Skip to contents

Prepare Care-Seeking Behavior Data by Wealth Quintile for MBG Analysis

Source: R/indicator_csb.R
calc_csb_by_wealth_mbg.Rd

Prepares cluster-level care-seeking behavior data stratified by wealth quintile for MBG analysis. Calculates proportions of febrile children who sought care, separately for each wealth quintile.

Usage

calc_csb_by_wealth_mbg(
  dhs_kr,
  gps_data,
  indicators = c("any", "public", "private", "none"),
  quintiles = 1:5,
  wealth_var = "v190",
  csb_priority_method = c("all", "first", "public", "private"),
  survey_vars = list(cluster = "v001", age = "hw1", fever = "h22"),
  gps_vars = list(cluster = "DHSCLUST", lat = "LATNUM", lon = "LONGNUM")
)

Arguments

dhs_kr

DHS Children's Recode (KR) dataset.

gps_data

DHS GPS dataset with cluster coordinates.

indicators

Character vector of indicators to calculate:

  • "any": Sought care anywhere (public or private)

  • "public": Public sector including CHW

  • "pub_nochw": Public sector excluding CHW

  • "chw": Community health worker only

  • "private": Any private sector

  • "priv_formal": Private formal sector only

  • "pharmacy": Pharmacy / drug shop only

  • "priv_informal": Private informal only

  • "priv_form_pha": Private formal or pharmacy

  • "trained": Trained provider (public + formal + pharmacy)

  • "none": Did not seek care

Default: c("any", "public", "private", "none").

quintiles

Numeric vector of wealth quintiles to include. Default: 1:5 (all quintiles). Use c(1) for poorest only, c(1,2) for poorest + poorer, etc.

wealth_var

Name of wealth quintile variable in dhs_kr. Default: "v190".

csb_priority_method

Character, one of "all" (default), "first", "public", or "private". Controls how overlapping care-seeking records are resolved so each individual is assigned to at most one sector. See calc_csb_mbg() for details. With non-"all" values, csb_public + csb_private + csb_none sums to 100% within each quintile.

survey_vars

Named list mapping DHS variable names.

gps_vars

Named list for GPS variable mapping.

Value

A nested list structure: first level keys are quintiles (e.g., "q1", "q2"), second level keys are indicators (e.g., "csb_public_q1"). Each leaf is a data.table with columns:

  • cluster_id: Cluster identifier

  • indicator: Numerator count (children who sought care)

  • samplesize: Denominator count (all febrile children in quintile)

  • x: Longitude

  • y: Latitude

Details

This function extends calc_csb_mbg() to provide wealth-stratified estimates. Each quintile produces separate MBG-ready outputs with cluster-level numerators and denominators.

Important: This is a standalone utility function for specialized wealth stratification analysis. It is NOT called by run_mbg_pipeline(). Use this function directly when you need wealth-disaggregated indicators for custom MBG modeling or equity analysis.

Methodology: https://github.com/ahadi-analytics/sntmethods/blob/master/inst/methods/csb_dhs.yml

Output Structure

The function returns a list where each indicator-quintile combination gets its own data.table. For example, with indicators = c("public", "private") and quintiles = c(1, 5):


list(
  csb_public_q1 = data.table(...),
  csb_public_q5 = data.table(...),
  csb_private_q1 = data.table(...),
  csb_private_q5 = data.table(...)
)

See also

Examples

if (FALSE) { # \dontrun{
# Care-seeking by wealth for poorest quintile only
csb_poorest <- calc_csb_by_wealth_mbg(
  dhs_kr = kr_data,
  gps_data = gps_data,
  indicators = c("public", "private"),
  quintiles = 1
)

# Compare poorest vs richest quintiles
csb_comparison <- calc_csb_by_wealth_mbg(
  dhs_kr = kr_data,
  gps_data = gps_data,
  indicators = c("any", "public", "none"),
  quintiles = c(1, 5)
)
} # }