The primary risk factor for malaria infection in the low-transmission district of Richard Toll is travel. Most people (80%) with a malaria infection identified through passive case detection reported travel in the previous two weeks. RDT-positive prevalence among family and neighbours of the index case was significantly lower among those who reported no recent travel (0.2%) compared with those who reported travel within the previous 15 days with variations on risk based on location of travel – including travel within the Saint-Louis region (3.1%), Dakar region (10.5%), and other regions of Senegal (33.3%). Index malaria cases were disproportionately male (82%) and infection was more common among male family/neighbours (0.6%) compared with females (0.2%). Although confirmation through additional and longer term surveillance methods is needed, results from this 12-week pilot are consistent with either no or very low levels of local malaria transmission in the Richard Toll district of Senegal and infections appear to be acquired primarily during travel. With marked malaria transmission reduction, burden commonly shifts from young children and pregnant women to adult men with occupational or behavioural factors that put them in contact with infectious mosquitoes [
7]. In other contexts, this has included men who migrate to higher risk areas for work, including forest and plantation work [
16]. Results from a recent case control study in Ethiopia suggest that travel can be a risk factor for malaria infection even when it occurs within areas considered to be endemic [
19].
Intervention and monitoring strategies are needed to effectively target this population of travellers residing in northern Senegal. Reduction of malaria transmission in areas within Senegal to the south of Richard Toll may be the most important factor in reducing malaria infections in Richard Toll as it is travel to these areas that is apparently the risk that reintroduces malaria to the district. Until this malaria transmission reduction is achieved, intervention strategies targeting travellers could include specific communication to use ITNs for prevention when travelling and to seek prompt evaluation and treatment for suspected malaria. Promoting prevention and treatment among this population may be challenging given that many infections will be asymptomatic. Clearly, the index cases were identified based on presentation with fever to a health worker who then tested the case and identified the infection. The finding of additional infections among household members or neighbours must go beyond simple screening for history of fever. On the one hand, asking household and neighbours about recent fever was helpful in identifying a group with increased risk of malaria infection, yet most RDT-positive family and neighbours of index cases reported no current or recent fever (70% asymptomatic). Importantly, a history of recent travel was also helpful in identifying cases. People who frequently travel can be identified – for example, people travelling for school or work in certain occupations, and they can be targeted for blood testing and treatment to clear any identified infections. As malaria transmission has dropped dramatically in northern Senegal, past methods of identifying and monitoring at-risk populations (e.g., young children and pregnant women [
1]) are no longer useful for tracking transmission. Alternatively, measuring prevalence among the newly recognized population at risk through more targeted surveys could provide relevant malaria burden estimates.
Optimizing case investigation
The success of malaria case investigations for transmission reduction will be facilitated by operations that are both effective in identifying all (or nearly all) additional cases and efficient with human and financial resources. The Richard Toll case investigation experience can be used to optimize the efficiency of this strategy. The overall effort tested 5,520 individuals and identified 23 infections – a lot of work for a few cases. Similar to findings in Swaziland [
13], a substantial fraction of infections (14/23) were located within the index case compound. Thirteen infected individuals reported recent travel, and seven reported recent fever. Modification of the case investigation strategy could include selective testing limited to people: 1) residing with the index case compound; 2) neighbours with any recent travel; and, 3) neighbours with any recent fever. This strategy would have identified 20 (87%) of the 23 infections by screening 1,173 individuals – 79% fewer people tested. Information on the remaining three infections suggests that screening for boarding school students would have identified all cases – as the three remaining cases were males attending secondary school or university away from home. Although additional pilot evaluation is needed, school enrolment away from home may serve as a useful screening question for identifying infections. Locally developed evidence to shape reactive case detection components of case investigation is needed [
16]; this study highlights the need to incorporate demographic risk factors in screening during reactive case detection [
7]. More specific testing criteria in this context will save time and resources that would otherwise be spent on unnecessary malaria blood testing.
The current experience in northern Senegal must be taken to scale and evaluated once more to confirm the gain in efficiency. This study reports findings from the pilot phase of a programme during which demographic factors were identified to optimize the investigation protocol. The pilot was characterized by substantial investment in resources; investigation staff were identified from community, facility, district, national, and control programme partner levels, and all levels were equipped with resources necessary for field investigations. Detailed data on all index cases, family, and neighbors were collected and analysed to guide scale-up. When taken to scale, efforts to cost the programme and identify efficiencies at scale will be needed. It will also be critical to monitor operations and identify strategies to maintain high rates of individual participation as well as timely investigations and follow-ups. Information to optimize continued roll-out and scale-up reported here strengthens an initial foundation for this surveillance system component. Continued systems strengthening to ensure a functional, effective, and efficient operation at scale will be needed.
