Abstract
BACKGROUND: Social engagement (SE) has been consistently shown to improve survival among community-dwelling older people, but the evidence in nursing home residents is inconclusive and prone to short-term reverse causation and confounding by major health determinants. Our main objective was to study the potential causal effect of within-the-facility social engagement (SE) on long-term all-cause mortality in care home residents. METHODS: A representative cohort of 382 nursing home residents in Madrid without severe physical and cognitive impairments at baseline was followed up for 10-year all-cause mortality. Standardized mortality curves for residents with low/null, moderate, and high levels of SE at baseline were estimated using Kaplan-Meier methods and spline-based survival models with inverse probability of exposure weights conditional on baseline sociodemographic characteristics, facility features, comorbidity, and disability. Standardized 5-year mortality risks and median survival times were compared across levels of SE. RESULTS: The baseline prevalences of low/null, moderate, and high SE were 36, 44, and 20%, respectively. Compared with residents with low/null SE at baseline, the standardized differences (95% confidence intervals) in 5-year mortality risk were - 2.3% (- 14.6 to 10.0%) for moderately engaged residents and - 18.4% (- 33.8 to - 2.9%) for highly engaged residents. The median survival time increased by 0.4 (- 1.4 to 2.2) and 3.0 (0.8 to 5.2) years, respectively. CONCLUSION: Residents with high SE within the nursing home had an 18% lower 5-year mortality risk and a 3-year increase in their median survival, as compared with residents with similar health determinants but low/null SE. The development of adequate tailored intervention programs, addressed to increase SE in nursing home residents, could improve their long-term survival, in addition to expected gains in quality of life.