A linear models application of competing risks to multiple causes of death

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• Investigating the joint occurrence of acute myocardial infarct and stroke is crucial for understanding complex mortality patterns.
• Previous models of competing risks may not fully capture scenarios with multiple causes of death.

Purpose of the Study:

• To analyze the joint incidence of acute myocardial infarct and stroke in Massachusetts and North Carolina in 1969.
• To extend competing risks theory to model multiple causes of death.
• To apply categorical data analysis to survival parameters.

Main Methods:

• Development of a biological model to assess the association between acute myocardial infarct and stroke.
• Extension of Chiang's (1968) theory of competing risks for multiple causes of death.
• Application of Grizzle, Starmer, and Koch's (1969) categorical data procedures.

Main Results:

• Higher incidence of joint acute myocardial infarct and stroke on death certificates in North Carolina compared to Massachusetts.
• A clear age gradient observed in the joint occurrence of these two diseases.
• Differential patterns of age variation between males and females, with more prominent state-by-age interactions in females.

Conclusions:

• The findings are consistent with higher stroke mortality rates in North Carolina.
• The developed methodology allows for the modeling of survival parameters in the presence of multiple causes of death.
• Age and sex are significant factors influencing the joint occurrence of acute myocardial infarct and stroke.