To the best of our knowledge, this study is the longest follow up of the durability and functionality of house modifications designed to reduce the entry of malaria mosquitoes. The doors and windows themselves proved very robust, with no damage recorded to either them or their frames. Most looked new and were clean suggesting that the owners took good care of them. Although minor cracks were noted in the mortar around doors and windows, these were most likely present from the time they were first installed (“Hairlike cracks were apparent in some concreting around the door…...” ). Hairline cracks are a common occurrence when cement and mud are married, as in most rural construction work in The Gambia. They appear as the cement dries in the heat.
There were several features of the doors that should be improved or removed. Firstly, only 39% of the door lock handles were present and in working order. The lock mechanisms commonly failed, resulting in homeowners removing the handles so that the doors could be secured with a chain and padlock. Stronger lock handles are needed, or they could be substituted with simpler solutions, such as using bolts internally and hasp and staple locks with padlocks externally. Although the use of padlocks externally is a common practice in The Gambia and other countries in sub-Saharan Africa, it is potentially dangerous since individuals could be locked indoors and unable to exit in a fire. Secondly, the blinds were not fit for purpose, with only 4% of door blinds and 18% of window blinds still present and working. Of the four prototype doors it is the concertinaed door (Figure 1A), which has no blinds and could be mass-produced from a single sheet of metal, which would be the most robust prototype recommended for village use. Thirdly, although 82% of the self-closing doors closed smoothly only 41% of doors fully shut automatically. Of the few doors which stuck (11 in total), most 7/11 (64%) were catching on the bottom right-hand side. Such damage may have been caused by children swinging on the doors and pulling them away from their hinges on the opposite side or even bending the door blade itself. Alternatively, it could be due to slight shifting of the doors after they were first installed, before the cement set. Room owners were asked not to open doors for at least 48 h following installation , but this may not have always been strictly adhered to. The doors are designed to close automatically by using spring-loaded hinges. The idea behind this is to facilitate closing the doors at night to reduce the number of mosquitoes entering the house. Yet, in keeping with observations in the original study, we found most front doors (68%) propped open with sticks or a heavy object (brick or stool) to allow easy access into the home and to promote cross ventilation. Jawara and co-workers noted “that generally the self-closing prototype doors are open for much shorter periods than control doors, although after 06.00 h, with the prototype doors, some doors are propped open to allow housework to take place unimpeded.” In common with many interventions, health messaging is also required. In this case, house screening needs to be accompanied with strong messaging on keeping doors and windows closed at night.
The novel screened doors and windows were well liked in general and particularly for their ability to keep mosquitoes and other insects out and offer security. This was broadly in agreement with findings from the initial intervention where across all Focus Group Discussions (FGDs), participants mentioned that the doors provided privacy, kept out mosquitoes and were attractive to look at . Indeed, the general view amongst the community was that the doors and windows were beautiful and were a major improvement to their homes. The most frequent complaints concerned the poor quality of the lock handles and the blinds.
There were requests for windows which could be opened, five respondents said that their houses were too hot and one mentioned lack of ventilation. Some occupants told MCT that the houses became too hot when the metal doors had been closed all day, while they were away working in the fields. When they returned and opened the door, they were met by a wave of heat. The original study however, showed that there was no difference in indoor temperatures between control and intervention units. Study houses were almost entirely outside the human comfort index for most of the night, being too hot before midnight and too cold after midnight . Also, in the original study, most FGD respondents liked the windows because they allowed ventilation and looked attractive. Respondents noticed that “fresh air enters the house due to the holes” and it is “beautiful just like the doors.” Nevertheless in some FGDs, people were concerned about the windows not opening: “The major problem with the windows is their lack of opening but had it been they open and close it would be the best.”
Three respondents in our follow up survey complained of back doors allowing water ingress. Since all the houses had front verandas, front doors were protected from the rain by the overhanging roof. Only two of the three who complained about water inside the house had any sign of water being present (trace inside back doors), also they were responding during a particularly heavy downpour so were conceivably influenced by the heavy rain, strong wind and flooding outdoors when interviewed.
Most people hung curtains in their doors (82%) and in many of their windows (43%). The most common explanation for the curtains was to provide privacy, especially when the door was open. Seven said to keep light out, especially when there was lightning during the heavy storms that occur chiefly during the early rainy season at the time of interview. The fear of lightning probably stems from the danger of being physically injured or the house being burnt down as can happen with thatched-roofed houses. Although not stated, it appeared to the investigators that curtains were mainly used to decorate the house. Each curtain was bright, clean and unique, probably serving as markers of wealth, beauty, individuality and hospitality. It is also likely that people use curtains for privacy, rather than closing the door which may be considered rude, and to allow small children to come in and out of the house more easily. Having the curtains down inside the doors will reduce the amount of dust that enters the house during the dry season when strong Harmattan winds bring dust down from the Sahara but will also impede ventilation in the house. A well-ventilated house is important since it will help keep the house cool at night and increase the likelihood of people using a mosquito net [8, 9], it is also likely to reduce mosquito house entry  and reduce the incidence of respiratory disease .
This study has several limitations. Firstly, the sample size is small with new doors and windows installed in just 31 house units. Nonetheless the findings were replicated across units with different prototype doors and windows. Secondly, the acceptability survey is susceptible to social desirability bias since some house owners may have either given answers they believed the researchers expected or just wished to be polite. Having been given brand new doors and windows, they may have been unwilling to offend by voicing criticisms of the intervention.