Association between loneliness and social isolation and health outcomes among cancer survivors and non-cancer controls.

Introduction: The present analyses take advantage of an existing cohort of cancer survivors (cases) and women without a history of cancer (controls) to examine whether older cancer survivors experience greater loneliness or social isolation than women without a history of cancer and how loneliness and social isolation are related to health outcomes. Methods: Cross-sectional analyses were conducted from Visit 17 of The Study of Women's Health Across the Nation (SWAN), a multiracial/ethnic cohort study initiated to study the biological and psychosocial changes occurring during the menopausal transition. We identified 276 women who had developed cancer over the 30 years of SWAN (eligible cases) and 1,123 women who never had cancer. Loneliness was measured using the UCLA 3-item loneliness scale. Social isolation was assessed with a modified version of the Social Network Index. Outcomes included health-related quality of life (HRQL) as assessed by the SF-36 (MCS and PCS), depressive symptoms (CES-D), and anxiety. Covariates included sociodemographic, health-related, and psychosocial variables. Associations of case-control status, years since diagnosis in cases, loneliness, and social isolation with MCS and PCS were estimated using analysis of covariance, and with depressive symptoms and high anxiety using logistic regression. Effect modification of loneliness and social isolation by case-control status was assessed by adding relevant interaction terms to models. Results: Case-control status was not significantly related to loneliness or social isolation. Loneliness and social isolation were negatively related to PCS before, but not after covariate adjustment. Corresponding negative associations with MCS remained statistically significant after covariate adjustment. Loneliness and social isolation were positively associated with depressive symptoms. Loneliness, but not social isolation, was positively associated with high anxiety. With the exception of a significantly stronger unadjusted association of PCS with loneliness in controls than in cases, there was no statistically significant effect modification by case-control status before or after covariate adjustment. Conclusion: Neither loneliness nor social isolation was related to cancer survivorship status. Although both loneliness and social isolation were related to HRQL and mental health outcomes in a cohort of older women, these associations were similar for cancer survivors and those without a history of cancer.