Child anthropometric assessments are the cornerstones of child nutrition and food security surveillance around the world. Ensuring the quality of data from these assessments is paramount to obtaining accurate child under nutrition prevalence estimates. Additionally, the timeliness of reporting is, as well, critical to allowing timely situation analyses and responses to tackle the needs of the affected population.

mwana, term for child in Elómwè, a local language spoken in the central-northern regions of Mozambique, with a similar meaning across other Bantu languages, such as Swahili, spoken in many parts of Africa, is a package that streamlines data quality checks and wasting prevalence estimation from anthropometric data of children aged 6 to 59 months old through a comprehensive implementation of the SMART Methodology guidelines in R.

Motivation

mwana was borne out of the author’s own experience of having to work with multiple child anthropometric data sets to conduct data quality appraisal and prevalence estimation as part of the analysis Quality Assurance Team of the Integrated Phase Classification (IPC) Global Support Unit. The current standard child anthropometric data appraisal workflow is extremely cumbersome, requiring significant time and effort utilizing different software tools - SPSS, Excel, Emergency Nutrition Assessment or ENA software - for each step of the process for a single data set. This process is repeated for every data set needing to be processed and often needing to be implemented in a relatively short period of time. This manual and repetitive process, by its nature, is extremely error-prone.

mwana simplifies this cumbersome workflow into a programmable process particularly when handling multiple-area data set.

[!NOTE]

mwana was made possible thanks to the state-of-the-art work in nutrition survey guidance led by the SMART initiative. Under the hood, mwana bundles the SMART Methodology guidance, for both survey and non survey data, through the use of the National Information Platforms for Nutrition Anthropometric Data Toolkit (nipnTK) functionalities in R to build its handy function around plausibility checks and wasting prevalence estimation. Click here to learn more about the {nipnTK} package.

What does `mwana` do?

It automates plausibility checks, prevalence analyses, and summary outputs, providing particular advantages when handling data sets with multiple areas.

Plausibility checks.

mwana performs plausibility checks on weight-for-height z-score (WFHZ) data by mimicking the SMART plausibility checkers in ENA for SMART software, their scoring and classification criterion. Read guide here.
It performs, as well, plausibility checks on MUAC data. For this, mwana integrates recent advances in using muac-for-age z-score (MFAZ) for checking the plausibility and the acceptability of MUAC data. In this way, when the variable age is available: mwana performs plausibility checks similar to those in WFHZ, with a few differences in the scoring criteria for the percent of flagged data. Otherwise, when the variables age is missing, a similar test suit used in the current version of ENA is performed. Read guide here.

A tidy workflow for plausibility check using `mwana`

Prevalence estimation

mwana prevalence estimators were built to take decisions on the appropriate analysis procedure to follow based on the quality of the data, as per the SMART rules. They return output tables with summarized results based on the data quality test results. Fundamentally, the functions loop over the survey areas in the data set whilst doing quality checks and taking decisions on the appropriate prevalence analysis path that best fits the data.

mwana estimates wasting prevalence on the basis of:

WFHZ and/or edema. Read the guide here
Raw MUAC values and/or edema. When variable age is available, detection and removal of outliers is based on MFAZ, otherwise based on the raw MUAC values. This is simply to exclude outliers; the actual prevalence estimation is based on the raw MUAC values. Read the guide here.
MFAZ and/or edema. Read the guide here.
Combined prevalence. A concept of combined flags is used to streamline the flags removed in WFHZ and those in MUAC. Read the guide here.

In the context of IPC Acute Malnutrition (IPC AMN) analysis workflow, mwana provides a handy function for checking whether the minimum sample size requirements of a given area were met, on the basis of the methodology used to collect the data, be it a survey, a screening or a sentinel site data. Read the guide here.

[!TIP]

If you are undertaking a research and you want to wrangle your data before using it in your statistical models, mwana is a great helper.

[!WARNING]

Please note that mwana is still highly experimental and is undergoing a lot of development. Hence, any functionalities described above have a high likelihood of changing interface or approach as we aim for a stable working version.

Installation

mwana is not yet on CRAN but can be installed from the nutriverse R Universe as follows:

install.packages(
  "mwana",
  repos = c('https://nutriverse.r-universe.dev', 'https://cloud.r-project.org')
)

Then load to in memory with

library(mwana)

Citation

If you were enticed to use mwana package and found it useful, please cite using the suggested citation provided by a call to citation function as follows:

citation("mwana")
#> To cite mwana: in publications use:
#> 
#>   Tomás Zaba, Ernest Guevarra (2024). _mwana: An Efficient Workflow for
#>   Plausibility Checks and Prevalence Analysis of Wasting in R_. R
#>   package version 0.2.0, <https://github.com/nutriverse/mwana>.
#> 
#> A BibTeX entry for LaTeX users is
#> 
#>   @Manual{,
#>     title = {mwana: An Efficient Workflow for Plausibility Checks and Prevalence Analysis of Wasting in R},
#>     author = {{Tomás Zaba} and {Ernest Guevarra}},
#>     year = {2024},
#>     note = {R package version 0.2.0},
#>     url = {https://github.com/nutriverse/mwana},
#>   }

Community guidelines

Feedback, bug reports and feature requests are welcome; file issues or seek support here. If you would like to contribute to the package, please see our contributing guidelines.

This project is releases with Contributor Code of Conduct. By participating in this project you agree to abide by its terms.

mwana: An efficient workflow for plausibility checks and prevalence analysis of wasting in R