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.

• Guardian™ long-lasting spatial emanators (Device), also known as Spatial repellents: Arm Description: Guardian™ (SC Johnson \& Son, Racine, WI, USA) is a long-lasting spatial emanator. The active ingredient is transfluthrin (C15H12Cl2F4O2, 2.5 g/Guardian™, EPA Reg. number 432-1588). Each unit has a mean emanation rate of 4.7 mg/day, with 2 units installed per regular size room (12 m2/30 m3; 9.4 mg/day emanation per room). This intervention is designed to remain efficacious for 12-months upon opening.

Primary outcomes

• All-age malaria infection prevalence at 6 months following Phase 1 Guardian™ deployment [6 months after Phase 1 Guardian™ deployment.]: Percentage of randomly selected participants of all ages with a positive malaria rapid diagnostic test, irrespective of symptoms, calculated as the number with positive results divided by the number with valid results, multiplied by 100. Two separate Phase 1 comparisons will be conducted: Yida intervention versus Yida control, and Ajuong Thok intervention versus Pamir control. The effect measure will be the intervention-to-control prevalence ratio, adjusted for baseline malaria prevalence in the corresponding geographic sampling cluster and prespecified participant and household covariates.

Secondary outcomes

• Indoor female Anopheles landing rate [Baseline and monthly during months 1-6 after distribution in each phase; baseline and monthly during months 1-3 for newly arriving refugees.]: Number of female Anopheles mosquitoes collected indoors by human landing catches divided by indoor collection effort, expressed as mosquitoes per person-hour. Phase 1 analyses will compare baseline-to-post-distribution changes between intervention and control populations, pooling the two population comparison strata. Phase 2 and newly arriving refugee analyses will assess changes from their respective reference periods.
• Outdoor female Anopheles mosquito landing rate [Baseline and monthly during months 1-6 after distribution in each phase; baseline and monthly during months 1-3 for newly arriving refugees.]: Number of female Anopheles mosquitoes collected outdoors by human landing catches divided by outdoor collection effort, expressed as mosquitoes per person-hour. Phase 1 analyses will compare baseline-to-post-distribution changes between intervention and control populations, pooling the two population comparison strata. Phase 2 and newly arriving refugee analyses will assess changes from their respective reference periods.

