Aging, Social Capital, and Utilization of Health Services in Canada

Audrey LaPorte, PhD

Purpose

This study focuses on the role of "social capital" both at the individual and the community level in maintaining health as a person ages. Social capital at the individual level refers to the networks of social relations that may provide individuals and groups with access to resources and supports. At the community level, social capital can be represented by the extent of outreach on the part of community-based organizations. We posit that the impact of social capital on probability of use as compared to intensity of use will differ. Furthermore, the influence of social capital on health service utilization may be expected to differ by age cohort.

Data

The Canadian Community Health Survey (CCHS) conducted in 2000-2001 contains information about health determinants, health status and health system utilization for 133,300 people from across Canada. In particular, it has detailed information on economic, social, demographic, occupational and environmental factors related to health. Of particular interest is the relationship between age, health care utilization, and information about both religious observance (attendance at church services) and other social supports (extended family, etc.). Canadian 2001 Census was used to derive community level social capital (employment level in religious and community-based organizations [NAICS code 813XX]-a.k.a., Petris Index).

Methods

Both the likelihood of service use (physician visits and hospital stays) and the intensity of such use is modeled using a two-part framework. We control for the effect of diet, substance abuse (smoking, alcohol consumption), exercise, baseline health status/physical disability/mental health (Health Utilities Index), income, education, and labor force participation. Separate profit and OLS regressions are used to model types of health services utilization (e.g., total number of annual hospital days and general practitioner (GP) visits) as a function of both forms of social capital, and the control variables listed above. These analyses are intended to establish baseline relationships and the relative importance of each factor in cross-section -- that is, the relationship between social capital and health care utilization within various age-groups (under 45 years, 45-64 years, 65 plus years).

Results

Higher levels of individual social capital are associated with lower numbers of days of hospitalization but higher numbers of primary care physician visits. As well, higher rates of individual social capital are associated with a greater tendency to use physician services. Higher levels of community social capital are only associated with lower rates of primary care physician visits. These effects are most pronounced among the population under the age of 45. Removal of individual social capital variables has only a marginal impact on the magnitude and statistical significance of community social capital variables. These results suggest that the impact of social capital varies by age and furthermore that the form of social capital (individual versus community) has a differential impact by age cohort.

Policy Relevance

Many researchers in the field of the effect of social capital on health have reached the conclusion that for the major debates in the literature to be resolved, future work will have to be at the individual level but also longitudinally and that better measures of social capital are required. This pilot project represents a first step toward addressing these objectives.

[Back To Top]

Associations Between Individual and Community Social Capital and
Self Assessed Health in Norwegian Level of Living Surveys

Tor Iversen, PhD

Objective

The objective is to estimate the association between individual and community social capital and health when other factors are controlled for.

Data Sources

We utilize Norwegian Level of Living data from Statistics Norway. There are available cross-sectional surveys from the two years 1998 and 2002. The Level of living surveys contain information of self-assessed health, socio-demographic characteristics and some information of individual social capital.

Data on community social capital are collected from the Norwegian Commune Database, from the Norwegian Confederation of Sports and from surveys conducted by TNS Gallup.

Study Design

Individual level data are merged with community level data at the county level. Analyses are performed with self-assessed general health and mental health as dependent variables. An ordered probit model that takes into account that data are clustered in 19 counties is used in the analysis of data.

[Back To Top]

Community Social Capital and Recurrence of Acute Coronary Syndrome

Richard M. Scheffler, PhD | Timothy T. Brown, PhD | Leonard Syme, PhD | Ichiro Kawachi, MD, PhD | Irina Tolstykh, MS | Carlos Iribarren, MD, MPH, PhD

Background

Social capital has been shown to be associated with reduced mortality due to cardiovascular disease, but no study to date has examined the association of time-varying community social capital (CSC) with recurrence of cardiovascular events with simultaneous consideration of extensive patient-level data and community-level data.

