The role of human and mosquito behaviour in the efficacy of a house-based intervention

Abstract

Housing improvement such as blocking eaves and screening windows can help in reducing exposure to indoor biting mosquitoes. The impacts of physical barriers could potentially be boosted by the addition of a mechanism that kills mosquitoes as they attempt to enter the house. One example is to combine household screening with EaveTubes, which are insecticide-treated tubes inserted into closed eaves that attract and kill host-searching mosquitoes. The epidemiological impact of screening + EaveTubes is being evaluated in a large cluster randomized trial in Cote d'Ivoire. The study presented here is designed as a complement to this trial to help better understand the functional roles of screening and EaveTubes. We began by evaluating householder behaviour and household condition in the study villages. This work revealed that doors (and to some extent windows) were left open for large parts of the evening and morning, and that even houses modified to make them more 'mosquito proof' often had possible entry points for mosquitoes. We next built two realistic experimental houses in a village to enable us to explore how these aspects of behaviour and household quality affected the impact of screening and EaveTubes. We found that screening could have a substantial impact on indoor mosquito densities, even with realistic household condition and behaviour. By contrast, EaveTubes had no significant impact on indoor mosquito density, either as a stand-alone intervention or in combination with screening. However, there was evidence that mosquitoes recruited to the EaveTubes, and the resulting mortality could create a community benefit. These complementary modes of action of screening and EaveTubes support the rationale of combining the technologies to create a 'Lethal House Lure'. This article is part of the theme issue 'Novel control strategies for mosquito-borne diseases'.

Keywords: EaveTubes; housing improvement; human behaviour; mosquito behaviour; screening; vector control.

Figure 1.

Plan of an experimental house.

Figure 2.

Pictures of the experimental houses. (a) The two experimental houses; (b) back of a house showing EaveTubes; (c) metallic window from the inside; (d) metallic front door; (e) chicken wire frame to put on the front door, which allowed mosquito entry but prevented entry of reptiles and rats when the doors were open.

Figure 3.

Proportion of houses with open doors (a), people awake (b) or open windows (c). It is given regarding the CRT treatment arm, control or EaveTubes+screening and it was assessed between December 2017 and June 2018.

Figure 4.

Mean (±s.e.) number of An. gambiae mosquitoes captured outside or inside experimental houses per night, depending on house type (Experiment 1). There were three house typologies: a standard village style house with closed eaves but no EaveTubes or screening; houses fitted with EaveTubes alone and no screening; houses fitted with screening + EaveTubes. In addition, householder behaviour was managed to either reflect typical behaviour in which doors and windows were open for part of the evening and morning (see Methods and figure 3), or behaviour was modified so that doors and windows were kept closed from 18.00 to 8.00. Mosquitoes were collected by HLC from 18.00 to 8.00. The means represent 48 nights of capture with individual treatments replicated 24 times.

Figure 5.

Mean (±s.e.) number of An. gambiae mosquitoes captured per hour outside or inside experimental houses, depending on house type (Experiment 1). There were three house typologies: a standard village style house with closed eaves but no EaveTubes or screening; houses fitted with EaveTubes alone and no screening; houses fitted with screening + EaveTubes. In addition, householder behaviour was managed to either reflect typical behaviour in which doors and windows were open for part of the evening and morning (see Methods and figure 3), or behaviour was modified so that doors and windows were kept closed from 18.00 to 8.00. Mosquitoes were collected by HLC from 18.00 to 8.00. The means represent 48 nights of capture with individual treatments replicated 24 times.

Figure 6.

Mean (±s.e.) number of An. gambiae mosquitoes captured outside or inside experimental houses per night, depending on house type (Experiment 2). There were four house typologies: a standard village style house with closed eaves but with no EaveTubes or screening; houses fitted with EaveTubes alone and no screening; houses fitted with no EaveTubes (closed eaves) and screening alone; or houses fitted with screening + EaveTubes. Householder behaviour was managed to reflect typical behaviour in which doors and windows were open for part of the evening and morning (see Methods and figure 3). Mosquitoes were collected by HLC from 18.00 to 8.00. The means represent 48 nights of capture with individual treatments replicated 24 times.

Figure 7.

Mean (±s.e.) number of An. gambiae mosquitoes captured outside or inside experimental houses per night, depending on house type (Experiment 3). There were five house typologies: a standard village style house with closed eaves but with no EaveTubes or screening; houses fitted with EaveTubes alone and no screening; houses fitted with no EaveTubes (closed eaves) and screening alone; houses fitted with screening + EaveTubes; or houses fitted with screening but with the EaveTubes inserts removed so that the eaves were open. Householder behaviour was managed to reflect typical behaviour in which doors and windows were open for part of the evening and morning (see Methods and figure 3). Furthermore, the window screening was deliberately damaged (four 4 × 4 cm holes added per window) and the doors modified to create a 1 cm gap above and below the door to make the house condition more representative of a typical village house. Mosquitoes were collected by HLC from 18.00 to 8.00. The means represent 40 nights of capture with individual treatments replicated 16 times.