Other outcomes

• Percentage of targeted households receiving Guardian™ within the planned distribution window [At completion of each distribution round, up to 2 weeks after the start of distribution, in Phases 1 and 2]: Number of targeted households documented as receiving Guardian™ within the prespecified distribution window divided by the total number of households targeted for that distribution, multiplied by 100. Receipt dates will be obtained from distribution registers and verified through household monitoring. Results will be reported by beneficiary group and implementation phase.
• Percentage of targeted households receiving at least one Guardian™ unit [Phase 1: 1-2 weeks and 1, 3, 6, 9 and 12 months after distribution. Phase 2: 1-2 weeks and 1, 3 and 6 months. New arrivals: 1-2 weeks and 1 and 3 months in each phase.]: Percentage of sampled households in targeted intervention populations reporting receipt of at least one Guardian™ unit. Estimates will be obtained through repeated cross-sectional household surveys, supported by distribution records and direct observation where possible. Results will be reported separately for established camp residents, host communities and newly arriving refugees.
• Percentage of recipient households receiving an adequate number of Guardian™ units for shelter volume [Phase 1: 1-2 weeks and 1, 3, 6, 9 and 12 months after distribution. Phase 2: 1-2 weeks and 1, 3 and 6 months. New arrivals: 1-2 weeks and 1 and 3 months in each phase.]: Number of recipient households receiving at least the number of Guardian™ units required for their measured shelter volume divided by the number of recipient households assessed, multiplied by 100. Required unit numbers will follow the manufacturer's dosing instructions, based on one unit per less than 15 cubic metres of shelter volume.
• Percentage of household respondents correctly identifying Guardian™'s purpose and key use instructions [Phase 1: 1-2 weeks and 1, 3, 6, 9 and 12 months after distribution. Phase 2: 1-2 weeks and 1, 3 and 6 months. New arrivals: 1-2 weeks and 1 and 3 months in each phase.]: Percentage of interviewed adult household respondents who correctly identify Guardian™'s intended purpose of reducing mosquito exposure and describe at least two key use instructions. Instructions include unpacking/activation, appropriate placement, adequate unit numbers for shelter volume, continuous use, and safe handling or reporting procedures.
• Percentage of recipient households with correctly positioned Guardian™ units [Phase 1: 1-2 weeks and 1, 3, 6, 9 and 12 months after distribution. Phase 2: 1-2 weeks and 1, 3 and 6 months. New arrivals: 1-2 weeks and 1 and 3 months in each phase.]: Percentage of assessed recipient households whose deployed Guardian™ units meet the manufacturer's placement instructions, determined through direct shelter observation using a standardized checklist. Placement will be assessed separately from whether the household received sufficient units for its shelter volume.
• Percentage of originally issued Guardian™ units retained and in use [Phase 1: 1-2 weeks and 1, 3, 6, 9 and 12 months after receipt. Phase 2: 1-2 weeks and 1, 3 and 6 months. New arrivals: 1-2 weeks and 1 and 3 months in each phase.]: Number of originally issued Guardian™ units verified as remaining in use in the monitored household divided by the number originally issued to that household, multiplied by 100. Unit identifiers and repeated household observations will distinguish original units from replacements. Loss, transfer, removal, disposal and replacement will be recorded separately.
• Percentage of household respondents reporting a perceived benefit from Guardian™ use [Phase 1: 1-2 weeks and 1, 3, 6, 9 and 12 months after receipt. Phase 2: 1-2 weeks and 1, 3 and 6 months. New arrivals: 1-2 weeks and 1 and 3 months in each phase.]: Percentage of interviewed household respondents reporting that Guardian™ use has provided a perceived benefit, such as reduced mosquito presence, biting or nuisance. Structured survey responses will be complemented by interviews and focus group discussions exploring the nature of perceived benefits.
• Percentage of observed Guardian™ units remaining physically intact [Phase 1: 3, 6, 9 and 12 months after deployment. Phase 2: 3 and 6 months after deployment.]: Number of Guardian™ units observed to be physically intact divided by the number of units inspected, multiplied by 100. Field teams will document visible damage and the type of damage using a standardized observation checklist. Findings will be linked to product age, placement location, shelter characteristics and exposure conditions.
• Residual transfluthrin content in retrieved Guardian™ units [Phase 1: unused unit baseline and 3, 6, 9 and 12 months after deployment. Phase 2: unused unit baseline and 3 and 6 months after deployment.]: Transfluthrin content in sampled Guardian™ units, measured by gas chromatography and expressed as mass of transfluthrin per unit. Unused units will provide baseline quality control measurements. Retrieved field units will be assessed by duration of deployment, beneficiary group and relevant exposure conditions. Residual content will also be expressed relative to baseline content.
• Implementation cost per Guardian™ unit delivered [From procurement through 12 months after Phase 1 distribution and 6 months after Phase 2 distribution.]: Total humanitarian implementation cost divided by the number of Guardian™ units delivered to households, expressed in US dollars per unit. Costs will include procurement, shipment, customs clearance, storage, transport, distribution, training, community engagement, supervision and routine implementation monitoring. Research-only costs will be identified separately. Estimates will be reported by implementation phase, beneficiary group and distribution modality.
• Implementation cost per household receiving Guardian™ [From procurement through 12 months after Phase 1 distribution and 6 months after Phase 2 distribution.]: Total humanitarian implementation cost divided by the number of households receiving at least one Guardian™ unit, expressed in US dollars per household reached. Research-only costs will be identified separately. Estimates will be reported by implementation phase, beneficiary group and distribution modality.
• Implementation cost per person living in a household with adequate Guardian™ coverage [From procurement through 12 months after Phase 1 distribution and 6 months after Phase 2 distribution.]: Total humanitarian implementation cost divided by the estimated number of people living in households receiving adequate numbers of Guardian™ units for their shelter volume, expressed in US dollars per person. This operational coverage measure does not assume that every covered individual is protected from malaria infection. Estimates will be reported by implementation phase and beneficiary group.
• Implementation cost per correctly protected person-month [From household receipt through 12 months in Phase 1, 6 months in Phase 2, and 3 months for newly arriving refugees in each phase.]: Total humanitarian implementation cost divided by cumulative person-months in households with adequate Guardian™ dosing for shelter volume, correct placement, and retained units in use. Person-months will combine household occupancy with observed implementation status over time. Interpolation between monitoring visits and alternative assumptions will be examined in sensitivity analyses. Results will be expressed in US dollars per correctly protected person-month.
• Incremental implementation cost per malaria clinical case averted [Effectiveness and cost observations through 12 months after Phase 1 distribution and 6 months after Phase 2 distribution.]: Incremental humanitarian implementation cost divided by the estimated number of malaria clinical cases averted, expressed in US dollars per case averted. This exploratory analysis will be undertaken only where the effectiveness data support a defensible estimate of clinical cases averted. Assumptions and uncertainty in costs, coverage and effectiveness will be examined through sensitivity analyses.

