Abstract
OBJECTIVES: Self-rated health is an instrumental variable to assess the overall health status of a population. However, it remains questionable whether it is still useful for cognitively impaired individuals. Therefore, this study aims to analyze whether self-rated health by the cognitively impaired predicts mortality reliably. METHODS: This study used 7,881 community-dwelling individuals, aged 45 and above, from the Korean Longitudinal Study of Aging (2006-2016). It used the Cox proportional hazard models for analysis. Cognitive status was classified based on the Korean Mini Mental State Examination score and a stratified analysis was used to determine whether the predictability of self-rated health varies according to cognitive status. RESULTS: For cognitively intact individuals, the adjusted hazard ratios (aHR) of mortality were 2.00 (95% confidence interval [CI], 1.18 to 3.41, model 4) for those with 'bad' self-rated health and 2.40 (95% CI, 1.35 to 4.25, model 4) for those with 'very bad' self-rated heath, respectively, compared with those with 'very good' health. The results remain statistically significant even after adjusting for socio-demographic factors, health status, and health-related behaviors. For cognitively impaired individuals, the aHR of mortality was statistically significant for those with 'very bad' self-rated health, compared with those with 'very good' health, when socio-demographic factors were accounted for (aHR, 3.03; 95% CI, 1.11 to 8.28, model 2). CONCLUSIONS: Self-rated health by cognitively impaired individuals remains useful in predicting mortality. It appears to be a valid and reliable health indicator for the rising population with cognitive impairment, especially caused by aging population.