Abstract P70: Social Network Characteristics Are Associated With Long Term Outcomes After Myocardial Infarction

Abstract

Background: Lack of social support is associated with worse outcomes after myocardial infarction (MI). Social support is a complex concept that includes the quality of perceived support and the size and quality of one's social network. It is not known if having a geographically close social network affects outcomes post-MI. We hypothesized that patients with a greater number of close network contacts would have better post-MI outcomes. Methods: From contacts listed by subjects in TRIUMPH, a prospective registry of MI patients from 24 US centers, we characterized social network size (number of contacts listed) and closeness (number of nuclear family contacts and number of contacts residing in the same area code). We assessed univariate relationships between these indices and outcomes (mortality and health status by the EQ-5D Visual Analogue Scale [VAS], an overall assessment of patients' quality of life), and scores from the ENRICHD Social Support Instrument (ESSI). We created multivariable Cox proportional hazards and linear regression models with mortality and VAS as dependent variables and with demographic, clinical, treatment and social support measures as independent variables. Results: Of 4340 subjects enrolled, 472 died over a median of 28.6 months. Mean 12-month VAS score was 75.4 ± 21.1. Having no same area code contacts (19.3% vs. 15.9%) (p=0.025 for trend), no nuclear family contacts (40.9% vs. 33.8%) (p=0.010 for trend), and living alone (33.0% vs. 23.6%, p<0.001) were associated with higher mortality on univariate analysis; total number of contacts and ESSI tertile were not. Univariate relationships with VAS were similar. In a multivariable model, age, prior stroke, diabetes, kidney disease, ejection fraction <40%, and living alone were significantly associated with mortality. In the multivariable VAS model, fewer area code contacts was additionally significant (p=0.023 for trend across number of contacts). Subjects with no area code contacts had a VAS score 3.2 ± 2.0 points lower than those with 4 near contacts. Conclusions: Geographic closeness of patients' social networks is weakly associated with outcomes after MI. Further investigation is warranted before proposing interventions designed to compensate for low network-based support.