Background: Despite the widespread interest in competing causes of death, empirical information on interrelationships between causes of death is scarce. We have used death certificate information to estimate the prevalence of competing causes of death at the moment of dying from specific underlying causes of death.
Materials and methods: In a stratified sample of 5975 deaths occurring in The Netherlands in 1990, information contained in the death certificate was used to determine the presence of diseases which, in the hypothetical case of elimination of the underlying cause of death, could develop into a new underlying cause of death. Poisson regression analysis was used to describe variation in age- and sex-adjusted prevalence of competing causes of death between different underlying causes.
Results: Per 100 deaths, 46.2 competing causes were identified (52.0 after reweighting to take away the effects of stratification). The most frequent competing causes, all occurring in more than 2% of deaths, were: senile dementia, diabetes mellitus, ischemic heart disease, cerebrovascular disease, chronic obstructive lung disease, hypertensive disease, and arteriosclerosis. The overall prevalence of competing causes is relatively high among deaths from respiratory diseases (relative risk for respiratory diseases as compared with all underlying causes (RR) = 1.42 (95% CI, 1.25-1.62)), relatively low among deaths from neoplasms (RR = 0.54 (95% CI, 0.47-0.62)), and in between among deaths from cardiovascular diseases (RR = 1.08 (95% CI, 0.95-1.22)).
Conclusion: Although it cannot be excluded that some of the variation in prevalence of competing causes by underlying cause is due to selective underregistration of coexisting diseases on death certificates, the results of this study suggest that conventional estimates of gains in life expectancy after elimination of neoplasms are much less biased by the effect of competing causes than the corresponding estimates for cardiovascular diseases and particularly respiratory diseases.