Reducing Malaria Transmission in Forest-going Mobile and Migrant Populations in Myanmar

Over the last decade, the burden of malaria has fallen dramatically in the Greater Mekong Subregion, with deaths falling by 95% and cases by 76% between 2012 and 2019.The declining malaria burden in GMS has largely been attributed to the deployment of interventions of malaria prevention tools such as long-lasting Insecticidal Nets (LLIN), and widespread availability of rapid diagnostic tests and artemisinin combination therapies. However, residual transmission exists among high risk populations, particularly mobile and migrant populations who enter forests for work. New interventions targetting these high risk groups are needed if countries of the Greater Mekong Subregion are to achieve malaria elimination by 2030. The study will be conducted in malaria endemic areas of Myanmar which have large forest going MMP populations. The aim of the study is to assess the effectiveness, acceptability, feasibility and cost-effectiveness of a personal protection package to reduce malaria transmission among mobile and migrant populations (MMPs) and the general population in their residing villages in Myanmar. The personal protection package includes WHO prequalified long-lasting insecticidal net (LLIN), insect repellent (N, N-Diethyl-meta-toluamide (DEET), insecticide-treated clothes (ITC), and a MMP-tailored behavioural change communication (BCC) package. The study design is a stepped-wedge cluster-randomized controlled trial with nested mixed methods study. The stepped-wedge cluster randomized trial will estimate the effectiveness of a personal protection package provided to forest-going MMPs (the intervention) delivered by village malaria volunteers at the site (village/worksite) level on reducing Plasmodium spp. infections (primary outcome). Whilst the personal protection package will be provided to mobile and migrant populations in each village in a step-wise manner, primary and secondary outcomes relating to malaria testing will be collected in all consenting individuals in the village who present for malaria rapid diagnostic tests, whether they are mobile and migrant people or not. The study will also include investigation of antibody responses to Plasmodium spp. and mosquito vectors and molecular markers of antimalarial drug resistance.

The study will be implemented between July 2021 (M1) to June 2022 (M12). The personal protection package for MMPs will be implemented sequentially in a minimum of 100 villages serviced by an ICMV. The villages will be grouped into 11 blocks of 9 villages, with each block transitioned from control phase (before introduction of the personal protection package) to intervention phase (after introduction of the personal protection package) at monthly intervals in random order. This follows an initial one-month baseline period at the start of the study period where all villages will not be exposed to the personal protection package exclusively.

For this proposed stepped-wedge cluster randomized controlled study, it is estimated that approximately 30 RDTs per month will be tested in each study site (village/worksite) and there will be 36,000 MMP tests over 12 months in 100 villages/worksites. It will be undertaken yielding a relative minimum detectable difference of 40% in odds of RDT-detectable malaria infection (assuming a village intraclass correlation [ICC]= 0.42 [6]; 5% significance; 80% power and 1% RDT malaria prevalence).

For the stepped-wedged cluster randomized trial, both descriptive and primary outcome trial analyses will be performed. In both analyses, sampling weights will be derived and applied in analyses where village selection probabilities are not equal - possible if the achieved number of villages sampled across townships varies. Furthermore, cluster robust standard errors will be used in all descriptive analyses given the complex sampling design. Differences in prevalence of malaria infections will be estimated across intervention and control periods using generalized linear mixed modelling (e.g. logit link function and binomial distribution).