Abstract
We assessed whether social isolation (SI), social support (SS), and subtypes of SS were associated with self-rated health trajectories and clinical heart failure (HF) outcomes among participants with incident HF hospitalizations. We included 2967 Atherosclerosis Risk in Communities study participants with incident HF hospitalization after Visit 2 (1990-1992). SI, SS, and subtypes of SS were measured at Visit 2. We identified incident HF hospitalization as ICD-9 code 428 and physician adjudicated events; on average HF occurred 17 (SD 8) years after Visit 2. We assessed associations with trajectories of annually measured self-rated health in the 4 years prior to and after incident HF hospitalization (excellent/good self-rated health on a 0-100 scale), using linear mixed effects models. We calculated hazard ratios (HR) and 95% confidence intervals (CIs) for associations with time to first all-cause rehospitalization and all-cause mortality using Cox proportional hazard models. Low overall SS had a 5.8 point (95% CI 7.8, 3.8) lower self-rated health value over time than high SS; associations of subtypes of SS with this outcome were similar. Low belonging SS was associated with greater days to first rehospitalization (HR 0.85; 95% CI 0.79, 0.96) compared to the highest tertile; however, belonging SS was not associated with mortality (HR 1.05; 95% CI 0.95, 1.17). Being socially isolated/high risk for SI was associated with greater hazard of all-cause mortality among females (HR 1.57; 95% CI 1.20, 2.06) but not males (HR 0.95; 95% CI 0.75, 1.19), compared to low SI. SI and SS were not associated with number of hospitalizations in the first year or percent of first year spent at home.