The increase recorded in Plasmodium prevalence rate from November 2007 to November 2008 could be explained by the fact that during the latter period, conditions such as rainfall and humidity were more conducive to mosquito development. Generally, when environmental conditions are favourable (monthly rainfall up to 150 mm and humidity around 85%), more productive An. gambiae breeding sites are formed, contributing to the increase of Plasmodium infection . Previous studies carried out in sub-Saharan Africa, in areas of stable malaria characterized by high EIRs, showed that the protective efficacy of LLINs was lower . The reduction recorded in LLIN protective efficacy is confirmed by the fact that several studies performed in three malaria stable West African countries (Ghana, Côte d'Ivoire and Burkina Faso) reported very low reduction on Plasmodium prevalence rate (13% reduction when the control group did not have any nets and 10% reduction when the control group had untreated nets) [19–21]. Increased records in Plasmodium prevalence rate could also be due to the fact that LLINs were not properly used by householders, specifically because we recorded a reduction in malaria transmission in April and July 2008, 9 months after net distribution (unpublished data). During this low transmission and biting period, the perceived benefits of reduction in mosquito bites and of malaria were considered not to be important by household members. Recent studies carried out in Burkina Faso  showed that LLINs were not used during the dry season (period of low biting rates) and when the perceived benefits of reduction in mosquitoes bites and of malaria were considered not to be important, even during the high transmission period. Additionally, both studies have shown that the usefulness of LLINs in reducing malaria prevalence rate and morbidity was moderated by the fact that mosquitoes were considered to be only one of several factors which caused malaria .
From April 2007 to April 2008, a statistically significant reduction was observed in the overall Plasmodium infection rates recorded among children aged between 6 and 59 months. Indeed, activities carried out few months after net distribution by CHWs and mass sensitization campaigns undertaken in the study villages contributed to reduce drastically Plasmodium prevalence rate. It is conceivable that the motivation for the use of LLINs is high shortly after mass and household-to-household sensitization campaigns. This argument is confirmed by a recent study carried out in Burkina Faso showing that the motivation for the use of LLINs decreased after less than a year . The increase in motivation for the use of LLINs could also be explained by the fact that nets were distributed during the short rainy season (from mid-September to mid-November), at a time An. gambiae biting rates are supposed to be high. During this study period, particularly in April, during the rainy season, previous studies carried out in three urban cities of West Africa (Abidjan, Cotonou and Ouagadougou) showed that sleeping under an ITN the night before the survey was protective against Plasmodium infection [23–25]. Additionally, a recent study showed that Plasmodium prevalence rate was strongly influenced by socioeconomic status, and the methods used for prevention depended on their perceived cost . In our study, however, no link was found between the proportion of protective measures and Plasmodium infection rates, perhaps because in villages where we recorded the highest Plasmodium infection rates, the use of fumigating coils and insecticides was high.
Another important finding of our study is that we observed a significant decrease in the overall proportions of clinical malaria cases among children less than 5 years old during both study periods, from April 2007 to April 2008 and from November 2007 to November 2008. This observation might be explained by the considerable number of household members sleeping under LLINs instigated by the household-to-household and mass sensitization campaigns undertaken before and after net distribution. Of note, in this project, although the price of nets was relatively low, we did not carry out free distribution of nets. Similarly, in Tanzania, 10 years ago, social marketing of treated nets was identified as an effective means for malaria control in rural settings. Eighteen months after launching promotion of treated nets, 46% of 312 families with children aged less than 5 years reported that their children were sleeping under treated nets . In a demographic surveillance system area of Tanzania, after a social marketing campaign, ITN coverage of infants rose from < 10% at baseline to > 50% some 3 years later. Treated nets were associated with a 27% increase in survival in children aged 1 month to 4 years . A reduction in the number of clinical malaria cases was also recorded in Kafine, a village located in the northern rice-growing region of Côte d'Ivoire, where the main malaria vector, An. gambiae s.s., is resistant to permethrin and other pyrethroids . In this village, the rate of malaria attacks was twice as high among non-users of mosquito nets. Meanwhile, previous studies performed in several West African malaria stable countries (Côte d'Ivoire, Gambia and Sierra Leone), the differences in treated nets protective efficacies against uncomplicated malaria were low (11%) [20, 27, 28].
Finally, in our study, no significant statistical difference was found between users and non-users of LLINs concerning the prevalence of high parasitaemia levels (> 5000 parasites/μl blood). This observation is in agreement with previous studies carried out in the savannah village of Kafine, northern Côte d'Ivoire, where no difference was found between user and non-user groups of mosquito nets with regard to the prevalence of high parasitaemia, or gametocytes or to the mean parasite load . Generally, in a highly endemic area, there is no marked variation in malaria prevalence rates and morbidity during the season of high prevalence rate. Indeed, in some previous studies performed in East Africa treated nets did not appear to reduce malaria prevalence rate, parasite density and all cause morbidity [29, 30]. Conversely, treated nets have proved to have a significant protective efficacy against high parasitaemia in Kenya among children less than 5 years .
To conclude, training and sensitization activities undertaken before and after net distribution contributed to reducing Plasmodium prevalence from April 2007 to April 2008 and to significantly reducing the number of clinical malaria cases. However, 8-9 months after net distribution, a decrease in the motivation for the use of nets probably favoured a significant increase in the overall Plasmodium prevalence rate from November 2007 to November 2008. Thus, continued long-term social marketing for promoting LLINs holds promise for effective malaria control in rural and urban malaria endemic countries.