PHEindicatormethods (version 1.1.5)

phe_life_expectancy: Calculate Life Expectancy using phe_life_expectancy

Description

Compute life expectancy for a given age, and its standard error

Usage

phe_life_expectancy(data, deaths, population, startage,
  age_contents = c(0L, 1L, 5L, 10L, 15L, 20L, 25L, 30L, 35L, 40L, 45L,
  50L, 55L, 60L, 65L, 70L, 75L, 80L, 85L, 90L), le_age = "all",
  type = "full", confidence = 0.95)

Arguments

data

data.frame or tbl containing the deaths and population data

deaths

field name from data containing the number of deaths within age band; unquoted string; no default

population

field name from data containing the population within age band; unquoted string; no default

startage

field name from data containing the age band; no default

age_contents

vector; describes the contents of startage in the ascending order. This vector is used to check whether each group in data contains the complete set of age bands for the calculation to occur. It is also used to reorder the data based on the startage field

le_age

the age band to return the life expectancy for. The default is "all", where the function returns the life expectancy values for all ages appended onto the input table. Any other value (or vector of values) must be age bands described by the age_contents input

type

type of output; can be "standard" or "full" (full contains added details on the calculation within the dataframe); quoted string; default full

confidence

the required level of confidence expressed as a number between 0.9 and 1 or 90 and 100; numeric; default 0.95

Details

This function aligns with the methodology in Public Health England's Life Expectancy Excel Tool.

The function is for an abridged life table using 5 year age intervals with a final age interval of 90+. The table has been completed using the methods described by Chiang.[1],[2] This age structure and methodology is used by The Office for National Statistics to produce life expectancy at national and local authority level.[3]

This function includes an adjustment to the method for calculating the variance of the life expectancy estimate to include a term for the variance associated with the final age interval. In the Chiang method the variance of the life expectancy is the weighted sum of the variance of the probability of survival across all the age intervals. For the final age interval the probability of survival is, Chiang argues, zero and has zero variance. However, Silcocks et al argue[4] that in the case of the final age interval the life expectancy is dependent not on the probability of survival but on the mean length of survival \((1/M_omega)\). Therefore the variance associated with the final age interval depends on the age-specific mortality rate \(M_omega\).

Life expectancy cannot be calculated if the person-years in any given age interval is zero. It will also not be calculated if the total person-years is less than 5,000 as this is considered to be the minimum size for robust calculation of life expectancy.[5] Zero death counts are not a problem, except for the final age interval - there must be at least one death in the 90+ interval for the calculations to be possible.

The methodology used in this function, along with discussion of alternative options for life expectancy calculation for small areas, were described Eayres and Williams.[6]

References

[1] Chiang CL. The Life Table and its Construction. In: Introduction to Stochastic Processes in Biostatistics. New York, John Wiley & Sons, 1968:189-214. [2] Newell C. Methods and Models in Demography. Chichester, John Wiley & Sons, 1994:63-81 [3] Office for National Statistics Report. Life expectancy at birth by health and local authorities in the United Kingdom, 1998 to 2000 (3-year aggregate figures.) Health Statistics Quarterly 2002;13:83-90 [4] Silcocks PBS, Jenner DA, Reza R. Life expectancy as a summary of mortality in a population: statistical considerations and suitability for use by health authorities. J Epidemiol Community Health 2001;55:38-43 [5] Toson B, Baker A. Life expectancy at birth: methodological options for small populations. National Statistics Methodological Series No 33. HMSO 2003. [6] Eayres DP, Williams ES. Evaluation of methodologies for small area life expectancy estimation. J Epidemiol Community Health 2004;58:243-249

See Also

Other PHEindicatormethods package functions: phe_dsr, phe_isr, phe_mean, phe_proportion, phe_quantile, phe_rate, phe_sii, phe_smr

Examples

# NOT RUN {
library(dplyr)

## A simple example
df <- data.frame(startage = c(0L, 1L, 5L, 10L, 15L, 20L, 25L, 30L, 35L, 40L, 45L, 50L, 55L,
                              60L, 65L, 70L, 75L, 80L, 85L, 90L),
                 pops = c(7060L, 35059L, 46974L, 48489L, 43219L, 38561L, 46009L, 57208L,
                          61435L, 55601L, 50209L, 56416L, 46411L, 39820L, 37978L,
                          37039L, 33288L, 23306L, 11936L, 11936L),
                 deaths = c(17L, 9L, 4L, 8L, 20L, 15L, 24L, 33L, 50L, 71L, 100L, 163L,
                            263L, 304L, 536L, 872L, 1390L, 1605L, 1936L, 1937L))
phe_life_expectancy(df, deaths, pops, startage)

## OR

phe_life_expectancy(df, deaths, pops, startage, le_age = c(5, 25), type = "standard")

## Unordered age bands example
df <- data.frame(startage = c("0", "1-4", "5-9", "10 - 14", "15 - 19", "20 - 24", "25 - 29",
                              "30 - 34", "35 - 39", "40 - 44", "45 - 49", "50 - 54",
                              "55 - 59", "60 - 64", "65 - 69", "75 - 79", "80 - 84",
                              "85 - 89", "90 +", "70 - 74"),
                 pops = c(7060L, 35059L, 46974L, 48489L, 43219L, 38561L, 46009L, 57208L,
                          61435L, 55601L, 50209L, 56416L, 46411L, 39820L, 37039L,
                          23306L, 11936L, 11936L, 37978L, 33288L),
                 deaths = c(17L, 9L, 4L, 8L, 20L, 15L, 24L, 33L, 50L, 71L, 100L, 163L,
                            263L, 304L, 872L, 1605L, 1936L, 1937L, 536L, 1390L))
phe_life_expectancy(df, deaths, pops, startage,
                    age_contents = c("0", "1-4", "5-9",
                                     "10 - 14", "15 - 19",
                                     "20 - 24", "25 - 29",
                                     "30 - 34", "35 - 39",
                                     "40 - 44", "45 - 49",
                                     "50 - 54", "55 - 59",
                                     "60 - 64", "65 - 69",
                                     "70 - 74", "75 - 79",
                                     "80 - 84", "85 - 89",
                                     "90 +"))

df <- data.frame(area = c(rep("Area 1", 20), rep("Area 2", 20)),
                 startage = rep(c(0L, 1L, 5L, 10L, 15L, 20L, 25L, 30L, 35L, 40L, 45L, 50L, 55L,
                                  60L, 65L, 70L, 75L, 80L, 85L, 90L), 2),
                 pops = rep(c(7060L, 35059L, 46974L, 48489L, 43219L, 38561L, 46009L, 57208L,
                              61435L, 55601L, 50209L, 56416L, 46411L, 39820L, 37978L,
                              37039L, 33288L, 23306L, 11936L, 11936L), 2),
                 deaths = rep(c(17L, 9L, 4L, 8L, 20L, 15L, 24L, 33L, 50L, 71L, 100L, 163L,
                                263L, 304L, 536L, 872L, 1390L, 1605L, 1936L, 1937L), 2))
df %>%
       group_by(area) %>%
       phe_life_expectancy(deaths, pops, startage)

# }