Freedom from infection: enhancing decision-making for malaria elimination

Assessing elimination of malaria locally requires a surveillance system with high sensitivity and specificity to detect its presence without ambiguity. Currently, the WHO standard criteria of observing the absence of locally acquired cases for 3 consecutive years, combined with a health systems assessment, are used to justify claims of malaria elimination. However, relying on a qualitative framework to support the application of this guideline can lead to early, over-optimistic relaxation of control measures with the potential for resurgence. Overcoming this challenge requires innovative approaches to model the coupled processes of malaria transmission and its clinical observation. We propose a novel statistical framework based on a state-space model to probabilistically demonstrate the absence of malaria, using routinely collected health system data (which is extensive but inherently imperfect). By simultaneously modelling the expected malaria burden within the population and the probability of detection, we provide a robust estimate of the surveillance system's sensitivity and the corresponding probability of local elimination (probability of freedom from infection). Our study reveals a critical limitation of the traditional criterion for declaring malaria elimination, highlighting its inherent bias and potential for misinterpreting ongoing transmission. Such oversight not only misrepresents ongoing transmission but also places communities at risk for larger outbreaks. However, we demonstrate that our integrated approach to data comprehensively addresses this issue, effectively detecting ongoing transmission patterns, even when local reports might suggest otherwise. Our integrated framework has far-reaching implications for malaria control but also for infectious disease control in general. Our approach addresses the limitations of traditional criteria for declaring freedom from disease and opens the path to true optimisation of the allocation of limited resources. Our findings emphasise the urgent need to reassess existing methods to accurately confirm malaria elimination, and the importance of using comprehensive modelling techniques to continually monitor and maintain the effectiveness of current surveillance systems, enabling decisions grounded in quantitative evidence.