Many vector-borne diseases are acquired in the home, usually through exposure to nocturnal, endophilic, and anthropophagic vectors 1. Fortunately, even very simple changes in house design can protect people against exposure to mosquito bites 2 and malaria infection 345. The primary malaria vectors of Africa prefer feeding on humans in the middle of the night when they are asleep. Thus, they usually have to find their way into the houses to obtain blood and survive 6.

Anopheles gambiae s.l. mosquitoes are well adapted for entering houses because they fly upwards when encountering a vertical surface 7. Attracted by human odour from inside the house they typically reach the wall, travel vertically along its surface and then enter through the eave gap between the wall and the roof. This observation is reinforced by studies showing that houses with open eaves and those lacking ceilings are associated with increased mosquito numbers and higher levels of malaria compared to the ones with closed eaves and the ones with ceilings 34. In The Gambia, children who lived in houses with closed eaves and metal roofs but slept without bed nets had fewer Plasmodium falciparum malaria attacks than children who slept in houses with open eaves and also had no bed nets 8.

In the early twentieth century, improved housing and screening were regarded as priority methods of controlling malaria. Italian field experiments on proofing houses against mosquitoes were the very first successful malaria control trials 45. People living in poor houses (incomplete or with walls and roofs made of palm thatch and mud) have a higher exposure to malaria than people occupying houses with complete brick and plaster walls and tile roofs. House screening was also found to reduce mosquito human biting rates as well as malaria infections in settings as diverse as the United States, Greece and Italy 4. More recently, clinical trials have shown that both full house screening and ceilings alone provide valuable protection against anemia and exposure to malaria transmission in rural parts of The Gambia 9. The broader potential of window screening, closed eaves and ceilings for preventing entry of a variety of culicine mosquitoes into houses has recently been established in both west and east Africa [10, Ogoma et al unpublished]. Culicines are vectors of a variety of viral and parasitic infections and crucially cause most of the biting nuisance which motivates uptake of household and personal protection measures. The most abundant of these mosquitoes is Culex quinquefasciatus, a vector of Wuchereria bancrofti, which causes lymphatic filariasis and arboviruses, such as West Nile Virus 11 and Chikungunya 12, mainly in Africa. Others like Mansonia sp. transmit Brugia malayi and B. timori more specifically in south Asian countries 13. Apart from being disease vectors they are also the most common human-biting culicines, consequently contributing the bulk of nuisance bites, especially in urban areas 14. The role of nuisance biting mosquitoes in the control of malaria requires particular consideration by control programmes relying upon community participation. For example, low levels of susceptibility of Cx. quinquefasciatus to insecticide as was reported in Tanzania 15 and the resulting low efficacy of ITNs against this widespread, nuisance-biting species has been linked to reduced public acceptance of ITNs 1617.

This study was carried out in Dar es Salaam, where overall ITN usage has remained consistently and disappointingly low (26% coverage), but window screening and ceiling boards became increasingly common between 2004 and 2006 18. Interestingly, this was coupled with a simultaneous decline in malaria prevalence, which could not be explained by changes in coverage of any other intervention (Figure 1). The results of these large, cluster-sampled, cross-sectional surveys, prompted us to further investigate the potential of window screening, ceiling boards and blocking of eaves, particularly considering complementary entomological studies showing that ITNs confer less protection against An. gambiae s.s in well-screened and ceilinged houses 18. Motives behind installation of different types of ceilings and screens by community members were evaluated. Human preference for blocking specific mosquito entry points was compared with the known preference of mosquitoes for various points of entry into the house. In addition, investigations on how much residents spent on ceilings and window screens were done. The overall goal was to understand and improve the acceptability of this intervention for incorporation into integrated vector management strategies.

Understanding the interactions of the main malaria vectors with changing self-protection behaviors' of humans is essential to success in programmatic settings. In this cross-sectional study, more people lived in houses with screened windows than in houses with ceilings or closed eaves. Nevertheless, the combined coverage of the latter two, both of which prevent house entry by mosquitoes through the eaves, was also high. The general perception of the community was that complete proofing of their houses was more beneficial than partial house proofing, as was illustrated by the positive correlation between ceilings, window screens and closed eaves. Community knowledge of the causative relationship between mosquito bites and malaria transmission was not investigated explicitly. The very low number of respondents who mentioned malaria prevention as a reason for installing any of these measures indicates that their main motivation was to protect against mosquito bites generally. Many residents of Dar es Salaam clearly understood that installation of ceilings protects them from mosquitoes and some even associate this with protection against malaria infection, yet this intervention has not been widely encouraged or promoted. Other motives for installation of ceilings mentioned included fashionability (especially the gypsum type of ceiling which was considered quite stylish; Figure 4) as well as keeping the house cool. This shows that ceilings are perceived to serve more than one function and may be promoted based on multiple benefits. Therefore, it should be relatively easy to promote them, particularly in urban areas where they are probably most appropriate. Indeed, screened ceilings have recently been shown to reduce densities of both the Anophelines that transmit malaria and the Culicines responsible for most of the indoor nuisance biting in both west and east Africa [3910, Ogoma et al., unpublished].

