Humanitarian Evaluation of Spatial Emanators for Malaria Control

Humanitarian emergencies increase malaria risk while fundamental logistical and biological constraints adversely affect the delivery and efficacy of conventional vector control tools. Long-lasting spatial emanators are lightweight, passive interventions potentially suited to strategic stockpiling and rapid deployment, but evidence on large-scale implementation and public health impact under emergency conditions remains limited. This study will evaluate the operational distribution and effectiveness of Guardian™, a World Health Organization recommended long-lasting transfluthrin spatial emanator, in northern South Sudan, to support its transition to a routine humanitarian vector control commodity.

This is a 26-month prospective, non-randomized, controlled, mixed-methods implementation-effectiveness evaluation study. The study will assess whether humanitarian delivery systems can achieve and sustain adequate household coverage, correct placement, continued use, acceptability, safety, and product durability, while evaluating malaria infection prevalence and human-vector contact.

The study will comprise two sequential operational phases, targeting 3 distinct beneficiary groups. Newly displaced refugees are recent arrivals living in reception, temporary, or transitional settings before or during settlement. Approximately 5,000 newly arrived refugees will be targeted for Guardian™ deployment in each study phase. Established camp residents are refugees living in Ajuong Thok and Pamir, where more stable settlement conditions support repeated monitoring and linkage with routine health services. Refugee-hosting communities and host-settled communities (former refugees) comprise the population included in the Yida host-community comparison and surrounding dispersed host communities. These settings provide information on delivery and use under different housing, access, and population conditions. Dispersed host communities will enter the evaluation during Phase 2.

Guardian™ will be delivered through existing humanitarian procurement, warehousing, transport, distribution, and community-engagement systems. UNHCR will coordinate in-country logistics and distribution. The MENTOR Initiative will support procurement and implementation and lead monitoring and evaluation with the University of Nevada, Las Vegas and national partners. Household intervention allocations will be determined by shelter volume according to the manufacturer's instructions. Distribution will be accompanied by education, communication and information on unpacking, placement, intended use, and reporting of concerns. Stock movements, distribution completeness, missed households, product damage, and implementation adaptations will be documented throughout the delivery pathway.

During Phase 1, Guardian™ will be distributed house-to-house in a designated intervention area within Yida and throughout Ajuong Thok refugee camp. A geographically separate area of Yida and Pamir refugee camp will serve as contemporaneous controls. Intervention allocation will be determined prospectively at the geographic area level without randomization. Existing malaria prevention, diagnosis, and treatment services will continue in all study populations. Two population-specific effectiveness comparisons will be evaluated separately: Yida intervention versus Yida control, and Ajuong Thok versus Pamir. Phase 1 monitoring will continue for 12 months after deployment.

Implementation monitoring will combine repeated cross-sectional household surveys with a nested longitudinal household cohort. During Phase 1, repeated cross-sectional surveys will sample 15 geographic clusters and 20 households per cluster in each of the two main intervention populations, providing approximately 300 households in Yida and 300 in Ajuong Thok per round. Six rounds will be conducted at 1-2 weeks and at 1, 3, 6, 9, and 12 months after distribution, generating approximately 3,600 household-round observations across the two populations. New random household samples will be selected at each round from updated sampling frames. The separate Phase 1 new-arrival population will contribute approximately 300 households per round at 1-2 weeks, 1 month, and 3 months, providing approximately 900 additional household-round observations.

The Phase 1 longitudinal implementation cohort will enroll 15 households per cluster within the same 15 clusters in each main intervention population, providing 225 households in Yida and 225 in Ajuong Thok (450 households overall). Allowing for approximately 20% loss to follow-up, around 360 households are expected to remain under observation through Phase 1. Repeated cross-sectional surveys will estimate population-level coverage, allocation adequacy, correct use, acceptability, and safety indicators. The longitudinal cohort will characterize trajectories of original-unit retention, placement, continued use, loss, damage, transfer, replacement, and discontinuation within the same households.

The primary effectiveness evaluation will assess all-age malaria infection prevalence 6 months after Phase 1 deployment using repeated cross-sectional household surveys. Surveys will be undertaken at baseline and at 3, 6 and 12 months post-distribution. The 6-month survey will provide the primary effectiveness assessment; the 3-month survey will assess intermediate changes, and the 12-month survey will measure persistence of impact across the operational lifespan of this intervention. Thirty-five fixed geographic sampling clusters will be selected in each of the four study areas and retained across survey rounds. New household and participant samples will be selected within these clusters, with up to two randomly selected eligible residents tested per household. Each area will contribute a target of 1,372 valid malaria rapid diagnostic test results per round, providing 5,488 results per round and 21,952 results across the four Phase 1 surveys. Participants with positive test results will receive anti-malarial treatment according to national guidelines.

