
Conclusions This study, one of the largest multilevel studies yet, found no statistically significant independent association of a structural measure of neighbourhood social capital with mortality—including suicide. Assuming social features of neighbourhoods are important determinants of health, future research should examine other features (e.g. social fragmentation) and other outcomes (e.g. behaviour).
Health varies between neighbourhoods, and characteristics of neighbourhoods (over and above characteristics of people) probably influence health.1 Macintyre and others have long argued that ‘unless we try to explore more systematically the ways in which different types of area differ, we are left without any suggestions for social or public health policies that might improve the health of those in the worst areas, other than those relating to individual improvements in lifestyle’ [p. 219 in Ref. (2)]. Three generally agreed domains to explore neighbourhood determinants include physical features (e.g. air pollution), availability of services (e.g. health care facilities), and social features (e.g. social capital).
Social capital is one potential aspect of the social features of neighbourhoods that may affect health. Its possible role has been robustly debated.3–5 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).6 Social capital has been associated with mortality in ecological studies.7–10 However, ecological studies are prone to error due to omission of individual-level confounders and other biases. Multilevel studies that control simultaneously for personal and neighbourhood-level characteristics are a more reliable study design,11,12 but they are more challenging: a large number of both individuals and neighbourhoods are needed. Multilevel studies of social capital and mortality thus far have used different methodologies and have had mixed findings.13–17 For example, no association was found between a range of social capital measures (political participation, trust in public and private institutions, social trust, neighbourhood integration, and isolation) and mortality in Tasmania, Australia, using routine mortality data to ‘create’ a multilevel study with age and sex at the individual-level.16 On the other hand, a weak association was found in Helsinki of a rather basic measure of social cohesion (standardized score of proportion of men living with a partner, voter turnout, and residential mobility) with all-cause, injury, and alcohol-related causes of death among 25–64 year olds.15 Subsequent work for all of Finland also found an ∼10–20% elevated suicide rate for each quartile decrease in voter turnout for areas with mean size of ∼60 000, after adjusting for a range of individual-level and area-level covariates.17
There are alternative theoretical perspectives on what social capital is and how to measure it. One perspective divides social capital into ‘structural’ and ‘cognitive’ aspects.18 The former represents the ‘store’ of social capital that may be built up in communities and can be measured directly (e.g. number of community organizations) or by behaviours and activities (e.g. participation, volunteering). The latter ‘cognitive’ aspect represents peoples' beliefs and assumptions, such as trust and attachment. More recent conceptual work on social capital has proposed a division of the construct into bonding (trusting and co-operative relations between members of a network who see themselves as being similar), bridging (between people who know that they are not alike in some socio-demographic sense), and linking (between people who are interacting across explicit, formal, or institutional power or authority gradients) forms of social capital.3 Specific measurement of these domains, however, is undeveloped.5 One recent study in Tasmania, Australia, found no evidence of an association of linking social capital (using a composite measure of trust in public institutions) with self-rated health, and some evidence of bonding social capital (using a composite measure of social trust in friends and relatives) being associated with poorer self-rated health.19 Finally, the level at which social capital is conceptualized to exist and operate varies. Some earlier theoretical perspectives emphasized social capital, or social cohesion more generally, being a property of the collective, with any putative association with health being over and above that due to, say, the individual-level manifestations of social networks and social support.6,20 Alternatively, the individual-level and ecological-levels may both be incorporated. For example, a reconsideration of Bourdieu might highlight the importance of relationships in defining social space (and, by extension, perhaps geographical space)21 or the accrual of resources to individuals as a result of their membership of social networks.22 That is, the demonstrated importance (in observational studies at least) of social networks and support to health23 are brought back into the social capital framework. Another recent study demonstrates that while an ecological measure of social trust is associated with individual self-rated health, it is largely mediated through individual perceptions of trust.24
In this study, we test the association of one measure of social capital, volunteerism measured at the neighbourhood-level, with mortality in New Zealand. A priori, an association with suicide seemed more likely than for other causes of death, because of the possibly shorter latency between exposure to a social context and suicide than for other causes of death, and sociological arguments dating back to Durkheim. Given the wide range of theoretical and empirical approaches to social capital, it is important at the outset to recognize what our conceptualization of social capital is, and what it is not. Using the ‘cognitive-structural’ distinction above, volunteerism is a structural measure of social capital; it is an indirect measure of community organizations occupying the space between personal networks and state or national organizations. Using the linking–bonding–bridging dimension, it arguably aligns more with linking social capital, although it also touches on bonding and bridging, i.e. we do not have good specificity on this dimension. Also, our volunteerism measure is an ecological measure; while it would be technically feasible to assess the role individual's volunteering behaviour has as an explanatory variable between the neighbourhood-level measure of volunteerism, it was beyond the scope of this study. Likewise, examination of other possible measures of neighbourhood social capital in New Zealand (e.g. political participation, trust) were beyond the scope and practicality of this study.
While cognizant of our above operationalization of social capital as volunteerism, several factors make this multilevel study internationally significant. First, analyses are conducted on linked census-mortality data for the entire adult population, making this one of the largest multilevel studies, internationally, of social capital and mortality. Second, neighbourhoods are identified with an average population size of ∼2000, compared with the larger aggregations often used in other studies. The actual neighbourhood boundaries, ‘area units’, ‘generally coincide with communities of interest or parts thereof’ and in towns and cities ‘generally coincide with suburbs or parts thereof’ (www.stats.govt.nz). Third, covariate data are available at the individual-level (e.g. ethnicity, income, and education) and the neighbourhood-level (socioeconomic deprivation), allowing thorough multilevel analyses to adjust for cross-level confounding. Fourth, neighbourhood volunteerism was measured using census questions on volunteering of all New Zealanders, not just a survey sample, allowing the quantification of a precise and stable neighbourhood-level measure.
Methods
Neighbourhood measure of social capital
The 1996 census included six questions on unpaid voluntary activities outside the respondent's home in the 4 weeks up to census night:
looking after a child
household work, cooking, repairs, gardening, or looking after a person who is aged, ill or has a disability
unpaid training, coaching, teaching, giving advice or counselling, helping at school, etc.
attending committee meeting, or organization, administration, policy work, etc., unpaid, for group, church, or marae (Maori tribal meeting place)
unpaid fund-raising work, selling, etc., for group, church, or marae
other unpaid work for a group, church, or marae.
We used the 1683 census area units (average population 2034; constructed by Statistics New Zealand to fit the natural boundaries of communities of interest) as our neighbourhood units. This level of aggregation was purposely selected as it approximates the suburb level at which much voluntary effort is focused; the smaller ‘meshblock’ level aggregation (∼100 people) is too small, capturing just streets or parts thereof; the larger ‘urban area’ or ‘territorial authority’ levels of aggregation captured whole cities (or large parts thereof) and in the latter case also mixed urban and rural zones. For each neighbourhood we calculated the proportion of census respondents aged 15 years or older responding positively to each of the above questions, and a summary proportion of people responding positively to any one of the six questions. Based on iterative principal components analyses, and a theoretical position that questions 1 and 2 above were not as relevant as the remaining four questions to our conceptualization of ‘structural’ or ‘linking’ social capital, we constructed an index using proportions for the last four questions above (loading on first component of 0.77, 0.92, 0.82, and 0.77, respectively) and the summary proportion variable (loading of 0.93; Eigen value 3.57). We then grouped the neighbourhoods into quintiles of social capital.
Linked census-mortality data
The linkage of census and mortality data in New Zealand has been described in detail elsewhere.25 A total of 79% of eligible mortality records (age 25-77 years at death, aged up to 74 years on census night, New Zealand residents) for the 1996–99 period were linked back to a 1996 census record, with at least 98% of these links estimated to be correct.26 The proportion of mortality records linked to a census record varied by sex, age, ethnicity, and neighbourhood deprivation. To adjust for any resultant linkage bias, we calculated inverse probability weights for use in all analyses in this paper. For example, if 20 out of 30 Māori male decedents were aged 45–64 years residing in moderately deprived small areas of New Zealand were linked to a census record, each of the 20 linked records received a weight of 1.5 (= 30/20). Similar weights were calculated and applied to numerous strata and have been shown elsewhere to be reliable.27
Covariates
Sex, age, marital status, and ethnicity (Maori, Pacific, and non-Maori non-Pacific) were included in all analyses as likely confounders. A range of potentially confounding individual-level socioeconomic variables were also available from the census data (Table 1). The household income variable was calculated by aggregating personal incomes in each household, then equivalizing (to adjust for economies of scale) for the number of adults and children. We used a logarithmic transformation of income as we have previously found it to be linearly associated with mortality risk (once negative and very low incomes are recoded to $1000).28 At the small-area-level, we measured deprivation using the NZDep96 index.29 This index was calculated from census data on socioeconomic characteristics (e.g. car access, tenure, and benefit receipt) at aggregations of ∼100 people and assigned to individual census respondents.