The Richard Toll pilot treated RDT-positive individuals according to national policy using artemether lumefantrine (AL). This artemisinin combination therapy (ACT) is appropriate for control in contexts where re-infection is likely. An elimination approach may consider employing longer acting drugs that clear infection and provide some duration of prophylaxis to further prevent transmission [
7,
20]. Applications in this context could entail using dihydroartemisinin piperaquine – an ACT with a post-treatment prophylactic effect that is longer than other ACT, including AL [
21]. An elimination approach may also include using a gametocytocidal drug such as a single low dose of primaquine recently recommended by WHO for use in combination with ACT for all patients with parasitologically confirmed
Plasmodium falciparum malaria [
22]. Infection management could further be optimized by using a single dose effective treatment; although not currently available, a single dose treatment could be given as directly observed therapy to optimize effectiveness [
20].
Improved diagnostics are another area for optimization in case investigation activities. This pilot used RDTs for reactive case detection. RDTs may miss infections detectable by more sensitive methods [
17], and submicroscopic infections may contribute to transmission in such settings [
18]; thus further evaluation of more sensitive diagnostic tests for case investigation is warranted. At present, the most sensitive diagnostic tests are not easily deployed in field settings. Diagnostic tools that are field-ready and capable of detecting low parasite densities may be needed to improve community-based surveillance [
7,
23], however, within this pilot effort, no emerging malaria infections were observed that had been missed by RDT testing in the initial investigation. In the absence of field-ready tools, attempts to deploy PCR in the field have been made in the context of intensive activities to control spread of artemisinin-resistant parasites in western Cambodia [
24]. The artemisinin resistance containment project deployed PCR for cross-sectional community screening and follow-up and treatment within high-incidence villages. The extent to which employing PCR for cross-sectional community screen and treat work could successfully be applied to ongoing case investigation programmes is unclear.
An alternative for halting transmission among malaria-affected households is the mass drug administration (MDA) approach whereby all household members are treated in lieu of any testing. While this is certainly an option to block transmission, this approach fails to provide any information on infection and precludes analyses that could identify extent of and risk for infection. Initial case investigation work benefits from gathering substantial data on infection and risk factors for infection as this information can inform protocol modifications for improved efficiency.
Study limitations
This 12-week pilot gathered substantial information from index cases and family and neighbours to understand patterns of risk and infection in a low-transmission setting. Limitations of the information gathered during the pilot include the limited scale and time frame. The activity targeted one district, and implementation occurred at a sub-district level due to the national data retention strike. Despite this limited scale, case investigation activities in this area were timely given the context, where recent routine and population-based survey data had indicated very low levels of transmission. Additionally, the pilot was an avenue for exploring risks in Richard Toll district that are likely relevant in neighbouring districts in northern Senegal. The duration of the pilot project was limited; however time was sufficient to generate enough cases and investigations to begin to draw conclusions about risk factors in this context. It is possible that risk factors identified during a limited time frame may not be generalizable to other time periods in the same context. However, the pilot period encompassed peak transmission season, thereby facilitating identification of the most critical risk factors for infection. While there is no reason to believe that the specific time period of this pilot introduced any bias in results, findings should be validated as case investigation strategies are expanded and implemented year-round.
RDTs were a practical tool for point-of-care testing and household and neighbourhood investigation in this pilot. However, more sensitive assays (PCR) would likely identify additional infections. While RDTs may be imperfect, there is no evidence to suggest that individuals who are RDT-negative but harbor a transmissible infection detectable by PCR will differ on the characteristics shown here to be associated with infection. Nonetheless, use of molecular methods to identify all infections would strengthen conclusions on risk factors for infection. Future studies using highly sensitive assays as an evaluation tool can further examine the frequency and transmission potential of RDT-negative but sensitive-assay-positive individuals.
The data collection tool used for this study assessed potential risk factors for infection, including age, sex, occupation, recent travel, and bed net use. It is possible that specific relevant risk factors were not measured because they have not yet been identified. Nonetheless, the risks that were investigated and identified in this study make sociologic sense and lend themselves to action. Travel was identified as an important risk factor, including travel to Dakar and other regions in Senegal. However, data collected in this study are unable to pinpoint the source of infection and should not be used to draw conclusions about specific geographic areas for malaria transmission in Senegal.
Evidence gaps
Travel is clearly a risk for infection; however this study provided insufficient evidence to effectively target travellers at risk for prevention, screening and treatment work. Further work is needed to better define this population at risk, the nature of the risk that they face, and strategies to mitigate risk of infection and treat infections. Recent advances and applications of mobile phone data offer possibility to map mobility and malaria infection risk [
25].
This activity identified potential for introducing relevant screening questions that would reduce human and financial resources needed for reactive case detection. The next iteration of case investigation in northern Senegal should include careful monitoring of such strategies. The work in Richard Toll district operated in the context of a national data retention strike whereby public health workers were withholding routine monitoring data – such as malaria case data – from national health information systems. Subsequently, the pilot activity was limited to a sub-district scale. There is need to monitor and document lessons learned for effective and efficient case investigation at larger scale – including the entire district and expansion to other low-transmission districts in northern Senegal. Evidence from other contexts is needed to build an evidence base around efficient and effective strategies for case investigation. The evidence agenda around case investigation programmes taken to scale should also incorporate measures to track impact on malaria transmission and prevention of reintroduction.