Extension of efficacy range for targeted malaria-elimination interventions due to spillover effects

Global efforts towards malaria eradication have plateaued despite renewed attention. Southern Africa, particularly the Elimination Eight Initiative, aims for malaria elimination by 2030. However, the annual number of malaria cases has slightly increased, and mortality rates remain nearly unchanged. The ideal intervention would not only prevent disease in recipients but also prevent onward transmission to nearby individuals through "spillover effects." Previous research has shown spillover effects for various interventions, including mass drug administration, insecticide-treated bed nets, and indoor residual spraying. However, these studies had limitations, such as lack of randomization or focus on high-transmission settings. This study addresses the need to estimate spillover effects in low-transmission settings approaching elimination using data from a cluster-randomized trial in Namibia that tested reactive, focal chemoprevention and vector control interventions.

Existing literature demonstrates varying degrees of spillover effects from malaria interventions. Studies from the 1990s estimated spillover effects from community-wide insecticide-treated bed net distributions, but these might differ due to changes in coverage, insecticide resistance, and mosquito behavior. A more recent study of indoor residual spraying lacked randomization and might suffer from confounding. No previous studies have quantified spillover effects of interventions designed for low-transmission settings approaching elimination. Accounting for spillover effects is crucial because it can substantially improve the cost-effectiveness of interventions, a key factor considering the projected increased cost of elimination and eradication efforts compared to existing control programs.

This study re-analyzed data from a previously reported cluster-randomized trial in the Zambezi region of Namibia (2017). The trial used a two-by-two factorial design, randomizing 56 clusters to four arms: (1) reactive case detection (RACD), (2) reactive focal mass drug administration (rfMDA) only, (3) RACD and reactive vector control (RAVC), and (4) rfMDA and RAVC. Interventions were delivered within 500 m of confirmed malaria cases. The re-analysis restructured the data into analytic cohorts mimicking a ring trial design to measure spillover effects. Each cohort included individuals within 1 km of index cases, with different follow-up periods for direct and spillover effects. Malaria incidence was the primary outcome, while prevalence and seroprevalence were secondary outcomes. Hierarchical targeted maximum likelihood estimation (TMLE), a doubly robust method, was used to estimate effects, adjusting for potential confounders using ensemble machine learning. Sensitivity analyses were conducted to assess robustness. Cost-effectiveness was evaluated using incremental cost-effectiveness ratios (ICERs).

The re-analysis revealed limited evidence of direct effects of any intervention on malaria incidence among recipients within 500 m. However, there was strong evidence of spillover effects from the combined chemoprevention and vector control intervention. Specifically, within 1 km of index cases, the combined intervention reduced malaria incidence by 43% (95% CI, 21–58%). Within 3 km, it reduced infection prevalence by 79% (6–95%) and seroprevalence by 34% (20–45%). Spillover effects were more pronounced in areas with lower pre-trial malaria incidence and favorable weather conditions for mosquito breeding. Analyses of prevalence showed stronger spillover effects for chemoprevention (72% reduction) and the combined intervention (79% reduction) among non-recipients within 3 km of interventions. Subgroup analyses indicated stronger effects closer to interventions. Cost-effectiveness analyses showed that accounting for spillover effects increased the cost-effectiveness of the combined intervention by 42%. There was no evidence of contamination between clusters.

This study demonstrates that a combined reactive, focal chemoprevention and vector control intervention can have substantial spillover effects, extending the reach of its impact beyond the targeted areas. The synergistic effect of reducing parasite reservoirs in both humans and mosquitoes likely explains the observed spillover. The stronger spillover effects observed in low-transmission areas suggest that this strategy may be particularly effective in settings approaching malaria elimination. The findings highlight the importance of considering spillover effects when evaluating and planning malaria interventions, particularly in the context of elimination efforts. The observed gender disparity in spillover effects, particularly for chemoprevention, requires further investigation to explore potential behavioral factors influencing malaria risk. The study highlights the potential for interventions to prevent transmission to others. Although the study suggests there may be a delay between interventions and the effect on incidence, this is consistent with the time required for mosquito incubation.

This re-analysis provides strong evidence for the significant spillover effects of a combined reactive, focal chemoprevention and vector control intervention for malaria. The increased cost-effectiveness when accounting for these spillover effects underscores the importance of considering indirect benefits in intervention design and resource allocation. Future research should investigate the optimal intervention timing and strategies to maximize spillover effects in various transmission settings and populations.

The study's limited sample size, particularly for direct effects analyses, could have reduced statistical power and increased the risk of type II error. The study was conducted in a low-transmission region, so findings may not be generalizable to all settings. Potential outcome dependence due to overlapping cohorts was addressed through analytical adjustments but remains a potential concern.