In Zanzibar, imported infections from mainland Tanzania are typically studied with the assumption that there is limited local transmission in the archipelago [7, 19]. However, this study shows that there are high-risk groups in relation to local transmission who have occupational and behavioural exposures that might benefit from tailored interventions. This study also observed a lower proportion of female malaria cases, which was expected. Recent epidemiologic trends in Zanzibar show that malaria is primarily affecting males between 15 and 45 years old. Characterizing groups that have higher occupational risk and other risk factors for malaria allows for the malaria control program (ZAMEP) to target effective interventions to halt local malaria transmission in Zanzibar. Given that the study was conducted during a surge year and in areas known to report relatively high case numbers, the findings offer valuable insights. Specifically, they suggest that outdoor transmission—potentially driven by occupational and behavioural risk factors—may have played a key role in sustaining local malaria transmission during this time.
Occupations with high levels of outdoor activities were associated with increased risk of malaria in urban areas: night watchmen/police, construction workers, and farmers were the salient high-risk occupational groups. Individuals who work in outdoor occupations are not protected from traditional vector control interventions such as insecticide-treated nets (ITNs) or indoor residual spraying (IRS) and are regularly exposed to mosquitoes for extended periods. Additionally, while sex was not assessed as a risk factor in this study, these identified higher-risk occupations are generally more common among men. It was more common to report more than one occupation in rural areas than in urban areas. The connection between occupations with outdoor or manual work and increased risk of malaria has been observed elsewhere, including in Madagascar and Namibia [10, 20]. The risk of malaria infection among farmers in particular has also been observed in Nigeria and South Africa, especially in rice fields that provide ecological conditions favourable for mosquito larva [10, 20, 21].
While the study did not include entomological data collection, assumptions regarding mosquito biting times were based on previous experience at the study sites and unpublished local data. In Zanzibar, the primary malaria vector is Anopheles arabiensis, which typically exhibits disproportionately between indoors, 22:00–02:00 h, and outdoors earlier in the evenings, 1800–2100 h [15]. While it is common for watchmen to stay outdoors which directly increases their exposure risk, construction workers in the study areas often live in temporary camps located on-site. These sites frequently include large man-made water storage dams, wells, or unfinished swimming pools, especially where tourist facilities are under construction. The local entomology team has regularly detected Anopheles larvae in these water bodies, indicating suitable breeding habitats. Additionally, many workers migrate from mainland Tanzania, where malaria transmission remains higher, raising the likelihood of parasite importation. Combined with factors such as low insecticide-treated net (ITN) usage and night shifts, construction workers may be exposed to mosquito bites during peak vector activity hours.
In both urban and rural areas of Zanzibar, a major behavioural risk factors for malaria is spending time participating in outdoor activities in the evening or early morning, as has been observed in other low transmission settings in Africa and Asia [10, 22]. Night-time activities and taking meals outside at night during the past month were associated with higher odds of malaria in this study. Outdoor activities at night in Zanzibar can include peri-domestic activities including childcare, animal care, cooking, cleaning, and fetching water while outdoor activities away from home can involve social activities, most often during the early evening and night [23].
The linkage between outdoor occupations and outdoor activities with increased risk of exposure to malaria is supported by recent studies in Zanzibar indicating transmission is occurring outdoors due to the mosquito population mainly biting and resting outdoors [19]. In many parts of the malaria endemic world, outdoor biting has been on the rise compared to indoor mosquito-biting behaviour [24,25,26]. These findings strengthen the evidence of outdoor transmission in these areas of Zanzibar, and further the need to find solutions for occupation and behaviour related malaria exposure [22]. More and better outdoor vector control tools are needed, such as topical repellents, long-sleeved clothing, and larval source management. Information and messaging on outdoor-related malaria risk in rural and urban communities is also required, and an approach to malaria prevention in urban areas specifically, with the global rise in urbanization [27, 28]. However, the transition of information and awareness raising to actual change in behaviour will be challenging as people are outside for many reasons and may perceive malaria risk as low in Zanzibar [1, 22].