Person-years | ||||||||||||
Males | Females | Deaths (weighted) | ||||||||||
Level | n | % | n | % | Males | Females | ||||||
Age | ||||||||||||
25–44 years | 1 176 423 | 51.0 | 1 274 004 | 52.2 | 1815 | 981 | ||||||
45–64 years | 835 959 | 36.2 | 843 759 | 34.5 | 5448 | 3597 | ||||||
65–77 years | 294 378 | 12.8 | 325 185 | 13.3 | 9183 | 5820 | ||||||
Ethnicity | ||||||||||||
Mäori | 221 616 | 9.6 | 244 464 | 10.0 | 2058 | 1440 | ||||||
Pacific people | 65 007 | 2.8 | 71 523 | 2.9 | 465 | 297 | ||||||
Non-Maori Non-Pacific | 2 020 137 | 87.6 | 2 126 958 | 87.1 | 13 917 | 8661 | ||||||
Income | ||||||||||||
Low income | 399 195 | 17.3 | 567 861 | 23.2 | 5757 | 4281 | ||||||
Medium-low income | 382 044 | 16.6 | 418 425 | 17.1 | 3990 | 2400 | ||||||
Medium income | 366 129 | 15.9 | 372 168 | 15.2 | 2355 | 1338 | ||||||
Medium-high income | 433 794 | 18.8 | 417 309 | 17.1 | 2058 | 1209 | ||||||
High income | 725 598 | 31.5 | 667 188 | 27.3 | 2283 | 1170 | ||||||
Education | ||||||||||||
No qualifications | 723 045 | 31.3 | 817 371 | 33.5 | 8049 | 5799 | ||||||
School qualifications | 572 181 | 24.8 | 782 355 | 32.0 | 3285 | 2550 | ||||||
Post-school qualifications | 1 011 534 | 43.9 | 843 222 | 34.5 | 5109 | 2046 | ||||||
Smoking | ||||||||||||
Smoker | 567 879 | 24.6 | 556 278 | 22.8 | 4758 | 2634 | ||||||
Ex-smoker | 647 913 | 28.1 | 539 211 | 22.1 | 6930 | 3054 | ||||||
Never smoke | 1 090 971 | 47.3 | 1 347 462 | 55.2 | 4755 | 4710 | ||||||
Car access | ||||||||||||
Nil cars | 117 747 | 5.1 | 174 498 | 7.1 | 1854 | 1827 | ||||||
1 car | 832 503 | 36.1 | 966 060 | 39.5 | 8046 | 5229 | ||||||
2 or more cars | 1 356 510 | 58.8 | 1 302 390 | 53.3 | 6543 | 3339 | ||||||
Labour force | ||||||||||||
Employed | 1 768 239 | 76.7 | 1 467 225 | 60.1 | 5367 | 2214 | ||||||
Unemployed | 91 590 | 4.0 | 87 795 | 3.6 | 450 | 147 | ||||||
Non-labour | 446 931 | 19.4 | 887 931 | 36.3 | 10 626 | 8037 | ||||||
Marital status | ||||||||||||
Married | 1 755 207 | 76.1 | 1 767 555 | 72.4 | 11 934 | 6177 | ||||||
Not married | 551 556 | 23.9 | 675 393 | 27.6 | 4509 | 4221 | ||||||
Rurality | ||||||||||||
Urban | 1 743 456 | 75.6 | 1 884 618 | 77.1 | 12 234 | 7950 | ||||||
Minor urban | 194 487 | 8.4 | 209 343 | 8.6 | 2022 | 1287 | ||||||
Rural and other | 368 820 | 16.0 | 348 984 | 14.3 | 2187 | 1155 | ||||||
Neighbourhood deprivation | ||||||||||||
Least deprived | 476 568 | 20.7 | 495 729 | 20.3 | 2286 | 1428 | ||||||
Quintile 2 | 497 466 | 21.6 | 520 383 | 21.3 | 2961 | 1809 | ||||||
Quintile 3 | 496 983 | 21.5 | 524 919 | 21.5 | 3438 | 2235 | ||||||
Quintile 4 | 465 084 | 20.2 | 500 073 | 20.5 | 3876 | 2481 | ||||||
Most deprived | 370 659 | 16.1 | 401 847 | 16.4 | 3876 | 2442 | ||||||
Neighourhood volunteerism | ||||||||||||
Lowest | 426 627 | 18.5 | 455 028 | 18.6 | 3018 | 1971 | ||||||
Medium-low | 468 840 | 20.3 | 507 024 | 20.8 | 3321 | 2142 | ||||||
Medium | 464 601 | 20.1 | 500 841 | 20.5 | 3438 | 2142 | ||||||
Medium-high | 464 898 | 20.2 | 498 069 | 20.4 | 3540 | 2262 | ||||||
Highest | 481 797 | 20.9 | 481 992 | 19.7 | 3123 | 1878 |
Person-years | ||||||||||||
Males | Females | Deaths (weighted) | ||||||||||
Level | n | % | n | % | Males | Females | ||||||
Age | ||||||||||||
25–44 years | 1 176 423 | 51.0 | 1 274 004 | 52.2 | 1815 | 981 | ||||||
45–64 years | 835 959 | 36.2 | 843 759 | 34.5 | 5448 | 3597 | ||||||
65–77 years | 294 378 | 12.8 | 325 185 | 13.3 | 9183 | 5820 | ||||||
Ethnicity | ||||||||||||
Mäori | 221 616 | 9.6 | 244 464 | 10.0 | 2058 | 1440 | ||||||
Pacific people | 65 007 | 2.8 | 71 523 | 2.9 | 465 | 297 | ||||||
Non-Maori Non-Pacific | 2 020 137 | 87.6 | 2 126 958 | 87.1 | 13 917 | 8661 | ||||||
Income | ||||||||||||
Low income | 399 195 | 17.3 | 567 861 | 23.2 | 5757 | 4281 | ||||||
Medium-low income | 382 044 | 16.6 | 418 425 | 17.1 | 3990 | 2400 | ||||||
Medium income | 366 129 | 15.9 | 372 168 | 15.2 | 2355 | 1338 | ||||||
Medium-high income | 433 794 | 18.8 | 417 309 | 17.1 | 2058 | 1209 | ||||||
High income | 725 598 | 31.5 | 667 188 | 27.3 | 2283 | 1170 | ||||||
Education | ||||||||||||
No qualifications | 723 045 | 31.3 | 817 371 | 33.5 | 8049 | 5799 | ||||||
School qualifications | 572 181 | 24.8 | 782 355 | 32.0 | 3285 | 2550 | ||||||
Post-school qualifications | 1 011 534 | 43.9 | 843 222 | 34.5 | 5109 | 2046 | ||||||
Smoking | ||||||||||||
Smoker | 567 879 | 24.6 | 556 278 | 22.8 | 4758 | 2634 | ||||||
Ex-smoker | 647 913 | 28.1 | 539 211 | 22.1 | 6930 | 3054 | ||||||
Never smoke | 1 090 971 | 47.3 | 1 347 462 | 55.2 | 4755 | 4710 | ||||||
Car access | ||||||||||||
Nil cars | 117 747 | 5.1 | 174 498 | 7.1 | 1854 | 1827 | ||||||
1 car | 832 503 | 36.1 | 966 060 | 39.5 | 8046 | 5229 | ||||||
2 or more cars | 1 356 510 | 58.8 | 1 302 390 | 53.3 | 6543 | 3339 | ||||||
Labour force | ||||||||||||
Employed | 1 768 239 | 76.7 | 1 467 225 | 60.1 | 5367 | 2214 | ||||||
Unemployed | 91 590 | 4.0 | 87 795 | 3.6 | 450 | 147 | ||||||
Non-labour | 446 931 | 19.4 | 887 931 | 36.3 | 10 626 | 8037 | ||||||
Marital status | ||||||||||||
Married | 1 755 207 | 76.1 | 1 767 555 | 72.4 | 11 934 | 6177 | ||||||
Not married | 551 556 | 23.9 | 675 393 | 27.6 | 4509 | 4221 | ||||||
Rurality | ||||||||||||
Urban | 1 743 456 | 75.6 | 1 884 618 | 77.1 | 12 234 | 7950 | ||||||
Minor urban | 194 487 | 8.4 | 209 343 | 8.6 | 2022 | 1287 | ||||||
Rural and other | 368 820 | 16.0 | 348 984 | 14.3 | 2187 | 1155 | ||||||
Neighbourhood deprivation | ||||||||||||
Least deprived | 476 568 | 20.7 | 495 729 | 20.3 | 2286 | 1428 | ||||||
Quintile 2 | 497 466 | 21.6 | 520 383 | 21.3 | 2961 | 1809 | ||||||
Quintile 3 | 496 983 | 21.5 | 524 919 | 21.5 | 3438 | 2235 | ||||||
Quintile 4 | 465 084 | 20.2 | 500 073 | 20.5 | 3876 | 2481 | ||||||
Most deprived | 370 659 | 16.1 | 401 847 | 16.4 | 3876 | 2442 | ||||||
Neighourhood volunteerism | ||||||||||||
Lowest | 426 627 | 18.5 | 455 028 | 18.6 | 3018 | 1971 | ||||||
Medium-low | 468 840 | 20.3 | 507 024 | 20.8 | 3321 | 2142 | ||||||
Medium | 464 601 | 20.1 | 500 841 | 20.5 | 3438 | 2142 | ||||||
Medium-high | 464 898 | 20.2 | 498 069 | 20.4 | 3540 | 2262 | ||||||
Highest | 481 797 | 20.9 | 481 992 | 19.7 | 3123 | 1878 |
Person-years | ||||||||||||
Males | Females | Deaths (weighted) | ||||||||||
Level | n | % | n | % | Males | Females | ||||||
Age | ||||||||||||
25–44 years | 1 176 423 | 51.0 | 1 274 004 | 52.2 | 1815 | 981 | ||||||
45–64 years | 835 959 | 36.2 | 843 759 | 34.5 | 5448 | 3597 | ||||||
65–77 years | 294 378 | 12.8 | 325 185 | 13.3 | 9183 | 5820 | ||||||
Ethnicity | ||||||||||||
Mäori | 221 616 | 9.6 | 244 464 | 10.0 | 2058 | 1440 | ||||||
Pacific people | 65 007 | 2.8 | 71 523 | 2.9 | 465 | 297 | ||||||
Non-Maori Non-Pacific | 2 020 137 | 87.6 | 2 126 958 | 87.1 | 13 917 | 8661 | ||||||
Income | ||||||||||||
Low income | 399 195 | 17.3 | 567 861 | 23.2 | 5757 | 4281 | ||||||
Medium-low income | 382 044 | 16.6 | 418 425 | 17.1 | 3990 | 2400 | ||||||
Medium income | 366 129 | 15.9 | 372 168 | 15.2 | 2355 | 1338 | ||||||
Medium-high income | 433 794 | 18.8 | 417 309 | 17.1 | 2058 | 1209 | ||||||
High income | 725 598 | 31.5 | 667 188 | 27.3 | 2283 | 1170 | ||||||
Education | ||||||||||||
No qualifications | 723 045 | 31.3 | 817 371 | 33.5 | 8049 | 5799 | ||||||
School qualifications | 572 181 | 24.8 | 782 355 | 32.0 | 3285 | 2550 | ||||||
Post-school qualifications | 1 011 534 | 43.9 | 843 222 | 34.5 | 5109 | 2046 | ||||||
