Neighborhood Characteristics, Size of Social Network and Social Cohesion, and Self-Reported Health Among Hispanics Aged 75 and over

Background: Many epidemiological studies support “Hispanic Paradox” by showing that Hispanics are healthier than non-Hispanic Whites and more acculturated Hispanics. One of theories attempts to explain the paradoxical finding argues that Hispanic immigrant enjoy better social network associated with Hispanic culture (e.g., familism) and practices (e.g., eating less processed food). Hispanics living in an ethic enclave, where many Hispanics live together, demonstrated better social support and are more integrated within a community. Recent studies also found that proportion of Hispanics in a census tract where respondents reside was inversely associated with likelihood of having chronic illnesses and alcohol abuse. Even, 7-year mortality rate was lower in Hispanic ethic enclaves. However, little is known that whether co-ethnic composition in a neighborhood is associated with health by virtue of more robust social network. We aims to examine two research questions. First, we examined whether neighborhood characteristics, proportion of Hispanics and Spanish speakers, are associated with social network’s size and social cohesion. Second, we also investigated whether adding the size and the social cohesion to a statistical model affect regression coefficients of those neighborhood-level variables in predicting variance of self-reported health.  

Methods: Individual-level attributes, including size of social network, social cohesion, self-reported general health, and confounders were retrieved wave from the wave five (2004-2005) data of the Hispanic Established Population for the Epidemiological Study of the Elderly (H-EPESE). The wave-five H-EPESE followed 1,167 participants aged 75 and over of 3,050 participants who initially started in the baseline survey (1993/1994).

 We retrieved neighborhood-level variables, the proportion of Latinos and Spanish speakers, from the 2000 U.S. Census Summary Files.

Since census tract number is available for most participants, we could link individual-level data and neighborhood-level data for analysis. Two primary analyses were conducted. First, hierarchical linear models (HLMs) were conducted to see whether neighborhood-level characteristics are associated with the size and the social cohesion. And, then, we used hierarchical generalized linear models (HGLMs) to investigate whether neighborhood-level variables and size and social cohesion are independently associated with an ordinal-level self-reported general health.

Results: Using HLMs, neighborhood-level variables were significantly associated with size of social network (p<.05) and social cohesion (p<.05). Hispanic older adults living in a census tracts where more Hispanics or Spanish speakers resided were more likely to reported larger size and increased social cohesion. In HGLMs, neighborhood-level variables were significantly associated with odds of satisfactory health (p<.05). Also, size (p<.05) and social cohesion (p<.05) were more likely to satisfactory health. Interestingly, when size and social cohesion were added to the statistical models, coefficients of neighborhood-level variables were subdued.

Conclusions and Implications: Our results suggest that Hispanic older adults benefit from living in an ethnic enclave and the benefit appears manifested by having better social network. Hispanic older adults living in a community with low co-ethnic composition appear to struggle with their social relationship. Thus, additional programs helping this group to have robust social network should be considered.