Bed nets for malaria prevention under field conditions in Ethiopia: : Durability, use, and impact on spatial variation of malaria

Background: Despite a remarkable achievement in reducing malaria and in scaling up vector control intervention using long-lasting insecticidal nets (LLINs) in the last 15 years, malaria remains a significant public health problem in Ethiopia. To maximize the benefit of LLINs for malaria reduction, LLINs should provide a serviceable life, and people at risk for malaria should use them. However, field studies show considerable variation in both LLIN serviceable life and utilization in different settings. Moreover, malaria transmission is highly heterogeneous in different geographic locations and over time due to variations in risk factors. To further reduce the malaria burden, it is important to understand the factors that affect malaria clustering. For example, at a micro-geographic scale, it is crucial to target interventions in so-called hotspots (areas with a higher proportion of malaria cases than the overall population). In Ethiopia, several studies have evaluated the spatiotemporal clustering of malaria; however, with the exception of a single study in a southern Ethiopian village with a high malaria infection rate, none quantified the impact of malaria control interventions on the observed clustering. Hence, spatial variation in malaria transmission according to different intervention types in areas with low malaria transmission is not yet fully understood. To fill this knowledge gap, we examined the impact of LLIN use, indoor residual spraying (IRS), and combined LLIN and IRS use on spatial clustering of malaria in settings with low malaria transmission. In summary, evaluating the lifespan of LLINs, the extent of LLIN utilization by people at risk of malaria infection and the impacts of LLIN use alone, IRS use alone or the combination of LLINs and IRS on spatial clustering of malaria, could aid in designing efficient and effective malaria control strategies. General objective: The overall aim of this study was to assess LLINs in real-life field conditions and their impact on spatial variation of malaria in an area targeted for a cluster- randomized controlled trial for malaria prevention in southern-central Ethiopia. The specific objectives: 1) To determine the durability of LLINs under field conditions in terms of attrition, physical integrity, functional survival and bio-efficacy; 2) to assess LLIN ownership and use over time and identify factors associated with LLIN use; 3) to assess malaria infection clusters in areas with LLIN use; and 4) to assess malaria infection clusters in relation to IRS alone or a combination of LLIN and IRS interventions. Methods: We followed up a cohort of 1,532 LLINs every six months from October 2014 to November 2016 to assess their attrition, physical integrity and functional survival under field conditions. In addition, 120 randomly selected LLINs were tested for bio-efficacy (Paper I). In Paper II, we followed a cohort of 17,142 individuals via weekly home visits for 121 weeks to document their LLIN use. We also conducted a survey at 110 weeks after LLIN distribution to determine LLIN ownership. In Paper III, we followed a cohort of 34,548 persons every week from October 2014 to January 2017 and used active and passive case detection mechanisms to identify clinical malaria episodes. A discrete Poisson model was used to identify high rates of spatial, temporal and spatiotemporal clustering of malaria using SaTScan software v9.4.2 (Paper III). Results: In Paper I, we observed high attrition and low functional survival of LLINs in the study period. LLIN attrition was mainly due to disposal. The median functional survival time of the LLINs was only 12 months. The PermaNet ® 2.0 LLIN met the criteria of effective bio-efficacy up to 24 months after distribution. In Paper II, we observed low LLIN ownership and use during the study period. After 110 weeks, only 8% of households owned at least one LLIN. The median proportion of LLIN use per individuals was only 14%. More frequent LLIN use was reported by older age groups, compared with children under five years old. Similarly, people residing in houses whose heads of household had better educations reported more frequent LLIN use, compared with those with higher rates of illiteracy. Having a family size of more than five persons was associated with less frequent use of LLINs, compared with smaller family sizes. In Paper III, we observed spatial, temporal and spatiotemporal clustering of malaria infections in the study area and period. The spatial clustering of malaria at the household level was detected in all study arms (LLIN + IRS, LLIN alone, IRS alone and routine arms) with no significant difference in the risk of clustering between the arms. The risk of malaria clustering was high among households located closer to potential vector breeding sites. Moreover, an overlap was observed between clusters with low rates of LLIN use and clusters with high rates of malaria infection. The risk of malaria infection among people living in clusters with low LLIN use was 2.20 times higher than for people living in areas outside of these clusters (adjusted hazard ratio = 2.20, 95% confidence interval (CI): 1.80–2.60). Conclusion: The bed nets given for malaria prevention did not last long and were utilized as expected under field conditions in southern-central Ethiopia. Therefore, strategies are needed to address these problems and maximize malaria control efforts. We demonstrated variations in the risk of malaria infection across micro-geographic areas. None of the malaria control interventions (LLIN alone, IRS alone or combined LLIN and IRS) affected the occurrence of spatial clustering of malaria. The risk of malaria clustering was high in areas nearer to potential malaria vector breeding sites and thus prioritizing malaria control intervention in these locations could optimize resources. However, locations with low rates of LLIN use also exhibited high rates of malaria infection. Reducing variations in LLIN use in different locations thus also could reduce the risk of infection. Trial registration: PACTR 201411000882128 (8 September 2014)Doktorgradsavhandlin