Smoking | ||||||||||||
Smoker | 567 879 | 24.6 | 556 278 | 22.8 | 4758 | 2634 | ||||||
Ex-smoker | 647 913 | 28.1 | 539 211 | 22.1 | 6930 | 3054 | ||||||
Never smoke | 1 090 971 | 47.3 | 1 347 462 | 55.2 | 4755 | 4710 | ||||||
Car access | ||||||||||||
Nil cars | 117 747 | 5.1 | 174 498 | 7.1 | 1854 | 1827 | ||||||
1 car | 832 503 | 36.1 | 966 060 | 39.5 | 8046 | 5229 | ||||||
2 or more cars | 1 356 510 | 58.8 | 1 302 390 | 53.3 | 6543 | 3339 | ||||||
Labour force | ||||||||||||
Employed | 1 768 239 | 76.7 | 1 467 225 | 60.1 | 5367 | 2214 | ||||||
Unemployed | 91 590 | 4.0 | 87 795 | 3.6 | 450 | 147 | ||||||
Non-labour | 446 931 | 19.4 | 887 931 | 36.3 | 10 626 | 8037 | ||||||
Marital status | ||||||||||||
Married | 1 755 207 | 76.1 | 1 767 555 | 72.4 | 11 934 | 6177 | ||||||
Not married | 551 556 | 23.9 | 675 393 | 27.6 | 4509 | 4221 | ||||||
Rurality | ||||||||||||
Urban | 1 743 456 | 75.6 | 1 884 618 | 77.1 | 12 234 | 7950 | ||||||
Minor urban | 194 487 | 8.4 | 209 343 | 8.6 | 2022 | 1287 | ||||||
Rural and other | 368 820 | 16.0 | 348 984 | 14.3 | 2187 | 1155 | ||||||
Neighbourhood deprivation | ||||||||||||
Least deprived | 476 568 | 20.7 | 495 729 | 20.3 | 2286 | 1428 | ||||||
Quintile 2 | 497 466 | 21.6 | 520 383 | 21.3 | 2961 | 1809 | ||||||
Quintile 3 | 496 983 | 21.5 | 524 919 | 21.5 | 3438 | 2235 | ||||||
Quintile 4 | 465 084 | 20.2 | 500 073 | 20.5 | 3876 | 2481 | ||||||
Most deprived | 370 659 | 16.1 | 401 847 | 16.4 | 3876 | 2442 | ||||||
Neighourhood volunteerism | ||||||||||||
Lowest | 426 627 | 18.5 | 455 028 | 18.6 | 3018 | 1971 | ||||||
Medium-low | 468 840 | 20.3 | 507 024 | 20.8 | 3321 | 2142 | ||||||
Medium | 464 601 | 20.1 | 500 841 | 20.5 | 3438 | 2142 | ||||||
Medium-high | 464 898 | 20.2 | 498 069 | 20.4 | 3540 | 2262 | ||||||
Highest | 481 797 | 20.9 | 481 992 | 19.7 | 3123 | 1878 |
Person-years | ||||||||||||
Males | Females | Deaths (weighted) | ||||||||||
Level | n | % | n | % | Males | Females | ||||||
Age | ||||||||||||
25–44 years | 1 176 423 | 51.0 | 1 274 004 | 52.2 | 1815 | 981 | ||||||
45–64 years | 835 959 | 36.2 | 843 759 | 34.5 | 5448 | 3597 | ||||||
65–77 years | 294 378 | 12.8 | 325 185 | 13.3 | 9183 | 5820 | ||||||
Ethnicity | ||||||||||||
Mäori | 221 616 | 9.6 | 244 464 | 10.0 | 2058 | 1440 | ||||||
Pacific people | 65 007 | 2.8 | 71 523 | 2.9 | 465 | 297 | ||||||
Non-Maori Non-Pacific | 2 020 137 | 87.6 | 2 126 958 | 87.1 | 13 917 | 8661 | ||||||
Income | ||||||||||||
Low income | 399 195 | 17.3 | 567 861 | 23.2 | 5757 | 4281 | ||||||
Medium-low income | 382 044 | 16.6 | 418 425 | 17.1 | 3990 | 2400 | ||||||
Medium income | 366 129 | 15.9 | 372 168 | 15.2 | 2355 | 1338 | ||||||
Medium-high income | 433 794 | 18.8 | 417 309 | 17.1 | 2058 | 1209 | ||||||
High income | 725 598 | 31.5 | 667 188 | 27.3 | 2283 | 1170 | ||||||
Education | ||||||||||||
No qualifications | 723 045 | 31.3 | 817 371 | 33.5 | 8049 | 5799 | ||||||
School qualifications | 572 181 | 24.8 | 782 355 | 32.0 | 3285 | 2550 | ||||||
Post-school qualifications | 1 011 534 | 43.9 | 843 222 | 34.5 | 5109 | 2046 | ||||||
Smoking | ||||||||||||
Smoker | 567 879 | 24.6 | 556 278 | 22.8 | 4758 | 2634 | ||||||
Ex-smoker | 647 913 | 28.1 | 539 211 | 22.1 | 6930 | 3054 | ||||||
Never smoke | 1 090 971 | 47.3 | 1 347 462 | 55.2 | 4755 | 4710 | ||||||
Car access | ||||||||||||
Nil cars | 117 747 | 5.1 | 174 498 | 7.1 | 1854 | 1827 | ||||||
1 car | 832 503 | 36.1 | 966 060 | 39.5 | 8046 | 5229 | ||||||
2 or more cars | 1 356 510 | 58.8 | 1 302 390 | 53.3 | 6543 | 3339 | ||||||
Labour force | ||||||||||||
Employed | 1 768 239 | 76.7 | 1 467 225 | 60.1 | 5367 | 2214 | ||||||
Unemployed | 91 590 | 4.0 | 87 795 | 3.6 | 450 | 147 | ||||||
Non-labour | 446 931 | 19.4 | 887 931 | 36.3 | 10 626 | 8037 | ||||||
Marital status | ||||||||||||
Married | 1 755 207 | 76.1 | 1 767 555 | 72.4 | 11 934 | 6177 | ||||||
Not married | 551 556 | 23.9 | 675 393 | 27.6 | 4509 | 4221 | ||||||
Rurality | ||||||||||||
Urban | 1 743 456 | 75.6 | 1 884 618 | 77.1 | 12 234 | 7950 | ||||||
Minor urban | 194 487 | 8.4 | 209 343 | 8.6 | 2022 | 1287 | ||||||
Rural and other | 368 820 | 16.0 | 348 984 | 14.3 | 2187 | 1155 | ||||||
Neighbourhood deprivation | ||||||||||||
Least deprived | 476 568 | 20.7 | 495 729 | 20.3 | 2286 | 1428 | ||||||
Quintile 2 | 497 466 | 21.6 | 520 383 | 21.3 | 2961 | 1809 | ||||||
Quintile 3 | 496 983 | 21.5 | 524 919 | 21.5 | 3438 | 2235 | ||||||
Quintile 4 | 465 084 | 20.2 | 500 073 | 20.5 | 3876 | 2481 | ||||||
Most deprived | 370 659 | 16.1 | 401 847 | 16.4 | 3876 | 2442 | ||||||
Neighourhood volunteerism | ||||||||||||
Lowest | 426 627 | 18.5 | 455 028 | 18.6 | 3018 | 1971 | ||||||
Medium-low | 468 840 | 20.3 | 507 024 | 20.8 | 3321 | 2142 | ||||||
Medium | 464 601 | 20.1 | 500 841 | 20.5 | 3438 | 2142 | ||||||
Medium-high | 464 898 | 20.2 | 498 069 | 20.4 | 3540 | 2262 | ||||||
Highest | 481 797 | 20.9 | 481 992 | 19.7 | 3123 | 1878 |
Mortality outcome
Mortality was treated as all-causes, cardiovascular disease (CVD), cancer (malignant), unintentional injury, and suicide.
Analyses
All analyses were conducted with the Glimmix macro in SAS (version 8.2), using Poisson regression. In the first instance we specified a random error term at the neighbourhood-level at which social capital was measured. However, when the average number of outcome events (i.e. deaths) becomes less than about five in each level-2 unit (i.e. neighbourhood), the multilevel regression output may become unreliable.30 Given the large number of neighbourhoods in our analyses, and the large amount of person-time, our models were probably robust to some violation of this general rule for all causes of death combined—but probably not for cause-specific mortality. For example, only 22.5 and 7.7% of neighbourhoods had at least one male or female suicide death, respectively. Moreover, analyses with a neighbourhood-level random error for injury and suicide deaths demonstrated an implausibly large variance at the neighbourhood-level (results not shown). Consequently, we only report cause-specific mortality analyses with the random term at a regional-level (n = 73 Territorial Authorities); all-cause analyses are presented with the random error term at both neighbourhood-level and regional-level. It is important to note, however, that the social capital variable was always measured at the neighbourhood-level or area unit-level.
Results
A total of 2.31 and 2.44 million person years of follow-up were available for 25–74 year olds during 1996–99 with complete data, with 16 446 and 10 398 male and female deaths (weighted up for linkage bias) occurring during the 3 year follow-up. The distribution of person time and deaths by volunteerism and covariates is shown in Table 1.
In the baseline regression model (model 1) that adjusts only for age, ethnicity, and marital status at the individual-level, and specifies a random effect at the neighbourhood-level, there was a modest association of low volunteerism with higher mortality rates (first column of results in Table 2 for males and Table 3 for females). Among males, the rate ratio comparing the people in the quintile of neighbourhoods with the lowest amounts of volunteerism to the quintile of neighbourhoods with the highest volunteerism was 1.16 [95% confidence interval (95% CI) 1.08–1.24] and among females it was 1.09 (1.01–1.18).
