Are long-lasting insecticidal nets effective for preventing childhood deaths among non-net users? A community-based cohort study in western Kenya.

BACKGROUND: Increasing the distribution and use of insecticide-treated nets (ITNs) in Sub-Saharan Africa has made controlling malaria with ITNs more practical. We evaluated community effects induced by ITNs, specifically long-lasting insecticidal nets (LLINs), under ordinary conditions in an endemic malaria area of Western Kenya. METHODS: Using the database from Mbita Health and Demographic Surveillance System (HDSS), children younger than 5 years old were assessed over four survey periods. We analyzed the effect of bed net usage, LLIN density and population density of young people around a child on all-cause child mortality (ACCM) rates using Cox PH models. RESULTS: During the study, 14,554 children were followed and 250 deaths were recorded. The adjusted hazard ratios (HRs) for LLIN usage compared with no net usage were not significant among the models: 1.08 (95%CI 0.76-1.52), 1.19 (95%CI 0.69-2.08) and 0.92 (95%CI 0.42-2.02) for LLIN users, untreated net users, and any net users, respectively. A significant increasing linear trend in risk across LLIN density quartiles (HR=1.25; 95%CI 1.03-1.51) and a decreasing linear trend in risk across young population density quartiles among non-net user children (HR=0.77; 95%CI 0.63-0.94) were observed. CONCLUSIONS: Although our data showed that current LLIN coverage level (about 35%) could induce a community effect to protect children sleeping without bed nets even in a malaria-endemic area, it appears that a better system is needed to monitor the current malaria situation globally in order to optimize malaria control programs with limited resources