Change in social ties and subsequent mortality in rural elders

James R. Cerhan, Robert B. Wallace

Research output: Contribution to journalArticle

46 Scopus citations


Prior studies assessing social ties and mortality risk in older persons have not considered the dynamic nature of social ties, health, and functional status in this age group. We evaluated the association between change in social ties (assessed in 1982 and 1985) and subsequent mortality through 1993 in a population-based cohort of 2,575 rural persons ages 65-102 years. The same summary measure of social ties was constructed for both interviews and included marital status, number of close friends and relatives, church attendance, and membership in a group. Compared with persons reporting higher levels of social ties at both interviews, a decrease to a low level of social ties was associated with an increased mortality risk for both men [hazard ratio (HR) = 1.8; 95% confidence interval (CI) = 1.2-2.7] and women (HR = 2.1; 95% CI = 1.5-3.0). Lower levels of social ties at both interviews were also associated with greater mortality risk in men (HR - 2.2; 95% CI = 1.5- 3.1) and women (HR = 2.2; 95% CI = 1.6-2.9), whereas an increase to a higher level of social ties was not associated with subsequent mortality risk. Adjustment for age, education, smoking status, cognitive function, onset of a major illness, change in physical function, change in self-perceived health status, and change in depressive symptoms eliminated the association between loss of social ties and survival, but both men (HR = 2.3; 95% CI = 1.5-3.5) and women (HR = 1.8; 95% CI = 1.2-2.6) reporting low levels of social ties at both interviews remained at elevated risk. The lack of an association between loss of social ties and mortality risk after adjusting for demographic, health, and functional status variables indicates that these are either confounding or intervening variables in a social ties and mortality association. In contrast, sustained low levels of social ties appear to influence mortality risk independent of these same variables and imply that continued social isolation may be a more important determinant of mortality risk than recent changes in social ties.

Original languageEnglish (US)
Pages (from-to)475-481
Number of pages7
Issue number5
StatePublished - Sep 1 1997



  • Aging
  • Longitudinal studies
  • Mortality
  • Population studies
  • Social networks

ASJC Scopus subject areas

  • Epidemiology

Cite this