Background: Self-rated health (SRH) predicts all-cause mortality in many studies; whereas, SRH has been inconsistently related to disease specific death, at least in part because often carefully documented cause of death is lacking.Methods: Physician-adjudicated cardiovascular disease (CVD), cancer, and other outcomes were evaluated in the Women's Health Initiative (WHI) multi-ethnic Observational Study (OS) cohort of 93,6756 postmenopausal women, aged 50 to79 years. SRH was assessed by the RAND36 at baseline and three years later.
Results:After adjusting for confounders, compared with women reporting excellent health, the risk of all-cause death among women reporting fair/poor health was significantly higher (HR=1.91, CI 1.68, 2.16) during a 7.6 year (1.6) follow-up, as were risks of death from CVD (HR=2.12, CI 1.65, 2.71) and from cancer (HR=1.40, CI, 1.15, 1.69) but not accidental death (HR=1.39, CI 0.69, 2.76). Compared with women whose scores did not change over the initial three years of follow-up, SRH that worsened significantly was associated with higher risk of all-cause (HR=2.06), CVD (HR=1.71) and cancer (HR=2.22) mortality; whereas, women with improved SRH had significantly lower all-cause, CVD and cancer mortality risks (HR: 0.78, 0.80, and 0.79, respectively),
Conclusions:Low SRH and a decrease in SRH over three years were strongly associated with increased risks of all-cause, CVD, cancer and other cause mortality after more than 7 years of follow-up in post-menopausal women. Lower SRH was also associated with incident CVD and cancer.SRH is a commonly used measure, so it is important to have a thorough understanding of its behavior, its biases, and what exactly it measures. In the WHI cohort studies that included SRH showed that participants reporting fair or poor health were nearly 12 times as likely to meet frailty criteria as those reporting excellent health [5].In the present analysis of WHI OS participants, SRH is examined as a J o u rn al of G e r o n to lo gy & G e r ia tr ic Resea rc h JGGR, an open access journal 6 Adjusted for age (linear) and race/ethnicity. Baseline hazard functions were allowed to vary by 5-year age groups. 7 Adjusted for age (linear), race/ethnicity, BMI (quintiles and linear), education, marital status, smoking status, alcohol consumption HT use and depressive symptoms. Baseline hazard functions were allowed to vary by 5-year age groups, number of chronic diseases, disability, current health care provider, mammogram within 2 years of enrollment and physical functioning (quintiles). 8 From a multivariable Cox proportional hazards models JGGR, an open access journal 12 Adjusted for age (linear), race/ethnicity, BMI (quintiles and linear), education, marital status, smoking status, alcohol consumption, HT use and depressive symptoms. Baseline hazard functions were allowed to vary by 5-year age groups, number of chronic diseases, disability, and current health care provider, mammogram within 2 years of enrollment, physical functioning (quintiles), and prior history of dis...