Self-rated health and physician-rated health as independent predictors of mortality in elderly men

Background: When assessing health status, physicians may focus on objective symptoms and diagnoses, whereas individuals may focus more on subjective symptoms, functional limitations and quality of life. Methods: In the Zutphen Elderly Study, 710 community-living men (aged 64-84 years) were followed until death for 15 years. Self-rated health was assessed through a single-item question. Physician-rated health was estimated on a Likert scale by physicians after medical history assessment and physical examination. Both health ratings were categorised into three groups. All-cause, cardiovascular and cancer mortality rates were analysed in Cox proportional-hazards models. Results: There were 352 (49.6%) men who felt healthy and 225 (31.7%) men with a good physician-rated health. During 15 years of follow-up 503 of 710 men (70.8%) died, of whom 229 (45.5%) from cardiovascular causes and 144 (28.6%) from cancer. Self-rated and physician-rated health both predicted independently all-cause mortality (hazard ratios [HR] for worst vs. best health category: 1.72; 95% confidence interval [CI]: 1.26-2.33, and 1.77; 95% CI: 1.36-2.29; respectively; P-values of < 0.005). When self-rated and physician-rated health were discordant, mortality risk was highest when physicians had a less favourable view on the health status than the participant. Self-rated health predicted independently cancer mortality (HR 2.41), whereas physician-rated health cardiovascular mortality (HR 2.13). Conclusion: Self-rated and physician-rated health status predicted both all-cause mortality, and showed a differential pattern for cancer and cardiovascular diseases mortality.