In caseybreen/childhoodmortality: Calculating Childhood Mortality Rates

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Replicating DHS Childhood Mortality Estimates

This vignette will provide step-by-step instructions to replicate the estimates provided by the DHS Final Reports. In this Vignette, we'll replicate the estimates reported in the Nigeria 2013 Final Report.

Importing Data

Data can be obtained from IPUMS-DHS or directly from the DHS Program.

Import data from IPUMS-DHS

If data is obtained from IPUMS-DHS:

• Select "Births" unit of analysis.
• Select following variables in the extract: KIDDOBCMC, KIDAGEDIEDIMP, INTDATECMC, PSU, PERWEIGHT, WEALTHQ, EDUCLVL, GEONG2013, URBAN
• The easiest way to import IPUMS-DHS data into R is the ipumsr package.
  • Create and download a .csv or .dat extract from IPUMS-DHS
  • On the extract page right click and use the "save link as" option to save the DDI to the same directory as the data file.

library(ipumsr)
ng2013 <- read_ipums_micro(data_file = "idhs_00019.csv", ddi = "idhs_00019.xml")

Import data from DHS Program

If data is obtained from the DHS Program:

• Download a DHS BR recode Stata file.
• Install the Haven Package

library(haven)
ng2013 <- read_dta("NGBR6AFL.DTA")

Replicating Tables

This table displays estimates of neonatal, post-neonatal, infant, child, under-five mortality rates and standard erros.

To replicate these tables, we'll just have to call the childhoodmortality function. The package defaults to computing mortality rates and standard errors for all rate types:

library(childhoodmortality)
childhoodmortality(ng2013)

|sample |rate_type | mortality_rate| SE| lower_confidence_interval| upper_confidence_interval| |:-------------------|:------------|--------------:|--------:|-------------------------:|-------------------------:| |[5665] Nigeria 2013 |neonatal | 37.28012| 1.483270| 34.31358| 40.24666| |[5665] Nigeria 2013 |postneonatal | 32.48209| 1.472823| 29.53644| 35.42774| |[5665] Nigeria 2013 |infant | 68.55127| 2.133319| 64.28463| 72.81791| |[5665] Nigeria 2013 |child | 63.88013| 2.676241| 58.52765| 69.23261| |[5665] Nigeria 2013 |underfive | 128.05234| 3.688539| 120.67526| 135.42942|

The next table displays mortality rates for several subpopulations.

To calculate subnational estimates, we need a variable to "disaggregate" over - a categorical variable in data which the mortality rates will be disaggregated (e.g. IPUMS-DHS integrated geography variables, wealth quintile, education variables, etc.)

Two important notes:

• This table uses the 10-year period preceding the survey by demographic characteristics. We will have to adjust the "period" argument of the function.
• Postneonatal is computed as the difference between infant and neonatal mortality rates.

library(childhoodmortality)
rates_by_residence <- childhoodmortality(ng2013, grouping = "urban", period = 10)

rates_by_residence %>% 
  select(urban
         , rate_type, mortality_rate) %>%
  spread(rate_type, mortality_rate) %>%
  mutate(postneonatal = infant - neonatal) %>%
  mutate_if(is.numeric, funs(round(., 0)))

|urban | child| infant| neonatal| postneonatal| underfive| |:---------|-----:|------:|--------:|------------:|---------:| |[1] Urban | 42| 60| 34| 26| 100| |[2] Rural | 89| 86| 44| 42| 167|

library(childhoodmortality)
rates_by_region <- childhoodmortality(ng2013, grouping = "geo_ng2013", period = 10)

rates_by_region %>% 
  select(geo_ng2013, rate_type, mortality_rate) %>%
  spread(rate_type, mortality_rate) %>%
  mutate(postneonatal = infant - neonatal) %>%
  mutate_if(is.numeric, funs(round(., 0)))

|geo_ng2013 | child| infant| neonatal| postneonatal| underfive| |:-----------------|-----:|------:|--------:|------------:|---------:| |[1] North Central | 36| 66| 35| 31| 100| |[2] North East | 90| 77| 43| 33| 160| |[3] North West | 105| 89| 44| 46| 185| |[4] South East | 54| 82| 37| 45| 131| |[5] South South | 35| 58| 32| 26| 91| |[6] South West | 31| 61| 39| 21| 90|

library(childhoodmortality)
rates_by_education <- childhoodmortality(ng2013, grouping = "educlvl", period = 10)

rates_by_education %>% 
  select(educlvl, rate_type, mortality_rate) %>%
  spread(rate_type, mortality_rate) %>%
  mutate(postneonatal = infant - neonatal) %>%
  mutate_if(is.numeric, funs(round(., 0)))

|educlvl | child| infant| neonatal| postneonatal| underfive| |:----------------|-----:|------:|--------:|------------:|---------:| |[0] No education | 100| 89| 44| 47| 180| |[1] Primary | 57| 74| 42| 34| 128| |[2] Secondary | 35| 58| 34| 25| 91| |[3] Higher | 13| 50| 30| 21| 62|

library(childhoodmortality)
rates_by_wealth <- childhoodmortality(ng2013, grouping = "educlvl", period = 10)

rates_by_wealth %>% 
  select(wealthq, rate_type, mortality_rate) %>%
  spread(rate_type, mortality_rate) %>%
  mutate(postneonatal = infant - neonatal) %>%
  mutate_if(is.numeric, funs(round(., 0)))

|wealthq | child| infant| neonatal| postneonatal| underfive| |:-----------|-----:|------:|--------:|------------:|---------:| |[1] Poorest | 108| 92| 45| 47| 190| |[2] Poorer | 103| 94| 45| 49| 187| |[3] Middle | 61| 71| 39| 31| 127| |[4] Richer | 38| 65| 37| 28| 100| |[5] Richest | 26| 48| 30| 18| 73|

caseybreen/childhoodmortality documentation built on June 8, 2020, 7:03 p.m.