Neighbourhood-level random error | Regional-level random error | |||||||||
Variable | Model 1: Baseline | Model 2: plus individual-level covariates | Model 3: plus neighbourhood-level covariates | Model 1: Baseline | Model 3: plus neighbourhood-level covariates | |||||
Ethnicity | ||||||||||
Mäori | 2.19 (2.09–2.29) | 1.93 (1.84–2.02) | 1.84 (1.76–1.94) | 2.23 (2.13–2.34) | 1.85 (1.76–1.94) | |||||
Pacific people | 1.74 (1.58–1.91) | 1.45 (1.32–1.59) | 1.37 (1.25–1.51) | 1.87 (1.70–2.05) | 1.38 (1.26–1.52) | |||||
Non-Mäori Non-Pacific | 1 | 1 | 1 | 1 | 1 | |||||
Marital status | ||||||||||
Married | 1 | 1 | 1 | 1 | 1 | |||||
Not married | 1.44 (1.39–1.49) | 1.22 (1.18–1.26) | 1.21 (1.17–1.25) | 1.45 (1.41–1.51) | 1.21 (1.16–1.25) | |||||
Logarithm of household income (min $1000) | 0.92 (0.89–0.94) | 0.93 (0.90–0.95) | 0.93 (0.90–0.95) | |||||||
Education | ||||||||||
No qualifications | 1.17 (1.13–1.21) | 1.15 (1.11–1.20) | 1.15 (1.11–1.19) | |||||||
School qualifications | 1.05 (1.01–1.10) | 1.05 (1.01–1.10) | 1.05 (1.01–1.10) | |||||||
Post-school qualifications | 1 | 1 | 1 | |||||||
Car access | ||||||||||
Nil cars | 1.52 (1.44–1.61) | 1.48 (1.39–1.56) | 1.48 (1.40–1.57) | |||||||
1 car | 1.13 (1.09–1.17) | 1.11 (1.07–1.15) | 1.11 (1.07–1.15) | |||||||
2 or more cars | 1 | 1 | 1 | |||||||
Labour force | ||||||||||
Employed | 1 | 1 | 1 | |||||||
Unemployed | 1.39 (1.26–1.52) | 1.37 (1.24–1.50) | 1.37 (1.24–1.51) | |||||||
Non-labour | 1.98 (1.90–2.07) | 1.96 (1.87–2.04) | 1.96 (1.88–2.05) | |||||||
Rurality | ||||||||||
Urban | 1 | 1 | ||||||||
Minor urban | 0.97 (0.91–1.03) | 0.98 (0.92–1.04) | ||||||||
Rural and other | 0.90 (0.85–0.96) | 0.89 (0.84–0.94) | ||||||||
Neighbourhood deprivation | ||||||||||
Least deprived | 1 | 1 | ||||||||
Quintile 2 | 1.11 (1.05–1.19) | 1.13 (1.07–1.19) | ||||||||
Quintile 3 | 1.17 (1.10–1.24) | 1.18 (1.12–1.25) | ||||||||
Quintile 4 | 1.29 (1.21–1.37) | 1.29 (1.22–1.36) | ||||||||
Most deprived | 1.45 (1.36–1.55) | 1.44 (1.36–1.53) | ||||||||
Neighbourhood volunteerism | ||||||||||
Lowest | 1.16 (1.08–1.24) | 1.07 (1.00–1.13) | 0.94 (0.88–1.01) | 1.37 (1.28–1.46) | 0.95 (0.89–1.02) | |||||
Medium-low | 1.13 (1.06–1.21) | 1.06 (1.00–1.13) | 0.98 (0.91–1.04) | 1.27 (1.20–1.34) | 0.98 (0.92–1.04) | |||||
Medium | 1.15 (1.08–1.22) | 1.08 (1.02–1.15) | 1.01 (0.95–1.07) | 1.22 (1.16–1.28) | 1.00 (0.95–1.06) | |||||
Medium-high | 1.12 (1.05–1.19) | 1.07 (1.01–1.13) | 1.01 (0.96–1.08) | 1.16 (1.10–1.22) | 1.01 (0.96–1.07) | |||||
Highest | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (standard error of variance) | ||||||||||
Individual-level | 0.925 (0.0012) | 0.923 (0.0012) | 0.928 (0.0012) | 0.963 (0.0012) | 0.952 (0.0012) | |||||
Neighbourhood-level | 0.065 (0.0060) | 0.042 (0.0049) | 0.030 (0.0044) | |||||||
Regional-level | 0.020 (0.0048) | 0.006 (0.0022) |
Neighbourhood-level random error | Regional-level random error | |||||||||
Variable | Model 1: Baseline | Model 2: plus individual-level covariates | Model 3: plus neighbourhood-level covariates | Model 1: Baseline | Model 3: plus neighbourhood-level covariates | |||||
Ethnicity | ||||||||||
Mäori | 2.19 (2.09–2.29) | 1.93 (1.84–2.02) | 1.84 (1.76–1.94) | 2.23 (2.13–2.34) | 1.85 (1.76–1.94) | |||||
Pacific people | 1.74 (1.58–1.91) | 1.45 (1.32–1.59) | 1.37 (1.25–1.51) | 1.87 (1.70–2.05) | 1.38 (1.26–1.52) | |||||
Non-Mäori Non-Pacific | 1 | 1 | 1 | 1 | 1 | |||||
Marital status | ||||||||||
Married | 1 | 1 | 1 | 1 | 1 | |||||
Not married | 1.44 (1.39–1.49) | 1.22 (1.18–1.26) | 1.21 (1.17–1.25) | 1.45 (1.41–1.51) | 1.21 (1.16–1.25) | |||||
Logarithm of household income (min $1000) | 0.92 (0.89–0.94) | 0.93 (0.90–0.95) | 0.93 (0.90–0.95) | |||||||
Education | ||||||||||
No qualifications | 1.17 (1.13–1.21) | 1.15 (1.11–1.20) | 1.15 (1.11–1.19) | |||||||
School qualifications | 1.05 (1.01–1.10) | 1.05 (1.01–1.10) | 1.05 (1.01–1.10) | |||||||
Post-school qualifications | 1 | 1 | 1 | |||||||
Car access | ||||||||||
Nil cars | 1.52 (1.44–1.61) | 1.48 (1.39–1.56) | 1.48 (1.40–1.57) | |||||||
1 car | 1.13 (1.09–1.17) | 1.11 (1.07–1.15) | 1.11 (1.07–1.15) | |||||||
2 or more cars | 1 | 1 | 1 | |||||||
Labour force | ||||||||||
Employed | 1 | 1 | 1 | |||||||
Unemployed | 1.39 (1.26–1.52) | 1.37 (1.24–1.50) | 1.37 (1.24–1.51) | |||||||
Non-labour | 1.98 (1.90–2.07) | 1.96 (1.87–2.04) | 1.96 (1.88–2.05) | |||||||
Rurality | ||||||||||
Urban | 1 | 1 | ||||||||
Minor urban | 0.97 (0.91–1.03) | 0.98 (0.92–1.04) | ||||||||
Rural and other | 0.90 (0.85–0.96) | 0.89 (0.84–0.94) | ||||||||
Neighbourhood deprivation | ||||||||||
Least deprived | 1 | 1 | ||||||||
Quintile 2 | 1.11 (1.05–1.19) | 1.13 (1.07–1.19) | ||||||||
Quintile 3 | 1.17 (1.10–1.24) | 1.18 (1.12–1.25) | ||||||||
Quintile 4 | 1.29 (1.21–1.37) | 1.29 (1.22–1.36) | ||||||||
Most deprived | 1.45 (1.36–1.55) | 1.44 (1.36–1.53) | ||||||||
Neighbourhood volunteerism | ||||||||||
Lowest | 1.16 (1.08–1.24) | 1.07 (1.00–1.13) | 0.94 (0.88–1.01) | 1.37 (1.28–1.46) | 0.95 (0.89–1.02) | |||||
Medium-low | 1.13 (1.06–1.21) | 1.06 (1.00–1.13) | 0.98 (0.91–1.04) | 1.27 (1.20–1.34) | 0.98 (0.92–1.04) | |||||
Medium | 1.15 (1.08–1.22) | 1.08 (1.02–1.15) | 1.01 (0.95–1.07) | 1.22 (1.16–1.28) | 1.00 (0.95–1.06) | |||||
Medium-high | 1.12 (1.05–1.19) | 1.07 (1.01–1.13) | 1.01 (0.96–1.08) | 1.16 (1.10–1.22) | 1.01 (0.96–1.07) | |||||
Highest | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (standard error of variance) | ||||||||||
Individual-level | 0.925 (0.0012) | 0.923 (0.0012) | 0.928 (0.0012) | 0.963 (0.0012) | 0.952 (0.0012) | |||||
Neighbourhood-level | 0.065 (0.0060) | 0.042 (0.0049) | 0.030 (0.0044) | |||||||
Regional-level | 0.020 (0.0048) | 0.006 (0.0022) |
All models additionally adjust for age in 5 year categories.
Neighbourhood-level random error | Regional-level random error | |||||||||
Variable | Model 1: Baseline | Model 2: plus individual-level covariates | Model 3: plus neighbourhood-level covariates | Model 1: Baseline | Model 3: plus neighbourhood-level covariates | |||||
Ethnicity | ||||||||||
Mäori | 2.19 (2.09–2.29) | 1.93 (1.84–2.02) | 1.84 (1.76–1.94) | 2.23 (2.13–2.34) | 1.85 (1.76–1.94) | |||||
Pacific people | 1.74 (1.58–1.91) | 1.45 (1.32–1.59) | 1.37 (1.25–1.51) | 1.87 (1.70–2.05) | 1.38 (1.26–1.52) | |||||
Non-Mäori Non-Pacific | 1 | 1 | 1 | 1 | 1 | |||||
Marital status | ||||||||||
Married | 1 | 1 | 1 | 1 | 1 | |||||
Not married | 1.44 (1.39–1.49) | 1.22 (1.18–1.26) | 1.21 (1.17–1.25) | 1.45 (1.41–1.51) | 1.21 (1.16–1.25) | |||||
Logarithm of household income (min $1000) | 0.92 (0.89–0.94) | 0.93 (0.90–0.95) | 0.93 (0.90–0.95) | |||||||
Education | ||||||||||
No qualifications | 1.17 (1.13–1.21) | 1.15 (1.11–1.20) | 1.15 (1.11–1.19) | |||||||
School qualifications | 1.05 (1.01–1.10) | 1.05 (1.01–1.10) | 1.05 (1.01–1.10) | |||||||
Post-school qualifications | 1 | 1 | 1 | |||||||
Car access | ||||||||||
Nil cars | 1.52 (1.44–1.61) | 1.48 (1.39–1.56) | 1.48 (1.40–1.57) | |||||||
1 car | 1.13 (1.09–1.17) | 1.11 (1.07–1.15) | 1.11 (1.07–1.15) | |||||||
2 or more cars | 1 | 1 | 1 | |||||||
Labour force | ||||||||||
Employed | 1 | 1 | 1 | |||||||
Unemployed | 1.39 (1.26–1.52) | 1.37 (1.24–1.50) | 1.37 (1.24–1.51) | |||||||
Non-labour | 1.98 (1.90–2.07) | 1.96 (1.87–2.04) | 1.96 (1.88–2.05) | |||||||
Rurality | ||||||||||
Urban | 1 | 1 | ||||||||
Minor urban | 0.97 (0.91–1.03) | 0.98 (0.92–1.04) | ||||||||
Rural and other | 0.90 (0.85–0.96) | 0.89 (0.84–0.94) | ||||||||
Neighbourhood deprivation | ||||||||||
Least deprived | 1 | 1 | ||||||||
Quintile 2 | 1.11 (1.05–1.19) | 1.13 (1.07–1.19) | ||||||||
Quintile 3 | 1.17 (1.10–1.24) | 1.18 (1.12–1.25) | ||||||||
Quintile 4 | 1.29 (1.21–1.37) | 1.29 (1.22–1.36) | ||||||||
Most deprived | 1.45 (1.36–1.55) | 1.44 (1.36–1.53) | ||||||||
Neighbourhood volunteerism | ||||||||||
Lowest | 1.16 (1.08–1.24) | 1.07 (1.00–1.13) | 0.94 (0.88–1.01) | 1.37 (1.28–1.46) | 0.95 (0.89–1.02) | |||||
Medium-low | 1.13 (1.06–1.21) | 1.06 (1.00–1.13) | 0.98 (0.91–1.04) | 1.27 (1.20–1.34) | 0.98 (0.92–1.04) | |||||
Medium | 1.15 (1.08–1.22) | 1.08 (1.02–1.15) | 1.01 (0.95–1.07) | 1.22 (1.16–1.28) | 1.00 (0.95–1.06) | |||||
Medium-high | 1.12 (1.05–1.19) | 1.07 (1.01–1.13) | 1.01 (0.96–1.08) | 1.16 (1.10–1.22) | 1.01 (0.96–1.07) | |||||
Highest | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (standard error of variance) | ||||||||||
Individual-level | 0.925 (0.0012) | 0.923 (0.0012) | 0.928 (0.0012) | 0.963 (0.0012) | 0.952 (0.0012) | |||||
Neighbourhood-level | 0.065 (0.0060) | 0.042 (0.0049) | 0.030 (0.0044) | |||||||
Regional-level | 0.020 (0.0048) | 0.006 (0.0022) |
Neighbourhood-level random error | Regional-level random error | |||||||||
