This study has shown that patients identified through passive surveillance and classified as imported cases by the Swaziland NMCP have networks of contacts who can easily be reached and screened for malaria infection. This preliminary study has implications for the mitigation of imported malaria in Swaziland. Individuals with similar high-risk travel patterns to those of imported malaria cases in Swaziland were able to be contacted and expressed their willingness to be tested for malaria. Key informants among the network contacts were identified and the locations were given where members of the Mozambican community could be accessed. These findings indicate the possibility of using new social network strategies within an active surveillance programme to reduce the risk of onward local transmission caused by imported malaria in elimination settings.
This study found that networks of high-risk travellers can be accessed. All of the index cases reported knowing at least someone who travelled frequently between Swaziland and Mozambique who might be at risk for carrying malaria parasites. Networks were made up mostly of family members and friends, co-workers, or acquaintances from the church. Networks were mostly comprised of Mozambicans living and working in Swaziland while their families remained in Mozambique. The majority were skilled workers in construction or farms or conducting small businesses as metal workers, traders, shoemakers and street vendors of fruits and vegetables. These findings complement the results of Tatem and Smith  who found that most of the population movements between malaria-endemic countries and countries working on elimination is related to trade or family visits; this population movement can contribute to the importation of parasites.
The majority of the network contacts in this particular setting travel to Maputo and Inhambane in Mozambique where they have their homes. This aligns with findings of the NMCP active surveillance programme demonstrating that the majority of imported cases come from these two areas (unpublished data). Many of the network contacts reported travelling to those two areas, showing the likelihood that those networks contain infected individuals. Since these contacts reside in receptive areas in Swaziland (see Figure 2), they could lead to onward transmission. Identifying where individuals have been traveling and linking this information with transmission maps might help to further target screening. Furthermore, cases now residing in more receptive areas in Swazi, which can be established from recent risk mapping work, would help to further prioritize networks.
The interviews determined that most of the travellers do not use malaria chemoprophylaxis prior to travelling. Additionally, some participants reported having auto-medicated themselves with painkillers when they had malaria symptoms. This could interfere with malaria detection by passive surveillance systems and can facilitate the onward transmission of malaria. It has been reported that the incubation time for some plasmodium species can last up to four months . Since the majority of the imported malaria cases come from Mozambique where Plasmodium vivax and Plasmodiun malariae, which are known to have longer incubation time, are present  there is a possibility that individuals infected with these species may transmit the infection to others even months after their travel.
The study showed that network contacts can be identified through chains of referral and be contacted through telephone numbers collected from the index cases. The study used traditional snowball sampling methods whereby the index case interviewed gives the names and contact details of his/her contacts. This method of sampling has been shown to be effective in HIV studies in contacting hard-to-reach populations, such as men who have sex with men and sex workers . Snowball sampling has also been used in studies of non-heterosexual women , and recreational drug users . This method may be an effective strategy in the future within active surveillance systems to rapidly identify imported cases of malaria in Swaziland.
Key informants were identified among the networks, including leaders from the Mozambican Association in Swaziland and other local businessmen who claimed to know many people who travel frequently to Mozambique (Mozambicans and Swazis) and whom they can refer if needed. These key informants claimed to have the ability to organize and assemble at a specific place and time those who travel frequently to Mozambique, opening the possibility for a TLS method in future intervention studies.
This study demonstrated the feasibility of identifying traveller networks for eventual screening and testing for malaria in order to reduce the impact of malaria importation in Swaziland. Once the travellers at risk are identified, special measures can be taken to reduce the risk of onward transmission resulting from importation, including targeted vector control among these groups, continuous screening and treatment of high-risk networks, and the information and education campaigns promoting person protection and treatment seeking behaviour.
This study cannot determine the risk of malaria among the network contacts since no testing was performed and the study period was short. The findings might be different if the study was done at a different time, for example during the peak of imported cases in January. It is possible that the participants might not really acknowledge where they have been or how they crossed the border, even though there was no indication that they did not. Because network contacts were not tested for malaria in this study, the prevalence of malaria among these groups could not be determined and the research could not ascertain how far to continue sampling. The further sampling would have cost and time implications for surveillance agents if adopted as a strategy. Further studies which include testing of individuals are planned and should help to resolve these issues. Since no study has yet evaluated traveller networks associated with malaria importation into Swaziland to date, there is no reference for comparable results.