Strategic priorities of national malaria control programs vary depending on many factors, including the natural potential for endemic malaria transmission, the levels of control that have been achieved, and the proximity to other malaria endemic countries. In countries where malaria is rare, malaria importation is of primary concern. Countries with well-developed health systems, good surveillance through national case reporting, and no local malaria can simply count the number of malaria cases that occur. In the United States in 2008, 1,298 detected malaria cases were imported mainly in tourists inadequately protected by chemoprophylaxis and immigrants, and none triggered any further transmission24
; autochthonous cases do occur, but only rarely25
. Countries that have recently reduced their malaria burden, but that have strong human travel connections with other malaria endemic countries26
with poorly developed health systems, or some level of ongoing transmission must devise alternative methods for estimating malaria importation rates. Key information is required to set up appropriate measures: contributions of various routes of malaria importation, reliable estimates of malaria importation rates and of the potential for local transmission, and the total impact of these quantities on the local endemic and non-endemic transmission. These estimates must also be made in a highly dynamic environment as neighboring countries also control malaria. Our analysis suggests that citizens of Zanzibar who travel to malaria endemic regions are likely to be the single most important source of importation. Mobile phone data analysis estimates that less than 1.6 malaria infections are imported per 1,000 Zanzibar residents every year, with a credible value of less than 3.7 infections per 1,000 people per year according to mathematical models. This analysis suggests that Rc
is below 1 in most places, as most local transmission is driven by imported malaria. Therefore, our results suggest that, if current levels of control are sustained, malaria would eventually disappear from most of the island, except possibly in a few residual foci if no imported infections occurred.
Therefore, enhanced prevention and control strategies should aim at mitigating pathogen importation to move towards malaria elimination and prevent reintroduction. The government could provide free chemoprophylaxis through the health system to Zanzibar residents travelling to endemic mainland using drugs that minimize the risk of developing resistance27
. As immunity to malaria parasite declines in low endemic settings, such travelers are even more vulnerable to infection, stressing the rationale to distribute prophylaxis. In addition, each case of infection should be reported to health authorities, which highlights the need for enhanced surveillance28
. Measures should include border screening especially at ports of landing, active case detection and investigation to set up appropriate reactive measures in the environment of the cases. Surveillance tools rely on the availability of sensitive and specific diagnostic tools such as rapid diagnostic tests (RDTs), which require further improvement29
. Finally, vector control measures could mitigate the importation of infectious anopheles, for instance by using insecticide in planes flying from malaria endemic region30
, even though this is likely to be a minor issue. Some of these measures may be technically, economically and operationally challenging and require prior feasibility assessment combined with public health education campaigns.
Imported malaria contributes differently to local dynamics depending on the number of incoming pathogens and the length of stay as addressed here, but also on when and where malaria is imported. Formal and informal human movements are increasingly becoming amenable to study as humans use mobile phones. Nevertheless, mobile phone usage data are subject to biases based on data availability, ownership and usage patterns. Principal malaria transmission season on mainland starts between November and January and ends approximately between April and June. Therefore, the current survey extending over a three-month period from October to December only captures a subset of potential malaria importation. Furthermore, the available Zantel mobile records did not provide further information on local cell tower usage, so it was not known where residents and visitors stayed in Zanzibar. As a result, even though malaria importation could be quantified for the main regions of Zanzibar, no local foci of residual transmission can be identified at higher spatial resolutions. Finally, even though general population movement to and from Zanzibar mainly rely on ferry transportation, informal movement on small fishing boats also represent a pathway for malaria importation and can only be partially captured through mobile phone records. These movements create hotspots of malaria importation mainly in coastal communities. As health authorities know these hotspots, the Zanzibar Malaria Control Programme (ZMCP) can intervene and mitigate malaria importation in these communities.
Some of these limitations have been addressed by cross-validating estimates with two completely independent methods using different data. Despite the uncertainties, mobile phone usage records have provided useful and reasonably accurate information about the lengths of stay in Zanzibar and the relative importance of various routes of malaria importation. These estimates could be improved through further studies focusing on the places where people stay in Zanzibar, and gathering mobile phone data over a longer study period to better take into account malaria seasonality. To move towards elimination, additional tools should be used to complement current assessment especially in terms of risk mapping. The Malaria Atlas Project (MAP) provides with endemicity maps based on prevalence surveys at a global, continental or country level. Using mathematical models31,32,33
, these maps were converted to EIR risk maps, a more relevant measure for risk transmission15
. Nevertheless, as Zanzibar is getting closer to elimination, finer scale mapping based on environmental, epidemiological, population, vector, social and economical factors is required to identify residual foci in order to implement rapid localized response.
As very low levels of transmission are reached, programs move from a focus on control to a focus on pre-elimination and elimination, and finally prevention of re-introduction. Quantitative approaches to human movement and the spread of malaria are increasingly useful for government to appropriately decide when shifting efforts to stemming importation. Mobility patterns can be used to estimate the import of infectious disease into geographic regions with low endemic states. The importance of control methods will depend on the vectors and immuno-epidemiology of the pathogen, as for dengue which has a shorter infectious period, daytime biting vectors, and life-long immunity34
. As methods improve for quantifying malaria importation and transmission, they can be applied more widely to advise countries on strategic planning across the region and worldwide.