Checking if IPC Acute Malnutrition sample size requirements were met

Tomás Zaba

library(mwana)

Evidence on the prevalence of acute malnutrition used in the IPC Acute Malnutrition (IPC AMN) can come from different sources, namely: representative surveys, screenings or community-based surveillance system (known as sentinel sites). The IPC set minimum sample size requirements for each of these sources. Details can be read from the IPC Manual version 3.1.

In the IPC AMN analysis workflow, the very first step of a data analyst is to check if these requirements were met. This is done for each area meant to be included in the IPC AMN analysis. For this, mwana provides a handy function: mw_check_ipcamn_ssreq().

To demonstrate its usage, we will use a built-in sample dataset anthro.01.

head(anthro.01)

# A tibble: 6 × 11
  area       dos        cluster  team sex   dob      age weight height edema
  <chr>      <date>       <int> <int> <chr> <date> <int>  <dbl>  <dbl> <chr>
1 District E 2023-12-04       1     3 m     NA        59   15.6  109.  n
2 District E 2023-12-04       1     3 m     NA         8    7.5   68.6 n
3 District E 2023-12-04       1     3 m     NA        19    9.7   79.5 n
4 District E 2023-12-04       1     3 f     NA        49   14.3  100.  n
5 District E 2023-12-04       1     3 f     NA        32   12.4   92.1 n
6 District E 2023-12-04       1     3 f     NA        17    9.3   77.8 n
# ℹ 1 more variable: muac <int>

anthro.01 contains anthropometry data from SMART surveys from anonymized locations. We can check further details about the dataset by calling help(anthro.01) in R console.

Now that we are acquainted with the dataset, we can proceed to execute the task. To achieve this, we simply do:

mw_check_ipcamn_ssreq(
  df = anthro.01,
  cluster = cluster,
  .source = "survey"
)

Or we can also choose to chain the data object to the function using the pipe operator:

anthro.01 |>
  mw_check_ipcamn_ssreq(
    cluster = cluster,
    .source = "survey"
  )

Either way, the returned output will be:

# A tibble: 1 × 3
  n_clusters n_obs meet_ipc
       <int> <int> <chr>
1         30  1191 yes     

A table (of class tibble) is returned with three columns:

• Column n_clusters counts the number of unique cluster IDs in the dataset.
  
• Column n_obs counts the number of children in the dataset.
  
• Column meet_ipc indicates whether the IPC AMN sample size requirements (for surveys in this case) were met or not.

The above output is not quite useful yet as we often deal with multiple area dataset. We can get a summarised table by area as follows:

## Load the dplyr package ----
library(dplyr)

## Use the group_by() function ----
anthro.01 |>
  group_by(area) |>
  mw_check_ipcamn_ssreq(
    cluster = cluster,
    .source = "survey"
  )

This will return:

# A tibble: 2 × 4
  area       n_clusters n_obs meet_ipc
  <chr>           <int> <int> <chr>
1 District E         30   505 yes
2 District G         30   686 yes     

For screening or sentinel site-based data, we approach the task the same way; we only have to change the .source parameter to “screening” or to “ssite” as appropriate, as well as to supply cluster with the right column name of the sub-areas inside the main area (villages, localities, comunas, communities, etc).