Skip to contents

Prepare ITN Data for MBG Analysis

Source: R/indicator_itn.R
calc_itn_mbg.Rd

Prepares cluster-level ITN ownership, access, and use data for Model-Based Geostatistics (MBG) analysis. Aggregates to cluster counts WITHOUT survey weights - MBG handles spatial smoothing internally.

Usage

calc_itn_mbg(
  dhs_hr,
  dhs_pr,
  gps_data,
  indicators = NULL,
  survey_vars = list(cluster = "hv001", hhid = "hhid", hhsize = "hv013", age = "hv105",
    sex = "hv104", pregnant = "hml18", itn_use = "hml12", itn_prefix = "hml10_",
    itn_treated_prefix = "hml7_"),
  gps_vars = list(cluster = "DHSCLUST", lat = "LATNUM", lon = "LONGNUM"),
  seed = NULL
)

Arguments

dhs_hr

DHS Household Records dataset.

dhs_pr

DHS Person Records dataset.

gps_data

DHS GPS dataset with cluster coordinates.

indicators

Character vector of indicators to calculate. Options from .itn_mbg_dictionary():

  • "with_itn": Households with at least one ITN (HR)

  • "enough_itn": Households with enough ITNs for every 2 people (HR)

  • "access_itn": Population with access to ITN – binary indicator (PR)

  • "use_itn": Population that used ITN last night (PR)

  • "use_itn_chu5": Under-5 children that used ITN (PR)

  • "use_itn_5_10": Children 5-9 years that used ITN (PR)

  • "use_itn_10_20": Adolescents 10-19 years that used ITN (PR)

  • "use_itn_20plus": Adults 20+ that used ITN (PR)

  • "use_itn_preg": Pregnant women that used ITN (PR)

  • "use_itn_if_access": Of those with access, proportion that used ITN (PR)

Default: all indicators.

survey_vars

Named list mapping DHS variable names.

gps_vars

Named list for GPS variable mapping.

seed

Deprecated. Previously used for probabilistic access assignment. Access is now calculated deterministically following standard DHS methodology.

Value

A list of data.tables (one per indicator), each with columns:

  • cluster_id: Cluster identifier

  • indicator: Numerator count

  • samplesize: Denominator count

  • x: Longitude

  • y: Latitude

Details

Uses a dictionary-driven approach matching the indicator codes from calc_itn_dhs. The dictionary mirrors the DHS .itn_conditions() – same outcome variables, same filters, same data sources (HR vs PR).

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

This function prepares data for MBG spatial modeling. Unlike the survey- weighted calc_itn_dhs() function, this uses simple cluster-level counts.

ITN access is calculated using the standard DHS deterministic assignment method:

  1. Calculate potential users per household: min(ITNs * 2, household_size)

  2. Sort individuals within each household by ITN use (users first)

  3. Assign access to the first N individuals where N = potential_users

This method guarantees that use <= access at the individual level, since anyone who used an ITN is prioritised for access assignment.

See also

calc_itn_dhs() for survey-weighted estimates, .itn_mbg_dictionary() for indicator definitions

Examples

if (FALSE) { # \dontrun{
itn_mbg <- calc_itn_mbg(
  dhs_hr = hr_data,
  dhs_pr = pr_data,
  gps_data = gps_data,
  indicators = c("access_itn", "use_itn_chu5")
)
} # }