Skip to contents

Prepare PfPR Data for MBG Analysis

Source: R/indicator_pfpr.R
calc_pfpr_mbg.Rd

Prepares cluster-level malaria parasite prevalence data for Model-Based Geostatistics (MBG) analysis. Uses a dictionary-driven approach matching the indicator codes from calc_pfpr_dhs.

Usage

calc_pfpr_mbg(
  dhs_pr,
  gps_data,
  indicators = NULL,
  test_type = NULL,
  age_groups = NULL,
  survey_vars = list(cluster = "hv001", age = "hc1", present = "hv103", mother = "hv042",
    rdt = "hml35", mic = "hml32"),
  gps_vars = list(cluster = "DHSCLUST", lat = "LATNUM", lon = "LONGNUM")
)

Arguments

dhs_pr

DHS Person Records dataset (data.frame or tibble).

gps_data

DHS GPS dataset with cluster coordinates.

indicators

Character vector of indicator codes to calculate. Available codes (<test> = rdt or mic):

  • "pfpr_<test>": 6-59 months (standard DHS PfPR)

  • "pfpr_<test>_u5": 0-59 months

  • "pfpr_<test>_5_10": 60-119 months

  • "pfpr_<test>_u10": 0-119 months

  • "pfpr_<test>_2_10": 24-119 months (PfPR2-10 reference)

Default: all indicators from the dictionary.

test_type

Deprecated. Character. Use indicators instead. When provided, translated to indicator codes for backward compatibility. One of "rdt", "mic", "both", or "either".

age_groups

Deprecated. Named list of age ranges. Use indicators instead. When provided alongside test_type, translated to indicator codes.

survey_vars

Named list mapping DHS variable names. Required keys:

  • cluster: Cluster ID (default: "hv001")

  • age: Age in months (default: "hc1")

  • present: Present in household (default: "hv103")

  • mother: Mother listed in household (default: "hv042")

  • rdt: RDT result variable (default: "hml35")

  • mic: Microscopy result variable (default: "hml32")

gps_vars

Named list for GPS variable mapping.

Value

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

  • cluster_id: Cluster identifier

  • indicator: Number of positive tests (numerator for MBG)

  • samplesize: Number of children tested (denominator for MBG)

  • x: Longitude

  • y: Latitude

Details

All dictionary-based indicators share the same data preparation pipeline via .prepare_pfpr_data(), the same shared helper used by the survey-weighted calc_pfpr_dhs() function. Positivity definitions are identical:

  • RDT positive: rdt_res == 1 (hml35 == 1)

  • Microscopy positive: mic_res == 1 (hml32 == 1, Pf only)

  • Either: positive on RDT OR microscopy

Unlike the survey-weighted function, this uses simple cluster-level counts because MBG handles spatial smoothing and uncertainty internally.

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

See also

calc_pfpr_dhs() for survey-weighted estimates

Examples

if (FALSE) { # \dontrun{
# New-style: specify exact indicator codes
pfpr_mbg <- calc_pfpr_mbg(
  dhs_pr = pr_data,
  gps_data = gps_data,
  indicators = c("pfpr_rdt_u5", "pfpr_mic_u5")
)

# Legacy style (still works, with deprecation warning)
pfpr_mbg <- calc_pfpr_mbg(
  dhs_pr = pr_data,
  gps_data = gps_data,
  test_type = "rdt",
  age_groups = list(u5 = c(6, 59))
)
} # }