Micro-epidemiology of Plasmodium falciparum malaria:: Is there any difference in transmission risk between neighbouring villages?

Background: Malaria control strategies are designed as a solution for either the whole region or the whole country and are assumed to suit every setting. There is a need to shift from this assumption because transmission may be different from one local setting to another. The aim of this study was to assess the risk of clinical malaria given the village of residence among under-five children in rural north-western Burkina Faso. Methods: 867 children ( 6 - 59 months) were randomly selected from four sites. Interviewers visited the children weekly at home over a one-year period and tested them for fever. Children with fever were tested for malaria parasites. An episode of clinical malaria was defined as fever ( axillary temperature >= 37.5 degrees C) + parasites density >= 5,000 parasites/mu l. Logistic regression was used to assess the risk of clinical malaria among children at a given site of residence. Results: Children accumulated 758 person years (PYs). Overall, 597 episodes of clinical malaria were observed, giving an incidence rate of 0.79 per PY. The risk of clinical malaria varied amongst the four sites. Taking one village as reference the odds ratio for the other three sites ranged from 0.66; 95% CI: 0.44 - 0.98 to 1.49; 95% CI: 1.10 - 2.01. Conclusion: Malaria control strategies should be designed to fit the local context. The heterogeneity of transmission should be assessed at the district level to allow cost-effective resource allocation that gives priority to locations with high risk. Functional routine health information systems could provide the necessary data for context specific risk assessment.