Given the coverage levels observed here, house-proofing may well prove to be a useful strategy for not only equitably protecting entire households against mosquito bites but also for achieving community-level suppression of malaria transmission so that even a remaining minority lacking them benefit from the "mass effect" 2122. In fact, in Dar es Salaam this may be already happening: (Figure 1). Steady decline of malaria prevalence between April 2004 and March 2007 was associated with increased coverage of screenings and ceilings while no other intervention experienced substantial increases in coverage except for larviciding which was restricted to only three of the 15 study wards in the final study year (Figure 1) 18.

Generally the residents of the city appeared to try as much as possible to protect themselves against mosquito bites by blocking entry to their houses, as depicted by the rarity of houses (13.8%), which did not have ceilings or windows screens and had open eaves. Perhaps most exciting is the prospect of what might be possible if these materials could be treated with effective, long-lasting insecticide formulations to achieve a substantially enhanced level of household and community-level transmission 21. This is further substantiated by a study which was carried out in Burkina Faso, West Africa, illustrating substantial reduction in the levels of malaria transmission when permethrin treated curtains were hanged on doors, windows and eaves 23. Protection of all members within a household, beyond merely those young children and pregnant women at great risk is essential to achieve maximum control and even elimination of malaria 22. Mosquito-proofing of houses therefore offers the significant advantage of equitably protecting all members of a particular household, even those that are not sleeping under a bed net.

The high coverage of screens and ceilings already attained in Dar es Salaam suggests that this is a vector control measure, which can be readily delivered to large populations in many towns and cities across Africa, particularly if the installment costs can be reduced. This study shows that the initial cost of installing window screens is comparable with that of providing bed nets for all occupants, but more studies should be carried out in order to ascertain the long term cost per person per year based on the durability of these two alternatives. Interestingly most houses which neither had ceilings nor window screens but had closed eaves were initially built this way, and, therefore, no additional cost was required for blocking eaves. Blocking of eaves might well be one the cheapest of the three options but schemes for promoting awareness and understanding of these accessible options for household-based control need to be developed and evaluated. The value of this approach is bolstered by the observation that residents of houses with ceilings, screened windows, and especially the combination of both, take advantage of this protection by spending more time indoors at night 24. It is particularly striking that while the existing window screening in Dar es Salaam has a greater market value than that of insecticide-treated nets that of ceiling dwarfs either one. With a total market value exceeding US $40 million, this is clearly a tool which the residents of Dar es Salaam have prioritized and invested in and, therefore, has great potential as an intervention tool in and beyond this particular setting.

Due to high coverage of screens and ceilings, it is concluded that people have readily accepted this method as a way of protecting themselves from mosquito bites and, perhaps inadvertently in many cases, reducing malaria transmission. The prioritization of ceilings to the extent that the residents have paid over US $40 million to install them, suggests ready opportunities for national and international programmes to developing this intervention strategy more proactively and deliberately.

Although cost is the most important constraint on the choice and degree of use of these methods, it is remarkable that coverage with a combination of closed eaves or ceilings equals the 2010 RBM target for ITNs of 80% 25, while that of ITNs in Dar es Salaam remains stagnant at a mere 26% (Figure 1) 18. This is all the more notable because this particular intervention does not feature in the National Medium Term Strategic Plan and is neither subsidized nor actively promoted.

It seems that most residents of Dar es Salaam know about the value of mosquito-proofing houses but need access to, and information about, cheaper and more durable materials which would ideally have insecticidal and/or excito-repellent properties, which would kill adult mosquitoes directly or act as a more effective barrier for preventing house entry 21. Moreover, since blocking of eaves seems to be a particularly effective and affordable option, netting materials suitable for window screening could also be used for screening eaves, since they would interfere less with airflow and indoor temperatures than simply blocking this gap.

In order to fully understand the cost-effectiveness of house screening, additional information is required on the durability of the materials used for ceilings and window screens so that the long term effectiveness and costs of this intervention can be determined. In addition, development and evaluation of effective models for promotion and subsidization should be prioritized. Lastly, there is an urgent need to engage policy makers in active consideration of mosquito-proofing houses as one of the tools for integrated control, and perhaps one that can be considered as "low hanging fruit" in the urban context.

The authors declare that they have no competing interests.

SBO designed the questionnaire used in the study, participated in the collection of data, analyzed the data and drafted the manuscript in consultation with the other authors. KK, implemented and managed the Urban Malaria Control Program in which this study was undertaken, including the household surveys, assisted in development of the questionnaire, and assisted in drafting the manuscript. MS, PPC, and NJG helped supervise collection of the data and drafting of the manuscript. WRM supervised the study and participated in drafting the manuscript; GFK conceived the study and supervised all aspects of the experimental design, data analysis, interpretation of data and drafting of the manuscript. All authors read and approved the final manuscript.