Risk factors for moderate to severe anemia among children in Benin and Mali: insights from a multilevel analysis.

BACKGROUND: Anemia currently affects 2 billion people throughout the world. Although the immediate causes of anemia among children are known (including malnutrition and infections), the importance of contextual determinants and their relationships with individual effects have rarely been explored.
OBJECTIVE: To identify anemia risk factors at the individual, household, and community levels among Beninese and Malian children, using simple and multilevel regression methods.
METHODS: An analysis was undertaken of nationally representative data collected in 2001 in Benin (n = 2,284) and Mali (n = 2,826) by the Demographic and Health Surveys. Sixteen potential risk factors for anemia were considered at the individual, household, and community levels. Comparative analyses were carried out using simple and multilevel logistic regression models.
RESULTS: Simple and multilevel logistic regression analyses yielded broadly similar results. Risk factors for moderate to severe anemia included incomplete immunization, stunted growth, recent infection, absence of bednet, low household living standard, rural residency (Mali), low maternal education, and low community development index (Benin). In addition, multilevel analysis indicated a clustering level of anemia in communities (intraclass correlation) of 14% and 19% in Benin and Mali, respectively.
CONCLUSIONS: Risk factors for child anemia appeared at all three levels (individual, household and community). Community-level clustering seemed to be low. Therefore, interventions to address anemia need not be village- or region-specific. Identifying a successful and replicable program is now a priority in child survival endeavors. It is likely that such a program would include a focus on improving immunization coverage, increased bednet usage, and reduced protein-energy malnutrition.