Boston University
Predictors of childhood stunting among children aged 2-11 months across districts in Zambia
Abstract
dc:description.abstractBackground: Childhood stunting is a pressing public health issue, predominantly in low- and middle-income countries, where it significantly impacts children’s primary health outcomes. Despite extensive research, a predictive model could guide possible interventions. Objectives: The primary goal of this study is to develop predictive models for prevalent and incident childhood stunting in Zambia. By evaluating baseline child, caregiver, and household characteristics and follow-up dietary variables. We aim to provide a comprehensive understanding of the factors contributing to this critical public health issue. Methods: The study utilized data from a cluster randomized trial conducted in Zambia. Initially, data were collected from 2291 caregiver-child pairs in Choma, Mansa, and Lusaka districts. At baseline to assess prevalent stunting, 2251 subjects were included, excluding participants with missing height measurements and participants with extreme length-age-z-score(<-6 or >6) as per WHO advice. Furthermore, incident stunting was evaluated by excluding participants receiving supplementary SQ-LNS, participants who did not attend follow-up, participants with extreme length-age-z-score at follow-up, and participants with prevalent stunting at baseline. Stunting was defined as length-age-z-score <-2. Logistic regression models were used to evaluate potential predictors by estimating the odds ratios (OR) and 95% confidence intervals (CI). Additionally, Akaike Information Criterion (AIC) to identify the best fitting multivariable model. Results: At baseline, nearly 20% of children were stunted. Key predictors evaluated in final predictive model at baseline included child factors: Low birth weight (OR 2.01, 95% CI: 1.42–2.85), HIV exposure (OR 1.53, 95% CI: 1.08–2.17), Weight for age z-score (OR 0.31, 95% CI: 0.27–2.85); Caregiver factors: Marital status (OR 1.24, 95% CI: 0.95–1.62); Household factors: Toilet facilities (Improved pit latrine OR 1.33, 95% CI: 0.92–1.93), (Basic pit OR 1.37, 95% CI: 0.94–1.99). Compared to model 1 (child age and sex), the final predictive model for prevalent stunting yielded Δ AIC=-501.01 and AUC=0.813. However, final predictive model at follow-up included different factors; child factors: Weight for age z-score (OR 0.38, 95% CI: 0.31–0.47); Caregiver factors: Education level (with primary education achievement OR 1.54, 95% CI: 1.04–2.27); Households factors: cooking fuel (OR 1.88 95% CI: 0.99-3.55); Nutrition: non-consumption of sugary drinks (OR 1.69, 95% CI: 1.23-2.34) and non-consumption of sweets (OR 1.17, 95% CI: 0.85.23-1.61) with Δ AIC=-140.10 when compared to model 1. Conclusions: this study identified child sex and weight-for-age z-scores at baseline and follow-up as significant potential predictors. Additionally, HIV exposure and low birth weight were found to be critical predictors for prevalent stunting. While caregiver education, cooking fuel type and non-consumption of sugary drinks were statistically significant at follow-up.
Degree
thesis:*- Name thesis:degree_name
- Master of Science
- Level thesis:degree_level
- masters
- Discipline thesis:degree_discipline
- Nutrition and Metabolism
- Grantor
- Boston University
- Year dc:date.issued
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Al-hamad, Khalifa Naser
- Advisors dc:contributor.advisor
-
- Locks, Lindsey
- Moore, Lynn L.
Subjects
dc:subject × 1Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/2144/53487