how-to-use-healthexpectancies

library(healthexpectancies)

Function CompleteDFLEtable

This function calculates life expectancies and disability-free life expectancies from a dataset containing (at least) information on age, mortality rates and disability prevalence.

As an option, a ‘sex’ and ‘categ’ variables can be added in the input dataset. If so, calculations are made for each separate values of ‘sex’ and ‘categ’. The latter is an undefinit category variables, for instance region or social category. If the initial dataset contains variable ‘year’, calculations are also made for each separate years.

Calculations are made using the Sullivan method. They completely follow example 1 provided in the Sullivan manual (June 2007 version), available at https://reves.site.ined.fr/en/resources/computation_online/sullivan/. One minor deviation is that the ad-hoc value for the conditional probability of death (qx) at age 0 is note taken into account.

The input table is a dataset containing at least the 3 variables above. The output table is a dataset containing all input variables and additionnal variables that are useful for the calculations of LE and DFLE. If those variables are already in the input table, they are not calculated and the input variables are used instead.

Definitions of variables in the input or output dataset:

  • year : Year
  • age : Age
  • Px : Mid-year population at age x
  • Dx : No. deaths
  • mx : central death rate
  • qx : conditional probability of death
  • lx : numbers surviving to age x
  • Lx : person years lived at age x
  • ex : total life expectancy
  • pix : proportion with disability
  • DFLx : person years lived without disability
  • DFTx : total years lived without disability from age x
  • DFLEx : disability-free life expectancy
  • pctDFLEx : proportion of life spent disability-free
  • DLEx : years lived with disability (= ex - DFLEx)
  • DLx : person years lived with disability,
  • DLEx : with-disability life expectancy
  • pctDLEx : proportion of life spent with disability
  • MeanDAx : average conjonctural age of person years lived with disability
  • MedianDAx : median conjonctural age of person years lived with disability
  • ModalDAx : modal conjonctural age of person years lived with disability

Function prevalence_to_polynomial

This function approximates a vector of prevalences (according to age) by a polynomial function.

The degree of the polynomial can be chosen. Default value is 4.

Weights can also be defined (eg. number of survivors at each each). By default, all ages are weighted identically.

Function polynomial_to_prevalence

This function calculates prevalences at each age, given a polynomial function.

Parameters are the following :

  • coefficients : a vector of coefficients for the polynomial.
  • agemin : minimal age in the output vector (default is 60)
  • agemax : maximal age in the output vector (default is 120)
  • ageclass : a vector defining age brackets. If ageclass is non null, the output vector are average prevalences within each agebracket

Function ForecastPrevalence

This function forecasts the values of disability prevalences and of DFLE according to age, given initial datasets containing age, disability prevalences at a reference year, forecasted mortality rates, and on hypothesis on the evolution of disabilities.

Forecasting options are :

  • ‘cstDFLE’ : constant disability-free life expectancy over time (ie. every increase in life expectancy is spent in disability)
  • ‘cstDLE’ : constant in-disability life expectancy (ie. every increase in life expectancy is spent disability-free)
  • ‘cstPctDFLE’ : constant share of disability-free life expectancy in total life expectancy
  • ‘cstPrev’ : constant prevalences of disability (at every age and sex) over time

Datasets included in the package

Several datasets are included in the healthexpectancies package. All of them relates to France.

Insee’s population forecast (2016)

Insee is the French national institute for statistics and economic studies. It publishes population forecasts every 5 years (last publication was in 2016). Forecasted mortality rates by sex for the ‘central’ scenario are included in the package.

See https://www.insee.fr/fr/information/2546485#titre-bloc-3 for more information about the forecasting methods.

VQS (2014)

The Vie quotidienne et santé (VQS) survey is a large sample survey conducted in 2014 by DREES (the statistical directory of the French Ministry for Health and Social affairs).