Variable | Model 1: Baseline | Model 2: plus individual-level covariates | Model 3: plus neighbourhood-level covariates | Model 1: Baseline | Model 3: plus neighbourhood-level covariates | |||||
Ethnicity | ||||||||||
Mäori | 2.19 (2.09–2.29) | 1.93 (1.84–2.02) | 1.84 (1.76–1.94) | 2.23 (2.13–2.34) | 1.85 (1.76–1.94) | |||||
Pacific people | 1.74 (1.58–1.91) | 1.45 (1.32–1.59) | 1.37 (1.25–1.51) | 1.87 (1.70–2.05) | 1.38 (1.26–1.52) | |||||
Non-Mäori Non-Pacific | 1 | 1 | 1 | 1 | 1 | |||||
Marital status | ||||||||||
Married | 1 | 1 | 1 | 1 | 1 | |||||
Not married | 1.44 (1.39–1.49) | 1.22 (1.18–1.26) | 1.21 (1.17–1.25) | 1.45 (1.41–1.51) | 1.21 (1.16–1.25) | |||||
Logarithm of household income (min $1000) | 0.92 (0.89–0.94) | 0.93 (0.90–0.95) | 0.93 (0.90–0.95) | |||||||
Education | ||||||||||
No qualifications | 1.17 (1.13–1.21) | 1.15 (1.11–1.20) | 1.15 (1.11–1.19) | |||||||
School qualifications | 1.05 (1.01–1.10) | 1.05 (1.01–1.10) | 1.05 (1.01–1.10) | |||||||
Post-school qualifications | 1 | 1 | 1 | |||||||
Car access | ||||||||||
Nil cars | 1.52 (1.44–1.61) | 1.48 (1.39–1.56) | 1.48 (1.40–1.57) | |||||||
1 car | 1.13 (1.09–1.17) | 1.11 (1.07–1.15) | 1.11 (1.07–1.15) | |||||||
2 or more cars | 1 | 1 | 1 | |||||||
Labour force | ||||||||||
Employed | 1 | 1 | 1 | |||||||
Unemployed | 1.39 (1.26–1.52) | 1.37 (1.24–1.50) | 1.37 (1.24–1.51) | |||||||
Non-labour | 1.98 (1.90–2.07) | 1.96 (1.87–2.04) | 1.96 (1.88–2.05) | |||||||
Rurality | ||||||||||
Urban | 1 | 1 | ||||||||
Minor urban | 0.97 (0.91–1.03) | 0.98 (0.92–1.04) | ||||||||
Rural and other | 0.90 (0.85–0.96) | 0.89 (0.84–0.94) | ||||||||
Neighbourhood deprivation | ||||||||||
Least deprived | 1 | 1 | ||||||||
Quintile 2 | 1.11 (1.05–1.19) | 1.13 (1.07–1.19) | ||||||||
Quintile 3 | 1.17 (1.10–1.24) | 1.18 (1.12–1.25) | ||||||||
Quintile 4 | 1.29 (1.21–1.37) | 1.29 (1.22–1.36) | ||||||||
Most deprived | 1.45 (1.36–1.55) | 1.44 (1.36–1.53) | ||||||||
Neighbourhood volunteerism | ||||||||||
Lowest | 1.16 (1.08–1.24) | 1.07 (1.00–1.13) | 0.94 (0.88–1.01) | 1.37 (1.28–1.46) | 0.95 (0.89–1.02) | |||||
Medium-low | 1.13 (1.06–1.21) | 1.06 (1.00–1.13) | 0.98 (0.91–1.04) | 1.27 (1.20–1.34) | 0.98 (0.92–1.04) | |||||
Medium | 1.15 (1.08–1.22) | 1.08 (1.02–1.15) | 1.01 (0.95–1.07) | 1.22 (1.16–1.28) | 1.00 (0.95–1.06) | |||||
Medium-high | 1.12 (1.05–1.19) | 1.07 (1.01–1.13) | 1.01 (0.96–1.08) | 1.16 (1.10–1.22) | 1.01 (0.96–1.07) | |||||
Highest | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (standard error of variance) | ||||||||||
Individual-level | 0.925 (0.0012) | 0.923 (0.0012) | 0.928 (0.0012) | 0.963 (0.0012) | 0.952 (0.0012) | |||||
Neighbourhood-level | 0.065 (0.0060) | 0.042 (0.0049) | 0.030 (0.0044) | |||||||
Regional-level | 0.020 (0.0048) | 0.006 (0.0022) |
All models additionally adjust for age in 5 year categories.
. | Neighbourhood-level random error . | . | . | Regional-level random error . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Variable . | Model 1: Baseline . | Model 2: plus individual-level covariates . | Model 3: plus neighbourhood-level covariates . | Model 1: Baseline . | Model 3: plus neighbourhood-level covariates . | |||||
Ethnicity | ||||||||||
Mäori | 2.50 (2.36–2.65) | 2.26 (2.13–2.39) | 2.14 (2.01–2.27) | 2.50 (2.36–2.65) | 2.14 (2.02–2.28) | |||||
Pacific people | 1.76 (1.56–1.98) | 1.55 (1.38–1.75) | 1.47 (1.30–1.65) | 1.82 (1.62–2.05) | 1.46 (1.29–1.65) | |||||
Non-Mäori Non-Pacific | 1 | 1 | 1 | 1 | 1 | |||||
Marital status | ||||||||||
Married | 1 | 1 | 1 | 1 | 1 | |||||
Not married | 1.26 (1.21–1.31) | 1.11 (1.06–1.16) | 1.11 (1.06–1.16) | 1.26 (1.21–1.31) | 1.11 (1.06–1.16) | |||||
Logarithm of household income (min $1000) | 0.92 (0.89–0.95) | 0.94 (0.90–0.97) | 0.94 (0.90–0.97) | |||||||
Education | ||||||||||
No qualifications | 1.23 (1.17–1.29) | 1.20 (1.14–1.27) | 1.20 (1.14–1.27) | |||||||
School qualifications | 1.12 (1.06–1.19) | 1.12 (1.05–1.18) | 1.12 (1.05–1.18) | |||||||
Post-school qualifications | 1 | 1 | 1 | |||||||
Car access | ||||||||||
Nil cars | 1.45 (1.36–1.55) | 1.40 (1.31–1.50) | 1.40 (1.30–1.50) | |||||||
1 car | 1.11 (1.05–1.16) | 1.08 (1.03–1.14) | 1.08 (1.03–1.14) | |||||||
2 or more cars | 1 | 1 | 1 | |||||||
Labour force | ||||||||||
Employed | 1 | 1 | 1 | |||||||
Unemployed | 1.09 (0.92–1.28) | 1.08 (0.91–1.27) | 1.08 (0.91–1.28) | |||||||
Non-labour | 1.88 (1.78–2.00) | 1.87 (1.77–1.99) | 1.88 (1.77–1.99) | |||||||
Rurality | ||||||||||
Urban | 1 | 1 | ||||||||
Minor urban | 1.00 (0.93–1.08) | 1.00 (0.93–1.07) | ||||||||
Rural and other | 0.91 (0.84–0.99) | 0.91 (0.84–0.98) | ||||||||
Neighbourhood deprivation | ||||||||||
Least deprived | 1 | 1 | ||||||||
Quintile 2 | 1.07 (0.99–1.16) | 1.09 (1.01–1.17) | ||||||||
Quintile 3 | 1.20 (1.11–1.29) | 1.21 (1.12–1.30) | ||||||||
Quintile 4 | 1.27 (1.18–1.37) | 1.27 (1.18–1.37) | ||||||||
Most deprived | 1.45 (1.34–1.57) | 1.42 (1.31–1.53) | ||||||||
Neighbourhood volunteerism | ||||||||||
Lowest | 1.09 (1.01–1.18) | 1.03 (0.95–1.10) | 0.92 (0.85–1.01) | 1.33 (1.23–1.44) | 0.96 (0.88–1.04) | |||||
Med-low | 1.07 (0.99–1.15) | 1.01 (0.94–1.08) | 0.94 (0.87–1.02) | 1.23 (1.15–1.32) | 0.97 (0.90–1.04) | |||||
Medium | 1.05 (0.97–1.13) | 1.00 (0.93–1.07) | 0.94 (0.87–1.01) | 1.14 (1.06–1.22) | 0.95 (0.89–1.02) | |||||
Med-high | 1.07 (0.99–1.15) | 1.03 (0.96–1.11) | 0.99 (0.92–1.06) | 1.14 (1.07–1.21) | 1.01 (0.94–1.08) | |||||
Highest | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (standard error of variance) | ||||||||||
Individual-level | 0.946 (0.0012) | 0.967 (0.0012) | 0.975 (0.0012) | 0.979 (0.0012) | 0.996 (0.0012) | |||||
Neighbourhood-level | 0.063 (0.0075) | 0.043 (0.0066) | 0.031 (0.0060) | |||||||
Regional-level | 0.027 (0.0070) | 0.007 (0.0029) |
. | Neighbourhood-level random error . | . | . | Regional-level random error . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Variable . | Model 1: Baseline . | Model 2: plus individual-level covariates . | Model 3: plus neighbourhood-level covariates . | Model 1: Baseline . | Model 3: plus neighbourhood-level covariates . | |||||
Ethnicity | ||||||||||
Mäori | 2.50 (2.36–2.65) | 2.26 (2.13–2.39) | 2.14 (2.01–2.27) | 2.50 (2.36–2.65) | 2.14 (2.02–2.28) | |||||
Pacific people | 1.76 (1.56–1.98) | 1.55 (1.38–1.75) | 1.47 (1.30–1.65) | 1.82 (1.62–2.05) | 1.46 (1.29–1.65) | |||||
Non-Mäori Non-Pacific | 1 | 1 | 1 | 1 | 1 | |||||
Marital status | ||||||||||
Married | 1 | 1 | 1 | 1 | 1 | |||||
Not married | 1.26 (1.21–1.31) | 1.11 (1.06–1.16) | 1.11 (1.06–1.16) | 1.26 (1.21–1.31) | 1.11 (1.06–1.16) | |||||
Logarithm of household income (min $1000) | 0.92 (0.89–0.95) | 0.94 (0.90–0.97) | 0.94 (0.90–0.97) | |||||||
Education | ||||||||||
No qualifications | 1.23 (1.17–1.29) | 1.20 (1.14–1.27) | 1.20 (1.14–1.27) | |||||||
School qualifications | 1.12 (1.06–1.19) | 1.12 (1.05–1.18) | 1.12 (1.05–1.18) | |||||||
Post-school qualifications | 1 | 1 | 1 | |||||||
Car access | ||||||||||
Nil cars | 1.45 (1.36–1.55) | 1.40 (1.31–1.50) | 1.40 (1.30–1.50) | |||||||
1 car | 1.11 (1.05–1.16) | 1.08 (1.03–1.14) | 1.08 (1.03–1.14) | |||||||
2 or more cars | 1 | 1 | 1 | |||||||
Labour force | ||||||||||
Employed | 1 | 1 | 1 | |||||||
Unemployed | 1.09 (0.92–1.28) | 1.08 (0.91–1.27) | 1.08 (0.91–1.28) | |||||||
Non-labour | 1.88 (1.78–2.00) | 1.87 (1.77–1.99) | 1.88 (1.77–1.99) | |||||||
Rurality | ||||||||||
Urban | 1 | 1 | ||||||||
Minor urban | 1.00 (0.93–1.08) | 1.00 (0.93–1.07) | ||||||||
Rural and other | 0.91 (0.84–0.99) | 0.91 (0.84–0.98) | ||||||||
Neighbourhood deprivation | ||||||||||
Least deprived | 1 | 1 | ||||||||
Quintile 2 | 1.07 (0.99–1.16) | 1.09 (1.01–1.17) | ||||||||
Quintile 3 | 1.20 (1.11–1.29) | 1.21 (1.12–1.30) | ||||||||
Quintile 4 | 1.27 (1.18–1.37) | 1.27 (1.18–1.37) | ||||||||
Most deprived | 1.45 (1.34–1.57) | 1.42 (1.31–1.53) | ||||||||
Neighbourhood volunteerism | ||||||||||
Lowest | 1.09 (1.01–1.18) | 1.03 (0.95–1.10) | 0.92 (0.85–1.01) | 1.33 (1.23–1.44) | 0.96 (0.88–1.04) | |||||
Med-low | 1.07 (0.99–1.15) | 1.01 (0.94–1.08) | 0.94 (0.87–1.02) | 1.23 (1.15–1.32) | 0.97 (0.90–1.04) | |||||
Medium | 1.05 (0.97–1.13) | 1.00 (0.93–1.07) | 0.94 (0.87–1.01) | 1.14 (1.06–1.22) | 0.95 (0.89–1.02) | |||||
Med-high | 1.07 (0.99–1.15) | 1.03 (0.96–1.11) | 0.99 (0.92–1.06) | 1.14 (1.07–1.21) | 1.01 (0.94–1.08) | |||||
Highest | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (standard error of variance) | ||||||||||
Individual-level | 0.946 (0.0012) | 0.967 (0.0012) | 0.975 (0.0012) | 0.979 (0.0012) | 0.996 (0.0012) | |||||
Neighbourhood-level | 0.063 (0.0075) | 0.043 (0.0066) | 0.031 (0.0060) | |||||||
Regional-level | 0.027 (0.0070) | 0.007 (0.0029) |
All models additionally adjust for age in 5-year categories.