For each Phase 1 population comparison, the primary analysis will estimate the intervention-to-control prevalence ratio at 6 months, adjusted for observed baseline malaria prevalence in the corresponding geographic sampling cluster and prespecified covariates: age, sex, shelter type, previous night insecticide-treated net use, and usual sleeping location. Modified Poisson generalized estimating equations (GEEs) will account for clustering, with small-sample corrections for statistical inference. A baseline-to-6-month ratio of prevalence ratios will provide a sensitivity analysis. The two population comparisons will be reported separately. Geographic sampling clusters are sampling units rather than independent intervention allocation units, and interpretation will account for potential residual geographic confounding.

Phase 1 entomological monitoring will include 60 fixed sentinel households in each of the four study areas, providing 240 households overall. Each household will contribute two paired indoor/outdoor baseline collection nights and one paired collection night per month for 6 months after deployment, during the main malaria transmission season. This provides 1,920 paired household-nights, equivalent to 3,840 indoor and outdoor collector nights. Human landing catches will be supplemented by household Prokopack aspiration collections of resting mosquitoes the following morning. Laboratory investigations will characterize vector species, Plasmodium infection, verterbrate blood meal sources, and molecular changes in transfluthrin response.

Indoor and outdoor female Anopheles landing counts will be analyzed separately using negative binomial mixed-effects models that account for repeated household collections and collection effort. Each analysis will combine the host-community and refugee-camp comparison strata to estimate a pooled baseline-to-post-deployment ratio of landing rate ratios. Multiplicity adjustment will be applied across the two key secondary entomological comparisons.

Progression to Phase 2 will depend on review of Phase 1 prespecified operational, implementation, safety, and monitoring criteria. Phase 2 will expand Guardian™ distribution to the remaining study populations, including former control areas and dispersed host communities, and replace units in previously treated populations. Follow-up will continue for 6 months after scale-up to universal coverage.

Phase 2 implementation monitoring will sample Yida, Ajuong Thok, Pamir, and dispersed host communities separately. Each population will contribute approximately 300 households per repeated cross-sectional survey, selected through 15 geographic clusters and 20 households per cluster. Surveys at 1-2 weeks and at 1, 3, and 6 months will provide 1,200 household observations per round and 4,800 household-round observations overall across the four populations. A nested longitudinal cohort will enroll 225 households per population, providing 900 households at enrollment. Approximately 720 households are expected to remain under observation at 6 months after allowing for 20% loss to follow-up.

Phase 2 malaria surveys will be undertaken during the pre-Phase 2 reference period and at 6 months following universal deployment. Yida and dispersed host communities will each contribute 1,372 valid rapid diagnostic test results per round. Ajuong Thok and Pamir will each contribute 686 results per round to a combined refugee-population analysis. This provides 4,116 results per round and 8,232 across the two Phase 2 surveys. Analyses will account for geographic sampling clusters, study setting, camp, and previous Phase 1 exposure, where separately identifiable. Because Phase 2 has no untreated contemporaneous control, changes will be interpreted as supportive pre/post associations.

Phase 2 entomological monitoring will include 60 fixed sentinel households in each of Yida, Ajuong Thok, Pamir, and dispersed host communities, providing 240 households overall. Two paired baseline collection nights and 6 monthly paired follow-up nights per household will provide 1,920 paired household-nights and 3,840 collector-nights. Analyses will assess changes in indoor and outdoor landing rates within participating populations following universal deployment.

Newly arrived refugees will contribute separate short-duration epidemiological and entomological assessments in each study phase. Malaria surveys at enrollment or Guardian™ receipt and approximately 3 months later will target 600 valid test results per round, providing 1,200 results per new arrival population in each phase. Approximately 320 households, sampled through 16 clusters and 20 households per cluster, will be approached at each round. Entomological monitoring will include 60 sentinel households per new arrival population. Where operationally feasible, two paired reference period collection nights will precede 3 monthly paired follow-up nights, providing 300 paired household nights and 600 collector nights per phase. These assessments will provide descriptive evidence on early changes following arrival, settlement, and deployment.

Focus groups, interviews, community feedback, field-team debriefs, and observations of distribution processes will explore acceptability, barriers to sustained intervention use, and opportunities to improve delivery across beneficiary groups. Product durability will be assessed through household observation and chemical analysis. Baseline quality control will use unused stock in each phase, followed by field unit retrievals at 3, 6, 9, and 12 months in Phase 1 and at 3 and 6 months in Phase 2. Retrieved units will be replaced one-for-one, with replacements documented separately from originally deployed units. Activity-based costing will assess the resources required for procurement, delivery, community engagement, monitoring, and maintenance of coverage.

Retrospective and prospective routine health facility data will provide additional information on malaria burden, testing, treatment, reporting completeness, and service disruptions. Population movement, seasonality, rainfall, access constraints, and other malaria interventions will be documented to support interpretation. Consent procedures will distinguish participation in research from access to humanitarian commodities and services. Human landing catch collectors will provide separate consent and receive training, malaria chemoprophylaxis, regular malaria testing, and prompt diagnosis and treatment.

Quantitative and qualitative findings will be integrated across beneficiary groups to assess effectiveness alongside implementation fidelity, safety, durability, acceptability, and cost. The evaluation will inform practical guidance on the delivery and potential scale-up of long-lasting spatial emanators in humanitarian settings.