| Social isolation is a significant psychosocial problem for older adults that has been linked with numerous negative health morbidities such as increased cardiovascular risk (Brummett et al., 2001), falls (Faulkner, Cauley, Zmuda, Griffin, & Nevitt, 2003), cognitive decline (Fratiglioni, Paillard-Borg, & Winblad, 2004) and increased mortality (Berkman & Syme, 1979; Ceria et al., 2001).;The main goal of this dissertation was to determine the associations and/or direction of independent variables with social isolation and to use those associated independent variables to model prediction of social isolation over time. Additionally, point prevalence, period prevalence, and incidence rate of social isolation were examined.;The design for this study was a secondary data analysis of the Yale Health and Aging Project (YHAP) data. YHAP data were prospective longitudinal over a 12 year time period from 1982--1994. Using the Social Network Index (SNI) as the dependent variable, data from face-to-face interviews in, 1982, 1985, 1988, and 1994 were examined. For this study 2,057 participants met the baseline inclusion criteria.;Using generalized estimating equations (GEE), findings showed numerous predictor variables of social isolation. These predictor variables were religious engagement, depressive symptoms gender , marital status, overall health, smoking status, retirement status, life events (having a close friend die) and live alone. One of the most compelling predictors of social isolation was religious engagement. When examining religious engagement, for each increasing level of social isolation those who knew many had an OR of 1.81 (95% CI 1.04--2.19), a few had an OR of 4.96 (95% CI 3.97--6.22), and those who knew none had an OR of 20.55 (95% CI 16.28--25.94) when compared with those who knew almost all members of their congregation.;Point prevalence ranged from 21% at baseline to 32% in 1994 whereas 12-year period prevalence was 43%.;Through design of this study, the most glaring gaps in the literature, conceptual ambiguity, reverse causality and unknown predictors were all addressed. Findings from this study can help researchers to better understand the predictors of social isolation, allowing for construction of a profile to identify at-risk older adults and targeted interventions to reduce social isolation. |