. | Neighbourhood-level random error . | . | . | Regional-level random error . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Variable . | Model 1: Baseline . | Model 2: plus individual-level covariates . | Model 3: plus neighbourhood-level covariates . | Model 1: Baseline . | Model 3: plus neighbourhood-level covariates . | |||||
Ethnicity | ||||||||||
Mäori | 2.50 (2.36–2.65) | 2.26 (2.13–2.39) | 2.14 (2.01–2.27) | 2.50 (2.36–2.65) | 2.14 (2.02–2.28) | |||||
Pacific people | 1.76 (1.56–1.98) | 1.55 (1.38–1.75) | 1.47 (1.30–1.65) | 1.82 (1.62–2.05) | 1.46 (1.29–1.65) | |||||
Non-Mäori Non-Pacific | 1 | 1 | 1 | 1 | 1 | |||||
Marital status | ||||||||||
Married | 1 | 1 | 1 | 1 | 1 | |||||
Not married | 1.26 (1.21–1.31) | 1.11 (1.06–1.16) | 1.11 (1.06–1.16) | 1.26 (1.21–1.31) | 1.11 (1.06–1.16) | |||||
Logarithm of household income (min $1000) | 0.92 (0.89–0.95) | 0.94 (0.90–0.97) | 0.94 (0.90–0.97) | |||||||
Education | ||||||||||
No qualifications | 1.23 (1.17–1.29) | 1.20 (1.14–1.27) | 1.20 (1.14–1.27) | |||||||
School qualifications | 1.12 (1.06–1.19) | 1.12 (1.05–1.18) | 1.12 (1.05–1.18) | |||||||
Post-school qualifications | 1 | 1 | 1 | |||||||
Car access | ||||||||||
Nil cars | 1.45 (1.36–1.55) | 1.40 (1.31–1.50) | 1.40 (1.30–1.50) | |||||||
1 car | 1.11 (1.05–1.16) | 1.08 (1.03–1.14) | 1.08 (1.03–1.14) | |||||||
2 or more cars | 1 | 1 | 1 | |||||||
Labour force | ||||||||||
Employed | 1 | 1 | 1 | |||||||
Unemployed | 1.09 (0.92–1.28) | 1.08 (0.91–1.27) | 1.08 (0.91–1.28) | |||||||
Non-labour | 1.88 (1.78–2.00) | 1.87 (1.77–1.99) | 1.88 (1.77–1.99) | |||||||
Rurality | ||||||||||
Urban | 1 | 1 | ||||||||
Minor urban | 1.00 (0.93–1.08) | 1.00 (0.93–1.07) | ||||||||
Rural and other | 0.91 (0.84–0.99) | 0.91 (0.84–0.98) | ||||||||
Neighbourhood deprivation | ||||||||||
Least deprived | 1 | 1 | ||||||||
Quintile 2 | 1.07 (0.99–1.16) | 1.09 (1.01–1.17) | ||||||||
Quintile 3 | 1.20 (1.11–1.29) | 1.21 (1.12–1.30) | ||||||||
Quintile 4 | 1.27 (1.18–1.37) | 1.27 (1.18–1.37) | ||||||||
Most deprived | 1.45 (1.34–1.57) | 1.42 (1.31–1.53) | ||||||||
Neighbourhood volunteerism | ||||||||||
Lowest | 1.09 (1.01–1.18) | 1.03 (0.95–1.10) | 0.92 (0.85–1.01) | 1.33 (1.23–1.44) | 0.96 (0.88–1.04) | |||||
Med-low | 1.07 (0.99–1.15) | 1.01 (0.94–1.08) | 0.94 (0.87–1.02) | 1.23 (1.15–1.32) | 0.97 (0.90–1.04) | |||||
Medium | 1.05 (0.97–1.13) | 1.00 (0.93–1.07) | 0.94 (0.87–1.01) | 1.14 (1.06–1.22) | 0.95 (0.89–1.02) | |||||
Med-high | 1.07 (0.99–1.15) | 1.03 (0.96–1.11) | 0.99 (0.92–1.06) | 1.14 (1.07–1.21) | 1.01 (0.94–1.08) | |||||
Highest | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (standard error of variance) | ||||||||||
Individual-level | 0.946 (0.0012) | 0.967 (0.0012) | 0.975 (0.0012) | 0.979 (0.0012) | 0.996 (0.0012) | |||||
Neighbourhood-level | 0.063 (0.0075) | 0.043 (0.0066) | 0.031 (0.0060) | |||||||
Regional-level | 0.027 (0.0070) | 0.007 (0.0029) |
. | Neighbourhood-level random error . | . | . | Regional-level random error . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Variable . | Model 1: Baseline . | Model 2: plus individual-level covariates . | Model 3: plus neighbourhood-level covariates . | Model 1: Baseline . | Model 3: plus neighbourhood-level covariates . | |||||
Ethnicity | ||||||||||
Mäori | 2.50 (2.36–2.65) | 2.26 (2.13–2.39) | 2.14 (2.01–2.27) | 2.50 (2.36–2.65) | 2.14 (2.02–2.28) | |||||
Pacific people | 1.76 (1.56–1.98) | 1.55 (1.38–1.75) | 1.47 (1.30–1.65) | 1.82 (1.62–2.05) | 1.46 (1.29–1.65) | |||||
Non-Mäori Non-Pacific | 1 | 1 | 1 | 1 | 1 | |||||
Marital status | ||||||||||
Married | 1 | 1 | 1 | 1 | 1 | |||||
Not married | 1.26 (1.21–1.31) | 1.11 (1.06–1.16) | 1.11 (1.06–1.16) | 1.26 (1.21–1.31) | 1.11 (1.06–1.16) | |||||
Logarithm of household income (min $1000) | 0.92 (0.89–0.95) | 0.94 (0.90–0.97) | 0.94 (0.90–0.97) | |||||||
Education | ||||||||||
No qualifications | 1.23 (1.17–1.29) | 1.20 (1.14–1.27) | 1.20 (1.14–1.27) | |||||||
School qualifications | 1.12 (1.06–1.19) | 1.12 (1.05–1.18) | 1.12 (1.05–1.18) | |||||||
Post-school qualifications | 1 | 1 | 1 | |||||||
Car access | ||||||||||
Nil cars | 1.45 (1.36–1.55) | 1.40 (1.31–1.50) | 1.40 (1.30–1.50) | |||||||
1 car | 1.11 (1.05–1.16) | 1.08 (1.03–1.14) | 1.08 (1.03–1.14) | |||||||
2 or more cars | 1 | 1 | 1 | |||||||
Labour force | ||||||||||
Employed | 1 | 1 | 1 | |||||||
Unemployed | 1.09 (0.92–1.28) | 1.08 (0.91–1.27) | 1.08 (0.91–1.28) | |||||||
Non-labour | 1.88 (1.78–2.00) | 1.87 (1.77–1.99) | 1.88 (1.77–1.99) | |||||||
Rurality | ||||||||||
Urban | 1 | 1 | ||||||||
Minor urban | 1.00 (0.93–1.08) | 1.00 (0.93–1.07) | ||||||||
Rural and other | 0.91 (0.84–0.99) | 0.91 (0.84–0.98) | ||||||||
Neighbourhood deprivation | ||||||||||
Least deprived | 1 | 1 | ||||||||
Quintile 2 | 1.07 (0.99–1.16) | 1.09 (1.01–1.17) | ||||||||
Quintile 3 | 1.20 (1.11–1.29) | 1.21 (1.12–1.30) | ||||||||
Quintile 4 | 1.27 (1.18–1.37) | 1.27 (1.18–1.37) | ||||||||
Most deprived | 1.45 (1.34–1.57) | 1.42 (1.31–1.53) | ||||||||
Neighbourhood volunteerism | ||||||||||
Lowest | 1.09 (1.01–1.18) | 1.03 (0.95–1.10) | 0.92 (0.85–1.01) | 1.33 (1.23–1.44) | 0.96 (0.88–1.04) | |||||
Med-low | 1.07 (0.99–1.15) | 1.01 (0.94–1.08) | 0.94 (0.87–1.02) | 1.23 (1.15–1.32) | 0.97 (0.90–1.04) | |||||
Medium | 1.05 (0.97–1.13) | 1.00 (0.93–1.07) | 0.94 (0.87–1.01) | 1.14 (1.06–1.22) | 0.95 (0.89–1.02) | |||||
Med-high | 1.07 (0.99–1.15) | 1.03 (0.96–1.11) | 0.99 (0.92–1.06) | 1.14 (1.07–1.21) | 1.01 (0.94–1.08) | |||||
Highest | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (standard error of variance) | ||||||||||
Individual-level | 0.946 (0.0012) | 0.967 (0.0012) | 0.975 (0.0012) | 0.979 (0.0012) | 0.996 (0.0012) | |||||
Neighbourhood-level | 0.063 (0.0075) | 0.043 (0.0066) | 0.031 (0.0060) | |||||||
Regional-level | 0.027 (0.0070) | 0.007 (0.0029) |
All models additionally adjust for age in 5-year categories.