Sex: All, accepts healthy volunteers

Inclusion Criteria:

* Residents of the selected study populations in Yida, Ajuong Thok and Pamir, newly arrived refugee populations included in either phase, and dispersed host communities included during Phase 2.
* For malaria prevalence surveys: usual residents of sampled households who are randomly selected for testing, irrespective of age, sex, malaria symptoms or previous Guardian™ exposure.
* Adults aged 18 years or older who provide informed consent. For children, a parent or legal guardian must provide consent, with age-appropriate assent for children aged 7-17 years.
* For household implementation monitoring: households eligible for Guardian™ distribution within the selected intervention populations, with an adult representative available to provide informed consent.
* For household entomological monitoring: selected sentinel households whose adult representative provides consent for collection activities.
* For human landing catches: trained adult household occupants who provide separate written informed consent and agree to the collector safety procedures, including malaria chemoprophylaxis and regular malaria testing.

Exclusion Criteria:

* Individuals or households outside the sampling frame for the relevant evaluation component.
* Individuals who decline participation, or children for whom parent or guardian consent or required assent is not obtained.
* For household monitoring: households without an adult representative available to provide informed consent.
* Households that cannot be accessed safely.
* Individuals or households for whom participation would create an unacceptable safety, protection or confidentiality risk.

Participation in research evaluation activities is voluntary and does not affect access to Guardian™, health care or any other humanitarian assistance.

