Do proxy evaluations of health status predict mortality?

The ability of self-evaluations of health to predict mortality remains despite efforts to explain their contribution. To further the understanding of global evaluations, this study investigated whether proxy evaluations of health status also predict mortality. Data from the Longitudinal Study of Aging were used. For men and women, the unadjusted odds ratios for proxy evaluations exhibited a monotone relationship to 3-year and 7-year mortality rates and were predictive of 7-year mortality. This monotone relationship remained even after controlling for other factors when predicting 7-year mortality rates. However, only the fair and poor evaluations were significant, and only for men. The adjusted relationship when predicting 3-year mortality was not completely monotone, although good and poor proxy evaluations were significant. This suggests that proxy evaluations are useful, and that there are aspects of health status that are currently unmeasured but that should lend themselves to being quantified.