The analysis of mortality by the subject-years method

Area of Science:

• Epidemiology
• Biostatistics

Background:

• Comparing observed mortality to national expected rates is crucial for population health studies.
• The subject-years method is a standard approach for calculating expected deaths, accounting for follow-up duration.

Purpose of the Study:

• To describe the properties of the subject-years method for mortality analysis.
• To present a likelihood-based approach for calculating expected deaths.
• To generalize mortality analysis to complex group structures and individual covariates.

Main Methods:

• Utilized the subject-years method for calculating expected deaths.
• Applied a likelihood-based statistical framework.
• Treated observed deaths as Poisson variables for analysis.
• Employed generalized linear models (GLIM/GENSTAT) for calculations.

Main Results:

• The subject-years method's properties were elucidated.
• A likelihood approach was developed, offering a generalized framework.
• The method accommodates factorial structures and individual covariates.

Conclusions:

• The likelihood approach provides a flexible and generalizable method for comparative mortality analysis.
• Statistical software facilitates the implementation of these advanced methods.
• This enhances the accuracy of epidemiological studies comparing observed and expected mortality rates.