Health attitudes and behaviour as predictors of self-rated health in relation to mortality patterns (17-year follow-up in a Polish elderly population - Cracow study)

Abstract

The aim of this 17-year Cracow (Poland) study was to examine and identify the determinants most responsible for the relationship between self-rated health and mortality in non-institutionalized elderly people. Base-line data were collected in a simple random sample of 2,605 Cracow residents, aged 65 years and over. The vital status of all individuals under study was established by monitoring city records. Death certificates were obtained for the deceased and coded according to the underlying cause of death. Predictors related to self-rated health, developed using the results of principal component analysis, were modelled on three indexes: individual predisposition (to disease based on family history), caring about health in the past, and attitudes toward health. Cox multivariate analysis confirmed the significant role of self-rated health in the mortality patterns of women (HR = 1.18). Those who cared greatly about health in the past had a lower mortality risk than those who cared poorly (HR = 1.19). In men, a positive attitude towards health remained an independent predictor of mortality (HR = 1.20). Multivariate regression models found self-rated health to be a significant independent predictor of mortality only in women with a low level of individual predisposition (hazard ratio for self-rated health = 1.35) and in those with a positive attitude towards health (hazard ratio for self-rated health = 1.16).