This study used 23 factors (eight interval variables and 15 dummy variables) as proxies for health depreciation. We used 1248 older adults from the Center for Geriatrics and Gerontology of Taichung Veterans General Hospital (Taiwan) to examine the association among frailty, health depreciation, and mortality in older adults. This study found that a significant positive correlation existed between frailty and mortality in older adults. Further, we applied a recursive bivariate probit model to examine the association between health depreciation factors, frailty, and mortality. Our results showed that health depreciation factors, such as Charlson's comorbidity index, diabetes and hyperlipidemia, significantly increased older adults' frailty; in contrast, albumin and mini nutritional assessment significantly decreased older adults' frailty. Through the frailty regression, we confirmed not only that health depreciation factors significantly influenced mortality, but also that creatinine, myocardial infarction, and malignant tumors could directly and significantly increase older adults' mortality.
Keywords: frailty; health depreciation factors; mortality; recursive probit regression model.