Trust, self-rated health and mortality: a longitudinal study among ageing people in Southern Finland.

Poor self-rated health (SRH) predicts mortality significantly. High trust has been shown to associate with better SRH in cross-sectional studies and survival in longitudinal studies. However, little is known about the associations between trust, SRH and mortality among ageing people. The present study examined whether low trust at the baseline predicted higher all-cause mortality in a follow-up of over five years among ageing people, and whether the trust to mortality relationship varied by SRH. The study used longitudinal, questionnaire-based survey data gathered in 2002 (n = 2815; 66%) among three age cohorts (born in 1926-30, 1936-40 and 1946-50) living in the Lahti region, Southern Finland. Two survey follow-ups were done, the first in 2005 (n = 2476, 60%) and the second in 2008 (n = 2064, 73%). Deaths during the follow-up were obtained from the covering National Population Registry. Those who died within the first one year of follow-up were excluded from the analyses to reduce potential bias due to early deaths. Cox proportional hazard models were used to derive the results. Mortality proved to be higher among men with low trust, even if their SRH had been good at the baseline. Among women, no significant associations were found. The risk attenuated after adjustment for background health-related covariates, but the gradient remained statistically significant in all models. Initial SRH did not substantially explain the gradient of trust in mortality among men. Moreover, a Sobel test of indirect effects showed that SRH had no significant mediating role in links between trust and mortality. Thus, low trust is a sensitive indicator of higher mortality risk among ageing men.