• Chen I, Miller SL, Msellemu D, Lugenge AG, Swai JK, Achee N, Andres M, Bibbs CS, Chareonviriyaphap T, Charlwood JD, Devine G, Elman N, Fillinger U, Flores-Mendoza C, Gibson S, Govella N, Gowelo S, Horstmann S, Kawada H, Kline D, Lloyd A, Lobo NF, Maia MF, Mmbando A, Moreno-Gomez M, Morrison AC, Mponzi W, Mwanga EP, Njoroge M, Ogoma SB, Okumu FO, Opiyo M, Oumbouke WA, Paliga J, Pongsiri A, Ponlawat A, Saeaung M, Salazar F, Sangoro O, Stevenson JC, Sukkanon C, Syafruddin D, Tambwe MM, Tangena JA, Vajda EA, Vazquez-Prokopec G, Wagman JM, Yan C, Allen IE, Moore SJ. Volatile pyrethroid spatial repellents for preventing mosquito bites: a systematic review and meta-analysis. EBioMedicine. 2025 Sep;119:105891. doi: 10.1016/j.ebiom.2025.105891. Epub 2025 Aug 26. PMID 40866281
• Morrison AC, Reiner RC Jr, Elson WH, Astete H, Guevara C, Del Aguila C, Bazan I, Siles C, Barrera P, Kawiecki AB, Barker CM, Vasquez GM, Escobedo-Vargas K, Flores-Mendoza C, Huaman AA, Leguia M, Silva ME, Jenkins SA, Campbell WR, Abente EJ, Hontz RD, Paz-Soldan VA, Grieco JP, Lobo NF, Scott TW, Achee NL. Efficacy of a spatial repellent for control of Aedes-borne virus transmission: A cluster-randomized trial in Iquitos, Peru. Proc Natl Acad Sci U S A. 2022 Jun 28;119(26):e2118283119. doi: 10.1073/pnas.2118283119. Epub 2022 Jun 23. PMID 35737833
• Syafruddin D, Asih PBS, Rozi IE, Permana DH, Nur Hidayati AP, Syahrani L, Zubaidah S, Sidik D, Bangs MJ, Bogh C, Liu F, Eugenio EC, Hendrickson J, Burton T, Baird JK, Collins F, Grieco JP, Lobo NF, Achee NL. Efficacy of a Spatial Repellent for Control of Malaria in Indonesia: A Cluster-Randomized Controlled Trial. Am J Trop Med Hyg. 2020 Jul;103(1):344-358. doi: 10.4269/ajtmh.19-0554. Epub 2020 May 14. PMID 32431275
• Ochomo EO, Gimnig JE, Awori Q, Abong'o B, Oria P, Ashitiba NK, Polo B, Moshi V, Otanga H, Adung'o F, Ouma EA, Outa S, Ramaita E, Levine R, Odongo W, Harvey SA, Monroe A, Hudson A, Sandberg B, Hendrickson J, Zhao X, Zhou R, Liu F, Achee NL, Grieco JP. Effect of a spatial repellent on malaria incidence in an area of western Kenya characterised by high malaria transmission, insecticide resistance, and universal coverage of insecticide treated nets (part of the AEGIS Consortium): a cluster-randomised, controlled trial. Lancet. 2025 Jan 11;405(10473):147-156. doi: 10.1016/S0140-6736(24)02253-0. Epub 2024 Dec 19. PMID 39709979
• Allan R, Scherrer R, Kasap OE, Paris L, Scott T, Sauskojus H, Wetherill O, Estecha-Querol S, Alkhalaf Z, Karakus M, Yilmaz A, Alten B, Messenger LA. Enhancing protection against vector-borne diseases in forcibly displaced communities: evaluating the efficacy of spatial repellents for cutaneous leishmaniasis control in North-East Syria. BMC Med. 2025 Jul 3;23(1):401. doi: 10.1186/s12916-025-04244-2. PMID 40611204
• Allan RJ, Scherrer R, Estecha-Querol S, Weetman D, Paris L, Ba'abba Goni U, Idris Abdulhamid F, Nfornuh Alenwi B, Ahmad S, Cross CL, Ashir GM, Waziri M, Ntadom G, Messenger LA. The effectiveness of long-lasting spatial repellent emanators against malaria in humanitarian crisis settings in northern Nigeria: a two-arm pragmatic, open-label, controlled trial. Lancet Infect Dis. 2026 May;26(5):464-474. doi: 10.1016/S1473-3099(25)00684-X. Epub 2026 Jan 8. PMID 41520670
• Messenger LA, Furnival-Adams J, Chan K, Pelloquin B, Paris L, Rowland M. Vector control for malaria prevention during humanitarian emergencies: a systematic review and meta-analysis. Lancet Glob Health. 2023 Apr;11(4):e534-e545. doi: 10.1016/S2214-109X(23)00044-X. PMID 36925174

• Richard J Allan, PhD, OBE, Principal investigator, The Mentor Initiative