Adjusting for potential individual-level confounders in model 2 the association of volunteerism with mortality essentially halved for males and reduced almost to the null for females (Tables 2 and 3). Finally, the addition of rurality and neighbourhood deprivation caused any remaining association to disappear.
The final two columns of Tables 2 and 3 show the results for models 1 and 3 with the random error term specified at the regional-level. The model 1 association of volunteerism with mortality was stronger (e.g. a third higher than mortality in lower volunteerism neighbourhoods) than in the models with a random error at the neighbourhood-level. However, once covariates are adjusted for (i.e. model 3), there is essentially no difference in the volunteerism–mortality association between models with neighbourhood-level or regional-level random error terms. This latter equivalence gave us confidence that using a regional-level random error term for the cause-specific mortality analyses would be reliable—so long as covariates were adjusted for (i.e. model 3). Table 4 shows these results by cause of death. There was no association of volunteerism with any cause of death, with the possible exception of a protective association of low volunteerism with male unintentional injury. That said, there is little overall evidence that the association (null or otherwise) of volunteerism with all-cause and cause-specific mortality varies by sex given the similar patterns by sex and largely overlapping CIs between male and female rate ratios.
. | All-cause . | Cancer . | Cardiovascular disease . | Unintentional injury . | Suicide . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Males . | . | . | . | . | . | |||||
Neighbourhood volunteerism | ||||||||||
Low volunteerism | 0.95 (0.89–1.02) | 0.98 (0.88–1.10) | 1.00 (0.90–1.12) | 0.60 (0.44–0.82) | 0.89 (0.64–1.22) | |||||
Medium-low | 0.98 (0.92–1.04) | 1.01 (0.92–1.12) | 1.01 (0.92–1.11) | 0.83 (0.64–1.09) | 0.89 (0.67–1.18) | |||||
Medium | 1.00 (0.95–1.06) | 1.01 (0.92–1.11) | 1.03 (0.95–1.13) | 0.94 (0.74–1.20) | 0.87 (0.66–1.14) | |||||
Medium-high | 1.01 (0.96–1.07) | 0.98 (0.90–1.07) | 1.08 (0.99–1.17) | 0.95 (0.76–1.18) | 0.95 (0.74–1.23) | |||||
High volunteerism | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (SE of variance) | ||||||||||
Individual-level | 0.952 (0.0012) | 0.964 (0.0012) | 0.947 (0.0012) | 0.967 (0.0012) | 0.960 (0.0012) | |||||
Regional-level | 0.006 (0.0022) | 0.009 (0.0044) | 0.014 (0.0058) | 0.118 (0.0427) | 0.097 (0.0501) | |||||
Females | ||||||||||
Neighbourhood volunteerism | ||||||||||
Low volunteerism | 0.96 (0.88–1.04) | 1.00 (0.89–1.12) | 0.87 (0.75–1.02) | 0.85 (0.51–1.44) | 0.57 (0.31–1.05) | |||||
Medium-low | 0.97 (0.90–1.04) | 0.97 (0.87–1.08) | 0.94 (0.81–1.08) | 0.80 (0.50–1.28) | 0.93 (0.56–1.53) | |||||
Medium | 0.95 (0.89–1.02) | 0.94 (0.85–1.03) | 0.92 (0.80–1.06) | 1.12 (0.73–1.72) | 0.80 (0.50–1.29) | |||||
Medium-high | 1.01 (0.94–1.08) | 1.02 (0.93–1.12) | 0.91 (0.80–1.04) | 1.11 (0.74–1.67) | 0.92 (0.59–1.43) | |||||
High volunteerism | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (SE of variance) | ||||||||||
Individual-level | 0.996 (0.0012) | 0.977 (0.0012) | 1.049 (0.0013) | 0.926 (0.0011) | 0.825 (0.0010) | |||||
Regional-level | 0.007 (0.0029) | 0.006 (0.0042) | 0.008 (0.0058) | 0.165 (0.0992) | 0.397 (0.1590) |
. | All-cause . | Cancer . | Cardiovascular disease . | Unintentional injury . | Suicide . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Males . | . | . | . | . | . | |||||
Neighbourhood volunteerism | ||||||||||
Low volunteerism | 0.95 (0.89–1.02) | 0.98 (0.88–1.10) | 1.00 (0.90–1.12) | 0.60 (0.44–0.82) | 0.89 (0.64–1.22) | |||||
Medium-low | 0.98 (0.92–1.04) | 1.01 (0.92–1.12) | 1.01 (0.92–1.11) | 0.83 (0.64–1.09) | 0.89 (0.67–1.18) | |||||
Medium | 1.00 (0.95–1.06) | 1.01 (0.92–1.11) | 1.03 (0.95–1.13) | 0.94 (0.74–1.20) | 0.87 (0.66–1.14) | |||||
Medium-high | 1.01 (0.96–1.07) | 0.98 (0.90–1.07) | 1.08 (0.99–1.17) | 0.95 (0.76–1.18) | 0.95 (0.74–1.23) | |||||
High volunteerism | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (SE of variance) | ||||||||||
Individual-level | 0.952 (0.0012) | 0.964 (0.0012) | 0.947 (0.0012) | 0.967 (0.0012) | 0.960 (0.0012) | |||||
Regional-level | 0.006 (0.0022) | 0.009 (0.0044) | 0.014 (0.0058) | 0.118 (0.0427) | 0.097 (0.0501) | |||||
Females | ||||||||||
Neighbourhood volunteerism | ||||||||||
Low volunteerism | 0.96 (0.88–1.04) | 1.00 (0.89–1.12) | 0.87 (0.75–1.02) | 0.85 (0.51–1.44) | 0.57 (0.31–1.05) | |||||
Medium-low | 0.97 (0.90–1.04) | 0.97 (0.87–1.08) | 0.94 (0.81–1.08) | 0.80 (0.50–1.28) | 0.93 (0.56–1.53) | |||||
Medium | 0.95 (0.89–1.02) | 0.94 (0.85–1.03) | 0.92 (0.80–1.06) | 1.12 (0.73–1.72) | 0.80 (0.50–1.29) | |||||
Medium-high | 1.01 (0.94–1.08) | 1.02 (0.93–1.12) | 0.91 (0.80–1.04) | 1.11 (0.74–1.67) | 0.92 (0.59–1.43) | |||||
High volunteerism | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (SE of variance) | ||||||||||
Individual-level | 0.996 (0.0012) | 0.977 (0.0012) | 1.049 (0.0013) | 0.926 (0.0011) | 0.825 (0.0010) | |||||
Regional-level | 0.007 (0.0029) | 0.006 (0.0042) | 0.008 (0.0058) | 0.165 (0.0992) | 0.397 (0.1590) |
. | All-cause . | Cancer . | Cardiovascular disease . | Unintentional injury . | Suicide . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Males . | . | . | . | . | . | |||||
Neighbourhood volunteerism | ||||||||||
Low volunteerism | 0.95 (0.89–1.02) | 0.98 (0.88–1.10) | 1.00 (0.90–1.12) | 0.60 (0.44–0.82) | 0.89 (0.64–1.22) | |||||
Medium-low | 0.98 (0.92–1.04) | 1.01 (0.92–1.12) | 1.01 (0.92–1.11) | 0.83 (0.64–1.09) | 0.89 (0.67–1.18) | |||||
Medium | 1.00 (0.95–1.06) | 1.01 (0.92–1.11) | 1.03 (0.95–1.13) | 0.94 (0.74–1.20) | 0.87 (0.66–1.14) | |||||
Medium-high | 1.01 (0.96–1.07) | 0.98 (0.90–1.07) | 1.08 (0.99–1.17) | 0.95 (0.76–1.18) | 0.95 (0.74–1.23) | |||||
High volunteerism | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (SE of variance) | ||||||||||
Individual-level | 0.952 (0.0012) | 0.964 (0.0012) | 0.947 (0.0012) | 0.967 (0.0012) | 0.960 (0.0012) | |||||
Regional-level | 0.006 (0.0022) | 0.009 (0.0044) | 0.014 (0.0058) | 0.118 (0.0427) | 0.097 (0.0501) | |||||
Females | ||||||||||
Neighbourhood volunteerism | ||||||||||
Low volunteerism | 0.96 (0.88–1.04) | 1.00 (0.89–1.12) | 0.87 (0.75–1.02) | 0.85 (0.51–1.44) | 0.57 (0.31–1.05) | |||||
Medium-low | 0.97 (0.90–1.04) | 0.97 (0.87–1.08) | 0.94 (0.81–1.08) | 0.80 (0.50–1.28) | 0.93 (0.56–1.53) | |||||
Medium | 0.95 (0.89–1.02) | 0.94 (0.85–1.03) | 0.92 (0.80–1.06) | 1.12 (0.73–1.72) | 0.80 (0.50–1.29) | |||||
Medium-high | 1.01 (0.94–1.08) | 1.02 (0.93–1.12) | 0.91 (0.80–1.04) | 1.11 (0.74–1.67) | 0.92 (0.59–1.43) | |||||
High volunteerism | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (SE of variance) | ||||||||||
Individual-level | 0.996 (0.0012) | 0.977 (0.0012) | 1.049 (0.0013) | 0.926 (0.0011) | 0.825 (0.0010) | |||||
Regional-level | 0.007 (0.0029) | 0.006 (0.0042) | 0.008 (0.0058) | 0.165 (0.0992) | 0.397 (0.1590) |
. | All-cause . | Cancer . | Cardiovascular disease . | Unintentional injury . | Suicide . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Males . | . | . | . | . | . | |||||
Neighbourhood volunteerism | ||||||||||
Low volunteerism | 0.95 (0.89–1.02) | 0.98 (0.88–1.10) | 1.00 (0.90–1.12) | 0.60 (0.44–0.82) | 0.89 (0.64–1.22) | |||||
Medium-low | 0.98 (0.92–1.04) | 1.01 (0.92–1.12) | 1.01 (0.92–1.11) | 0.83 (0.64–1.09) | 0.89 (0.67–1.18) | |||||
Medium | 1.00 (0.95–1.06) | 1.01 (0.92–1.11) | 1.03 (0.95–1.13) | 0.94 (0.74–1.20) | 0.87 (0.66–1.14) | |||||
Medium-high | 1.01 (0.96–1.07) | 0.98 (0.90–1.07) | 1.08 (0.99–1.17) | 0.95 (0.76–1.18) | 0.95 (0.74–1.23) | |||||
High volunteerism | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (SE of variance) | ||||||||||
Individual-level | 0.952 (0.0012) | 0.964 (0.0012) | 0.947 (0.0012) | 0.967 (0.0012) | 0.960 (0.0012) | |||||
Regional-level | 0.006 (0.0022) | 0.009 (0.0044) | 0.014 (0.0058) | 0.118 (0.0427) | 0.097 (0.0501) | |||||
Females | ||||||||||
Neighbourhood volunteerism | ||||||||||
Low volunteerism | 0.96 (0.88–1.04) | 1.00 (0.89–1.12) | 0.87 (0.75–1.02) | 0.85 (0.51–1.44) | 0.57 (0.31–1.05) | |||||
Medium-low | 0.97 (0.90–1.04) | 0.97 (0.87–1.08) | 0.94 (0.81–1.08) | 0.80 (0.50–1.28) | 0.93 (0.56–1.53) | |||||
Medium | 0.95 (0.89–1.02) | 0.94 (0.85–1.03) | 0.92 (0.80–1.06) | 1.12 (0.73–1.72) | 0.80 (0.50–1.29) | |||||
Medium-high | 1.01 (0.94–1.08) | 1.02 (0.93–1.12) | 0.91 (0.80–1.04) | 1.11 (0.74–1.67) | 0.92 (0.59–1.43) | |||||
High volunteerism | 1 | 1 | 1 | 1 | 1 | |||||
Random variance (SE of variance) | ||||||||||
Individual-level | 0.996 (0.0012) | 0.977 (0.0012) | 1.049 (0.0013) | 0.926 (0.0011) | 0.825 (0.0010) | |||||
Regional-level | 0.007 (0.0029) | 0.006 (0.0042) | 0.008 (0.0058) | 0.165 (0.0992) | 0.397 (0.1590) |
Sensitivity analyses excluding people who had not been usually resident in the same neighbourhood for at least 5 years prior to census night did not substantially alter findings. Also, dropping either labour force status or car access from model 3 (variables that one might argue are perhaps influenced by volunteerism, and not ‘just’ confounders) did not substantially change the results. Finally, we found no evidence of significant variation in the rate ratio association of volunteerism with mortality by third variables of age, ethnicity, or household income (i.e. no cross-level interaction; results available from authors).
