Multilevel modelling of sociodemographic predictors of various levels of anaemia among women in Mali.

OBJECTIVE Anaemia currently affects 40-80% of women in Africa. While risk factors for anaemia have been intensively studied, research has rarely compared risk factors between mild anaemia and moderate/severe anaemia. Also, the contribution of neighbourhood to the prevalence of anaemia has been rarely studied. The aim of the present study was to identify and compare individual and contextual factors associated with various levels of anaemia among women. DESIGN A multilevel analysis of data from the 2001 Mali Demographic and Health Survey (n 3763) was carried out. Outcomes variables were mild anaemia, moderate-to-severe anaemia and any anaemia. Multilevel regression analyses were performed for each outcome. SETTING Mali, West Africa. SUBJECTS Women (n 3763) aged 15 to 49 years, including 512 pregnant women. RESULTS Among the eleven potential risk factors included in the models, two factors were associated with mild anaemia (BMI and education), three with any anaemia (pregnancy, BMI and education) and six with moderate-to-severe anaemia (pregnancy, BMI, education, wealth, childhood residency and region of residence). Clustering of anaemia within communities was 20% for moderate-to-severe anaemia and 13% for mild anaemia. Despite significant differences in the prevalence of anaemia across regions in Mali, no difference between regions in the risk of mild anaemia was found and only the region of Gao showed a significantly higher risk of moderate-to-severe anaemia. CONCLUSIONS The findings indicate that sociodemographic risk factors as well as clustering of anaemia varies with the severity of anaemia. Specific studies are needed to identify risk factors of mild anaemia as well as its consequences, as mild anaemia accounts for 20-40% of total prevalence of anaemia in Africa.