Previous studies have found travel to be associated with increased risk of malaria in elimination settings [11, 29, 30]; however, one study in 2015 in Zanzibar found no association between travel within Zanzibar and increased risk of malaria [19]. In this study, results showed that local travel within Zanzibar in the last 3 weeks was a risk factor for malaria among local cases who had not left Zanzibar in both urban and rural districts. Additionally, all three high-risk occupation groups in urban areas reported more recent travel among cases than controls. The higher risk may be due to increased exposure to mosquitoes while traveling, travel to areas of Zanzibar that are particularly high-risk for exposure, staying overnight in informal structures while traveling, or low use of prevention tools while travelling [10, 11, 31]. Previous research in Zanzibar has found that men spend more time away from home than women, indicating they could be at higher risk [23].
Not sleeping under a net has been found to be associated with increased risk of malaria in Zanzibar [19]. In this study, reporting prior night net use was strongly associated with lower odds of malaria in the rural districts, and trended towards being protective in urban districts, though the association did not reach statistical significance in urban districts. Overall, there was relatively low reported net use among the study population: 50% of cases and 58% of controls reported prior night net use. Barriers to ITN use may be heat, perception of low mosquito density during certain seasons, or inability to use ITNs while travelling [23]. A recent study on human behaviour and malaria transmission in Zanzibar (a pre-elimination setting) found that participants reported it was easier to protect themselves from mosquito bites inside than outside because of tools like ITNs that can be used inside [23].
This case–control study has several limitations. First, although it was conducted during a surge year with a marked increase in locally acquired malaria cases—and in areas known for high local transmission, the overall incidence remained relatively low in some districts. This may limit the generalizability of the identified risk factors to other transmission settings or seasons. Second, the small number of cases identified in rural districts reduced the statistical power to detect associations with specific occupations or exposures. The third limitation: generalizability is restricted to treatment-seeking populations who are ill (with malaria or some other cause) and who may differ from the general population in important ways. For example, cases that are not detected at health facilities may be lower density and differ socio-demographically. The test-negative design does have the advantage of avoiding selection bias by ensuring that controls are sampled from the same underlying treatment-seeking populations. Fourth, the use of RDTs for diagnostic testing is a limitation because of the possibility of false negatives, given the inability to detect low density infections, and false positives related to the persistence of HRP2. In the absence of HRP2/3 deletions this misclassification is unlikely to have a major impact on results, given the likelihood that symptomatic malaria generally is accompanied by high parasitaemia. In addition, the study primarily aimed to explore risk factors associated with testing positive for malaria—regardless of species—compared to testing negative; additionally, classification of a case as “locally acquired” versus “imported” was based on the surveillance system’s routine criteria, rather than species identification. Limitations of the data include defining prior-night net use as a dichotomous variable (a continuous measure would have allowed for addressing overnight exposure) and lack of data on indoor residual spraying (IRS) (due to an error in the questionnaire). Lastly, due to the operational nature of the study, it was not feasible to record detailed participant flow information as outlined in the STROBE checklist. Cases were identified and enrolled through the mandatory national surveillance system, with no separate research-led process to document refusals or ineligibility, and no follow-up component. Controls were selected from individuals seeking malaria testing at health facilities and enrolled if they tested negative and matched case profiles by age or sex. While some refusals or ineligible controls may have occurred, this was not systematically recorded, as controls are not typically followed in routine surveillance. Nevertheless, the study reflects a scalable, real-world approach embedded within national systems. It was estimated that over 95% of eligible cases were enrolled, and control recruitment was closely aligned with case enrolment.
There are programmatic implications of these study findings. Delineation of specific high-risk occupations allows for improved malaria surveillance, e.g., adapting case investigation forms to capture specific groups, and allows for targeting interventions that can be effective in these specific settings and to these specific populations. ZAMEP and partners have started initiating these changes: ZAMEP is including the reporting of identified high-risk populations of watchmen and construction workers in the urban districts, and is planning to scale up the monitoring these risk groups. Further information-gathering, such as through qualitative methods, of these priority high-risk populations (watchmen, construction workers, and students) in urban areas with local malaria transmission would build knowledge for targeting and tailoring interventions, such as where and when they congregate, risk perception, access and barriers to diagnosis and treatment, and acceptability of and preferences around potential future interventions.
Moreover, the frequency-matched case–control design used in this study is underutilized in malaria operational research, particularly in low-transmission or pre-elimination settings such as Zanzibar. This approach enabled efficient assessment of associations between exposures and locally acquired malaria while controlling for known confounders like age and sex. This study underscores the value of this design—especially its feasibility, cost-effectiveness, and suitability for identifying risk factors in low-incidence settings—particularly when embedded within routine surveillance activities.