# A four-year, mixed-family and community-based growth monitoring and promotion program using multi-level modeling to address undernutrition in children in Cambodia

## TL;DR

A family and community-based program in Cambodia improved child undernutrition over four years by using growth monitoring and multi-level modeling.

## Contribution

The study introduces a mixed-family and community-based growth monitoring program using multi-level modeling to address undernutrition in children.

## Key findings

- Older children showed significant improvements in height-for-age scores over time, converging with younger children.
- Multilevel modeling revealed that age at enrollment and time influenced height-for-age changes, while only age at enrollment affected weight-for-age.
- The program reduced significant baseline differences in growth indicators between age groups by the end of the study.

## Abstract

Child undernutrition persists as a formidable public health issue in developing countries. Children afflicted by undernutrition are susceptibility to both physical and neurological repercussions. For several decades, initiatives focused on growth monitoring and promotion have been instituted to mitigate this pressing issue. Nevertheless, the prevalence rates of undernutrition across developing nations continue to provoke concern.

Between the years 2016 and 2019, we executed a family- and community-oriented growth monitoring and promotion initiative within a rural Cambodian village, specifically aimed at children under the age of five. This initiative employed a hybrid workforce comprising both full-time health professionals and community volunteers. Leveraging this robust capacity, we delivered small-group nutrition education sessions, family-centered nutrition counseling, and regular anthropometric assessments. In contrast to a cross-sectional methodology, we used multi-level modeling to explore the growth trajectories of children utilizing longitudinal z scores for height-for-age and weight-for-age. A systematic taxonomy of models was developed in a sequential framework to ascertain the most appropriate final model.

Out of 533 enrolled children, 358 completed the growth monitoring program (GMP). At baseline, children older than 12 months had significantly lower height-for-age (HAZ) and weight-for-age (WAZ) scores, as well as higher rates of stunting and wasting, compared to younger children. These differences were no longer significant by the end of the program. Nonparametric trajectory analyses showed age-related differences in HAZ patterns, with initial declines followed by recovery in younger age groups, while WAZ trajectories remained relatively flat across all ages. Multilevel modeling indicated that both age at enrollment and time significantly influenced HAZ changes, while only age at enrollment affected WAZ. Older children exhibited steeper improvements over time, leading to convergence in growth outcomes with younger children.

An extended and efficacious growth monitoring and promotion program has the potential to ameliorate the issue of undernutrition in developing countries.