BACKGROUND Subjective and objective measures of isolation have been associated with increased risk of mortality in many studies, and some have found differential effects . DATA AND
METHODS Canadian Community Health Survey-Healthy Aging data (2008/2009) linked to the Canadian Vital Statistics-Death Database were used to estimate the prevalence of social isolation measured objectively (low social participation) and subjectively (feelings of loneliness and weak sense of community belonging). Associations with death during the 8 to 9 year follow-up period were examined with multivariate Cox proportional hazards models controlling for sociodemographic and health-related characteristics . Structural equation models (SEM) examined direct paths with survival time and indirect effects through health status controlling for covariates that were significant in the Cox models . Analyses were stratified by sex .
RESULTS An estimated 525,000 people (12 %) aged 65 or older felt socially isolated and over 1 million (1,018,000) (24 %) reported low participation . In multivariate Cox models, low participation was significantly associated with death for men and women even when the potential confounding effects of subjective isolation, socio-demographic characteristics, health status, and health behaviours were considered . Subjective isolation was not associated with death in final multivariate models for men or women . SEM revealed significant associations between low participation and survival time for men and women . In addition to the direct effects, there were significant indirect effects mediated by health status . There were no direct effects of subjective isolation on survival for men or women, only indirect effects mediated through health status .
DISCUSSION Subjective and objective isolation differed in their association with mortality.