Abstract
Anaemia is a widespread public health concern. It is estimated that globally 47 percent of young children are anaemic (WHO, 2005). In children, anaemia can impair development and increase susceptibility to infectious diseases. A lack of sufficient food rich in iron and other micronutrients are the commonest cause of the condition. Underweight is an indicator of both short and long term malnutrition and may reflect poor feeding practices or recent episodes of illness.
The paper investigates the link between children's nutritional status (weight-for-age) and anaemia. It aims to answer two the research questions:
(1) Is anaemia endogenous to weight-for-age?
(2) What are common observed socioeconomic and demographic risk factors associated with weight-for-age for anaemic and not anaemic children?
An observed correlation between unobserved covariates that might influence both anaemia and weight-for-age was tested to check whether endogeneity exists in order that the appropriate model could be used.
This study uses a cross-sectional study of 2,479 children aged between 6 and 54 months from the Democratic Republic of Congo. Initially, 24% of children were classified as underweight and 71% were anaemic. Endogenous switching regression models with Full Information Maximum Likelihood were fitted to the data.
The rationality of applying an endogenous switching regression model is that it
captures the direct effects of individual observed variable included in the models and the indirect influence of correlated unobserved factors for both weight-for-age and anaemia. If these were ignored then there could be bias in the estimates of the effect of the observed individual variables.
The results indicate that anaemia is significantly associated with child nutritional status (weight-for-age) and most of the risk factors associated with anaemia are also found to be related with weight-for-age. These include age, sex, maternal level of education, whether the mother is anaemic and household wealth status.
The paper investigates the link between children's nutritional status (weight-for-age) and anaemia. It aims to answer two the research questions:
(1) Is anaemia endogenous to weight-for-age?
(2) What are common observed socioeconomic and demographic risk factors associated with weight-for-age for anaemic and not anaemic children?
An observed correlation between unobserved covariates that might influence both anaemia and weight-for-age was tested to check whether endogeneity exists in order that the appropriate model could be used.
This study uses a cross-sectional study of 2,479 children aged between 6 and 54 months from the Democratic Republic of Congo. Initially, 24% of children were classified as underweight and 71% were anaemic. Endogenous switching regression models with Full Information Maximum Likelihood were fitted to the data.
The rationality of applying an endogenous switching regression model is that it
captures the direct effects of individual observed variable included in the models and the indirect influence of correlated unobserved factors for both weight-for-age and anaemia. If these were ignored then there could be bias in the estimates of the effect of the observed individual variables.
The results indicate that anaemia is significantly associated with child nutritional status (weight-for-age) and most of the risk factors associated with anaemia are also found to be related with weight-for-age. These include age, sex, maternal level of education, whether the mother is anaemic and household wealth status.
| Original language | English |
|---|---|
| Pages | 2132-2139 |
| Publication status | Published - Dec 2012 |
| Externally published | Yes |
| Event | 58th World Statistics Congress of the International Statistical Institute - Dublin, United Kingdom Duration: 21 Aug 2011 → 26 Aug 2011 |
Conference
| Conference | 58th World Statistics Congress of the International Statistical Institute |
|---|---|
| Country/Territory | United Kingdom |
| City | Dublin |
| Period | 21/08/11 → 26/08/11 |