Methods

We performed a retrospective cohort study of the association of CSC with recurrence of acute coronary syndrome (ACS). We identified 34,752 men and women, aged 30-85, hospitalized for ACS between Jan 1, 1998 and Dec 31, 2002 in Kaiser Permanente Northern California, an integrated healthcare delivery system. The primary outcome was recurrent non-fatal or fatal ACS; median follow-up was 19 months. We estimated random effects, multi-level Cox proportional hazard models adjusting for sex, age, race/ethnicity, median household income, comorbidities, medication use, revascularization procedures, income inequality, racial concentration and penetration of health maintenance organizations (HMO). Our measure of CSC was the previously validated Petris Social Capital Index (PSCI).

Findings

One standard-deviation increase in the PSCI after adjusting for sex, age, race/ethnicity, median household income, comorbidities, medication use, revascularization procedures, income inequality, racial concentration, and HMO penetration was associated with decreased recurrence of ACS (HR = 0.92, 95% CI, 0.88 - 0.97). Stratifying the sample by median income resulted in lower ACS recurrence only for those with less than median income ((HR = 0.91, CI 0.86 - 0.96) compared to those with median income or higher (HR = 0.96, CI 0.90 - 1.03).

Interpretation

Community social capital is inversely associated with recurrence of ACS among lower-income individuals.

[Back To Top]

The Demand for Religious Activity and Religious Social Capital:
An Application of the Grossman Model

Timothy Brown, PhD

I use a modified Grossman demand-for-health model to estimate the demand for spiritual health and religious practices. Religious health is defined as religious practices that are considered essential to spiritual health. In this case, regular attendance at church, temple, or mosque is the essential religious practice. Religious practices are defined as those which increase spiritual health, but which an individual has significant choice as to the frequency. In this case, frequency of prayer is used. The model predicts that wages will correlate negatively with both attendance and prayer, age will correlate positively with both attendance and prayer, and education will correlate positively with attendance and negatively with prayer. I use the General Social Survey for the years 1994 1996, 1998, 2000, 2002, and 2004. My findings are consistent with the theoretical predictions of the model.

[Back To Top]

The Demand for Social Interaction

Henry Saffer, PhD

In this paper social interaction is modeled as a consumer good. Social interaction may provide an externality in the form of social capital, but the primary reason that individuals engage in social interaction is that these activities directly yield utility. It is important to note that some measures of social interaction show declines while many do not. A model of household production is employed to derive the demand for social interaction. The model shows that the demand for social interaction is a function of its price, the price of other goods and income. The role of children and marriage in social interaction can also be explained in the model. The theory is tested with data from the General Social Survey (GSS) and the results show that social interaction can be explained as the consequence of utility maximizing behavior by individuals. Increases in education generally increase memberships but reduce visiting with relatives and friends. Increases in income generally increase memberships and some forms of visiting. The model predicts 70 percent, or more, of the time trends in social interaction. These results are in contrast to social capital theorists who have focused on the declines in social interaction and who have attributed these changes to factors such as increased community heterogeneity and increased television viewing.

[Back To Top]

The Empirical Relationship between Community Social Capital and the Demand for Cigarettes

Timothy T. Brown, PhD | Richard M. Scheffler, PhD | Sukyong Seo, MPH | Mary Reed, MPH

We show that the proportion of community social capital attributable to religious groups is inversely and strongly related to the number of cigarettes that smokers consume. We do not find overall community social capital or the proportion of community social capital attributable to religious groups to be related to the overall prevalence of smoking. Using a new validated measure of community social capital, the Petris Social Capital Index and three years (1998-2000) of US data on 39,369 adults, we estimate a two-part demand model incorporating the following controls: community-level fixed effects, price (including excise taxes), family income, a smuggling indicator, nonsmoking regulations, education, marital status, sex, age, and race/ethnicity. Copyright © 2006 John Wiley & Sons, Ltd.

[Back To Top]