The above lists show variables significant in Cox regression models after all variables in the previous slide were tested in stages. Analyses accounted for the complex sample design.

Nine factors were associated with increased hazard of mortality for women, but self-ratings of health were not included. The factors that predict mortality for women were primarily obtained by physical examination, not self-report.

SRH is associated with greater mortality hazard for men. Men with "excellent" health had a hazard of mortality that was 48% less than those who reported "poor" health. Men with "very good" health had a 44% reduced risk, and men with "good" health had a 32% reduced risk compared with "poor". Other risk factors included both measured and self-reported variables.