Evidence-Based Decision-Making for Malaria Elimination: Applying the Freedom from Infection Statistical Framework in Five Malaria Eliminating Countries

Abstract

Background: Routine surveillance is a key pillar of malaria programs and is the primary source of data relied upon for decision-making, particularly in areas striving to achieve elimination. However, any inference when relying on routine malaria data to inform decision-making is limited by how effective the system is at measuring the actual malaria burden. Here, we extended the freedom from infection (FFI) model framework to produce species-specific estimates and combined multiple surveillance components including community case management (CCM) and active case detection (ACD). We then applied the FFI model in five malaria eliminating settings.

Methods: Monthly routine malaria data on both P. falciparum and P. vivax and key informant interviews on health system factors were collected from 26 facilities in Cabo Verde, 34 in Dominican Republic, 474 in Peru, 60 in the Philippines, and 921 in Vietnam. Additionally, in the Dominican Republic routine data and health systems interviews from 12 community malaria health workers (CHWs) were conducted whereas in Cabo Verde, Peru, Philippines and Vietnam 10,767 individuals were sampled during cross-sectional surveys as part of the ACD. The data were analyzed using adapted FFI models that accounted for data from the multiple malaria species and surveillance components. The primary outcomes were the sensitivity of the malaria surveillance system (SSe) and the probability of freedom from malaria (Pfree).

Findings: Having testing and treatment supplies in stock, recent training on diagnostics and case management, recent supervision, and shorter estimated travel time to visit a facility were consistently associated with a strong surveillance system across the five study settings. Overall, 841 of 1,515 and 771 of 1,455 facilities, for P. falciparum and P. vivax, respectively, had a sufficient SSe to achieve and maintain a high Pfree, consistent with having achieved malaria elimination, with either passive case detection (PCD) data alone or when combined with ACD.

Interpretation: Applying the FFI model framework to existing malaria surveillance data can provide programs with important information to support decision-making, specific for each malaria species. Where routine malaria surveillance systems are strong, this is sufficient to achieve and maintain a high Pfree (i.e., having achieved elimination is likely). Including additional surveillance components like CCM and ACD with multiple diagnostic tools can help improve estimates where routine malaria data alone are not sufficient to ensure confidence in elimination.

Funding: This study was funded by the Bill and Melinda Gates Foundation awarded to GS and CD (OPP1177272) which supported the work in Cabo Verde, Peru, Philippines, and Vietnam. Funding from The Carter Center and the Global Institute for Disease Elimination awarded to GS supported the work in the Dominican Republic.

Declaration of Interest: We declare no competing interests.

Ethical Approval: Permission to collect the routine malaria data was obtained by the ministries of health in each country and written informed consent to conduct the health system interview was obtained from each participant. Next, for participation in the ACD, written informed consent was obtained from each participant. For those under the age of majority, consent was provided by the primary caregiver with children aged between 7 and 17 years also providing assent. The study protocols were approved by: CNEPS: 80/2020, CONABIOS: 1448, Emory: STUDY00003770, UPCH: 201615, RITM: 2020-31, NIMPE: 41/QD-VSR and, LSHTM: 17927/21886/26600/19167/21697.

Keywords: Malaria surveillance, surveillance sensitivity, Bayesian analysis, community case management, seroepidemiology