This study estimates the relative importance to child school performance (indicated by primary repetition) of fever, malaria and some social determinants at the cluster level. It uses individual, household and cluster surveys from the Demographic and Health Surveys conducted in Mali in 2001 and 2006 (MDHS). It also provides a discussion about the use of fever as an indicator of malaria in large cross-sectional surveys by comparing the 2001 and 2006 MDHS, which were realised during two different transmission seasons (dry and rainy seasons). Geographic Information System and DHS Global Positioning System datasets were used to extract age-specific malariological measures from reliable maps of the prevalence and transmission intensity of malaria. We show that fever is not a reliable proxy for malaria at the cluster level, and we recommend the use of spatial measures of malaria prevalence for future research. Cross-sectional regression analysis on data aggregated to the group-level suggests that a higher prevalence of malaria in a community is linked to higher primary repetition rates, but confirmatory studies are needed.
Copyright 2010 Elsevier Ltd. All rights reserved.