Competing causes of death: an analysis using multiple-cause-of-death data from The Netherlands

Area of Science:

• Demography
• Epidemiology
• Public Health

Background:

• Cause-elimination life tables commonly assume statistical independence between causes of death.
• This assumption implies that eliminating one cause does not alter the mortality risk from other causes.

Purpose of the Study:

• To investigate the validity of the statistical independence assumption in cause-elimination life tables.
• To analyze the prevalence of co-occurring conditions at death in the Netherlands using multiple-cause-of-death data.

Main Methods:

• Utilized 1990 multiple-cause-of-death data from the Netherlands.
• Calculated age-standardized prevalence of other conditions at death for four underlying cause groups: malignant neoplasms, cardiovascular diseases, respiratory diseases, and external causes.
• Performed two series of calculations: one including all co-occurring conditions, and another with conditions eligible to become a new underlying cause.

Main Results:

• Significant variations exist in the prevalence of co-occurring conditions across different underlying causes of death.
• Cardiovascular disease deaths showed a high prevalence of potential new underlying causes, while malignant neoplasms and external causes showed a low prevalence.
• Respiratory disease deaths had an average prevalence of potential new underlying causes.

Conclusions:

• The assumption of statistical independence in conventional cause-elimination life tables is likely invalid.
• Multiple-cause-of-death data is a valuable resource for refining mortality analyses.
• Further validation studies of multiple-cause-of-death data are warranted.