[The association of self-rated health and mortality--a 7-year follow-up study of a Japanese community].

This study aims to examine whether self-rated health is an independent predictor of cause-specific mortality even after the influence by a variety of factors in relation to mortality reported by previous studies have been excluded. This study included randomly selected 4,259 inhabitants (1,827 men and 2,432 women) in Y city, Fukuoka prefecture, Japan, aged from 30 to 79 in 1987. These subjects were surveyed in 1987 by self-administered questionnaires regarding self-rated health as well as life habits, and followed up for their vital status and underlying causes of death. After excluding the individuals who were lost to follow up or who didn't respond to the question on self-rated health, 4,046 individuals were analysed to see the relationship between self-rated health and cause-specific mortality by Cox proportional hazard models, controlling for sex, age, smoking, BMI, medical care use and ADL. It was shown that relative risks for all causes, cancer, circulatory disease and other causes among the unhealthy group were 2.95 (95% CI: 1.93-4.50), 2.96 (1.53-5.73), 2.32 (0.86-6.26) and 4.09 (2.12-7.89), relatively. In the analyses of the subgroup (subjects without diseases in 1987 or subjects excluding deceased cases within first 3 years after follow-up), to avoid selection bias, the association between self-rated health and mortality was substantially similar to the results obtained in the former analysis, even the association was weakened. Even after excluding both of the subjects with diseases in 1987 and the subjects who died in the first 3 years after follow-up, self-rated health could be associated with mortality from all causes (RR = 1.89, 95%CI; 0.91-3.94). From the results it is suggested that self-rated health itself can be the independent predictor of mortality.