Book contents
- Frontmatter
- Contents
- Overview
- 1 Rates and their properties
- 2 Life tables
- 3 Two especially useful estimation tools
- 4 Product-limit estimation
- 5 Exponential survival time probability distribution
- 6 Weibull survival time probability distribution
- 7 Analysis of two-sample survival data
- 8 General hazards model: parametric
- 9 General hazards model: nonparametric
- Examples of R
- Data
- Problem set
- References
- Index
2 - Life tables
Published online by Cambridge University Press: 03 February 2010
- Frontmatter
- Contents
- Overview
- 1 Rates and their properties
- 2 Life tables
- 3 Two especially useful estimation tools
- 4 Product-limit estimation
- 5 Exponential survival time probability distribution
- 6 Weibull survival time probability distribution
- 7 Analysis of two-sample survival data
- 8 General hazards model: parametric
- 9 General hazards model: nonparametric
- Examples of R
- Data
- Problem set
- References
- Index
Summary
The life table is perhaps the earliest statistical tool used to study human mortality rigorously. Early scientists Edmund Halley (1693) and John Graunt (1662) independently developed the first life tables from populations in Poland and England, respectively. A life table is essentially a highly organized description of age-specific mortality rates. Its importance has been reduced by modern methods (to be discussed), but it nevertheless remains fundamental to understanding survival data. A life table illustrates several basic statistical issues, particularly the roles of the two principal summaries of survival data, the survival function and the hazard rate.
Two kinds of life tables exist, called cohort and current life tables. Each kind has two styles, abridged and complete. A cohort life table is constructed from data accumulated by recording survival times from the birth of the first member of a population until the death of the last member. Collecting such cohort data is clearly impracticable in human populations. Cohort life tables primarily describe mortality patterns of small animal and insect populations. An abridged life table is based on a sequence of age intervals of any chosen length, typically five years. A current and complete life table is the subject of the following description. The life table components are derived from present-day observed mortality data (current) and applied to one-year age intervals (complete).
A life table describes the mortality experience of a cohort that does not exist. However, this theoretical cohort frequently provides valuable summaries of mortality patterns useful for comparing similarly constructed life table summaries from other groups or populations.
- Type
- Chapter
- Information
- Survival Analysis for Epidemiologic and Medical Research , pp. 27 - 52Publisher: Cambridge University PressPrint publication year: 2008