Discussion
This study presents to our knowledge the largest multilevel study of social capital and mortality, albeit for one possible measure of social capital (i.e. neighbourhood volunteerism). Adjusting for a range of possible confounders, we found no significant association of volunteerism with mortality among 25–77 year olds in New Zealand. Some of the variables we modelled as confounders might be considered as mediating variables between social capital and mortality. Excluding the two variables we identified as most likely to potentially also be mediating variables (labour force status and car access) did not substantively change the association of volunteerism with mortality.
Our measure of social capital is derived from questions on voluntary activities answered by the full population at census night. In contrast, other studies have used only survey data to create neighbourhood-level measures of social capital that—because of small numbers of respondents in each area, or not all neighbourhoods being included in the survey—reduces the accuracy of the measurement of social capital and may reduce the number of neighbourhoods that can be included in analyses. A relative weakness of our structural measure of social capital is that it only measures volunteering rather than the actual presence of organizations. It also does not specifically measure voluntary activity in one's neighbourhood, meaning that we are implicitly assuming that an aggregated measure of peoples' overall voluntary behaviour is highly correlated with: within neighbourhood volunteerism (assuming actual neighbourhood volunteerism is the target variable to measure) or neighbourhood social capital more generally construed. Finally, the voluntary activity variables we used in the construction of our index only record yes/no responses, not intensity of activity.
As indicated in the Introduction section, our study cannot refute the possibility that other aspects of social capital (e.g. trust, direct measures of linking social capital), or social capital measured at different levels of aggregation (e.g. regions or national), or neighbourhood-level volunteerism in another context might actually be associated with mortality or health more generally. Regarding context, New Zealand during the late 1990s was emerging from a period of major structural reform including deregulation and privatization, and state sector reform. Associated with this reform income inequality increased markedly from the late 1980s to the early 1990s31—but did not reach the levels of the US. While there had been some retrenchment by the state and devolution to communities in the provision of social services in New Zealand, neighbourhood and regional-level variation in social capital may not have been as important for health outcomes as in other countries. Indeed, just as the association of income inequality with mortality has only reliably been demonstrated in the US,32,33 it might also be that social capital is also emerging as a health determinant in the US.34 Put another way, context may be everything. Living in a country with a secure welfare system and universal provision of health and other social services may render neighbourhood social capital less important as a health determinant.
While mortality is a robust outcome to measure, it is likely to be removed in time from the relevant period of ‘exposure’ to social capital. Unless the volunteerism of the neighbourhood people lived in during 1996 was a good proxy for the social capital of neighourhoods over peoples' lifecourse, our null findings might be attributed to a misclassification bias of lifecourse social capital. We partially tested for this possibility by excluding people who had not lived at the same residence for at least 5 years, but no association of neighbourhood volunteerism with mortality was disclosed. Any association of social capital with health may only be detectable for morbidity and health-related behaviours. For example, studies have found an association of: social capital as an index of civic trust, reciprocity and voluntary group membership with self-reported health35; collective efficacy, and self-rated health36 and self-reported respiratory disease.37 A recent systematic review of social capital and mental health studies found seven studies conceptualizing and measuring social capital as an ecological-level variable that fitted their inclusion criteria.18 However, the majority of study findings were null.
The issue of what it is about neighbourhoods that influences health, and the likelihood that social features of neighbourhoods are important, remains a challenge. It still seems highly plausible that neighbourhood (or more generally ‘community’) social environments matter for the myriad of pathways and mechanisms that lead to poor or good health. Future quantitative research on the social features of neighbourhoods should, in our view, be cognizant of the wider regional or national contexts, and undertake improved and alternative measurement of the social environment (e.g. careful measurement of linking–bonding–bridging social capital, measures of social fragmentation). Using continuously measured outcomes (e.g. mental health rather than suicide death) would improve statistical power. Finally, given problems with time lags and long causal chains between social capital, and cancer and cardiovascular death, we also suggest considering intermediate outcomes (e.g. physiological variables, behaviours).
Ethical approval
Ethical approval for the NZCMS was obtained by the Wellington Regional Ethics Committee.
Summary statistics New Zealand security statement
(The full security statement is available at http://www.wnmeds.ac.nz/nzcms-info.html). The New Zealand Census Mortality Study (NZCMS) is a study of the relationship between socioeconomic factors and mortality in New Zealand, based on the integration of anonymized population census data from Statistics New Zealand and mortality data from the New Zealand Health Information Service. The project was approved by Statistics New Zealand as a Data Laboratory project under the Microdata Access Protocols in 1997. The datasets created by the integration process are covered by the Statistics Act and can be used for statistical purposes only. Only approved researchers who have signed Statistics New Zealand's declaration of secrecy can access the integrated data in the Data Laboratory. For further information about confidentiality matters with regard to this study please contact Statistics New Zealand.
The NZCMS is conducted in collaboration with Statistics New Zealand and within the confines of the Statistics Act 1975. Robyn Green, Jennifer Martin, and Ichiro Kawachi helped to develop the social capital index. Comments on a final draft of this paper were received from Gavin Turrell, Jamie Pearce, SV Subramanian, Sarah Wamala, Karen Witten, Anne Kavanagh, Anna Ziersch, and Ichiro Kawachi. This study was funded by the Health Research Council of New Zealand and the Ministry of Health.
References
Macintyre S, Maciver S, Sooman A. Area, class and health: Should we be focusing on places or people?
Szreter S, Woolcock M. Health by association? Social capital, social theory, and the political economy of public health.
Muntaner C. Commentary: Social capital, social class, and the slow progress of psychosocial epidemiology.
Kawachi I, Kim D, Coutts A, Subramanian SV. Commentary: Reconciling the three accounts of social capital.
Kawachi I, Berkman L. Social cohesion, social capital, and health. In: Berkman L, Kawachi I (eds). Social Epidemiology. New York: Oxford University Press,
Kawachi I, Kennedy BP, Lochner S, Prothrow-Stith D. Social capital, income inequality, and mortality.
Lochner KA, Kawachi I, Brennan RT, Buka SL. Social capital and neighborhood mortality rates in Chicago.
Skrabski A, Kopp M, Kawachi I. Social capital in a changing society: cross sectional associations with middle aged female and male mortality rates.
Skrabski A, Kopp M, Kawachi I. Social capital and collective efficacy in Hungary: cross sectional associations with middle aged female and male mortality rates.
Diez Roux A. Bringing context back into epidemiology: variables and fallacies in multilevel analysis.
Blakely T, Woodward A. Ecological effects in multi-level studies.
Mohan J, Twigg L, Barnard S, Jones K. Social capital, geography and health: a small-area analysis for England.
Franzini L, Spears W. Contributions of social context to inequalities in years of life lost to heart disease in Texas, USA.
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.
Turrell G, Kavanagh A, Subramanian SV. Area variation in mortality in Tasmania (Australia): the contributions of socioeconomic disadvantage, social capital and geographic remoteness.
Martikainen P, Maki N, Blomgren J. The effects of area and individual social characteristics of suicide risk: a multilevel study of relative contribution and effect modification.
De Silva MJ, McKenzie K, Harpham T, Huttly SRA. Social capital and mental illness: a systematic review.
Kavanagh A, Turrell G, Subramanian SV. Does area-based social capital matter for the health of Australians? A multilevel analysis of self-rated health.
Gatrell AC, Popay J, Thomas C. Mapping the determinants of health inequalities in social space: can Bourdieu help us?
Ziersch AM, Baum FE, MacDougall C, Putland C. Neighbourhood life and social capital: the implications for health.
Berkman L, Glass T. Social integration, social networks, social support, and health. In: Berkman L, Kawachi I (eds). Social Epidemiology. New York: Oxford University Press,
Subramanian SV, Kim DJ, Kawachi I. Social trust and self-rated health in US communities: a multilevel analysis.
Hill S, Atkinson J, Blakely T. Anonymous record linkage of census and mortality records: 1981, 1986, 1991, 1996 census cohorts. NZCMS Technical Report No. 3. Wellington: Department of Public Health, Wellington School of Medicine and Health Sciences, University of Otago,
Blakely T, Salmond C. Probabilistic record linkage and a method to calculate the positive predictive value.
Fawcett J, Blakely T, Atkinson J. Weighting the 81, 86, 91 & 96 census-mortality cohorts to adjust for linkage bias. NZCMS Technical Report No. 5. Wellington: Department of Public Health, Wellington School of Medicine and Health Sciences, University of Otago,
Blakely T, Kawachi I, Atkinson J, Fawcett J. Income and mortality: the shape of the association and confounding New Zealand Census-Mortality Study, 1981–1999.
Norm Breslow, “Whither PQL?” (January 24, 2003). UW Biostatistics Working Paper Series. Working Paper 192. http://www.bepress.com/uwbiostat/paper192
Subramanian SV, Kawachi I. Income inequality and health: What have we learned so far?
Lynch J, Davey Smith G, Harper S et al. Is income inequality a determinant of population health? Part 1. A systematic review.
Kawachi I, Kim D, Coutts A, Subramanian S. Commentary: Reconciling the three accounts of social capital.
Kawachi I, Kennedy BP, Glass R. Social capital and self-rated health: a contextual analysis.
Browning C, Cagney K. Neighbourhood structural disadvantage, collective efficacy, and self-rated physical health in an urban setting.