Application of Ordinal Logistic Regression Model to Nutritional Status of the Under-Five Children Indexed by Weight-for-Height

Anthony Ekpo; Waheed Babatunde  Yahya

doi:10.18502/jbe.v5i3.3620

Anthony Ekpo
Waheed Babatunde Yahya

DOI: https://doi.org/10.18502/jbe.v5i3.3620

Keywords: Under-five; Statistical surveillance; Anthropometrics; Bilateral edema; Odd-ratio; Nutritional status; log-odds

Abstract

Background and aim: In this paper, we present results regarding the outcomes of some anthropometric, epidemiological and demographic factors on the nutritional status of the under-five children which were categorized into three ordinal groups of Severe Acute Malnutrition (SAM), Moderate Acute Malnutrition (MAM) and Global Acute Malnutrition (GAM) in Kazaure Local Government Area in Nigeria.

Methods: An ordinal logistic model that depicted the log-odds in favour of GAM (normal) child was fitted to the data based on surveillance indexed by Weight-For-Height (WFH).

Results:The results showed that the proportional odd of measuring the nutritional status of a child in a nutrition survey using the WFH index has the OR= 7.43 (95% CI, 4.717 to 11.705) times greater, with Wald (1) 2  =74.81, p<0.001, hence a statistically significant effect.

Conclusion: Based on the results and summary of findings, it can be concluded that age is a major predictor of the nutrition status of a child in a nutritional study when the surveillance is based on WFH index unlike sex and measles that do not play a major role.