Supporting evidence-based decisions about the geographic and demographic extensions of seasonal malaria chemoprevention in Benin: A modelling study

Abstract

Seasonal malaria chemoprevention (SMC) has been implemented yearly in northern Benin since 2019 to reduce the malaria burden in children under 5 years of age. Its geographic scope was progressively extended until in 2022 two different extensions of SMC were considered: either demographic - children aged 5-10 in the currently targeted departments would also receive SMC, or geographic to children under 5 in new eligible departments to the south. As SMC had neither been implemented in the areas nor in the age groups suggested for expansion, modelling was used to compare the likely impact of both extensions. The model OpenMalaria was calibrated to represent the history of malaria interventions and transmission risk in administrative units of Benin. Currently planned future interventions and two scenarios for SMC extensions were simulated to inform where impact would be the highest. The model predicted that between 2024 and 2026 the geographic extension of SMC would avert at least four times more severe malaria cases and five times more direct malaria deaths per targeted child than the demographic extension. Indeed, most severe cases are concentrated in children under 5 in all departments of interest, as malaria burden remains high in this region. Numbers of severe cases averted per targeted child were similar between health zones eligible for geographic extension. The main limitations of this work are global model parameters due to lack of country-specific data on efficacy of interventions or development of immunity. SMC coverage was assumed to be uniform across rounds, zones, and age groups. Due to the high malaria burden in northern and central Benin, the geographic extension would be more impactful than the demographic extension both in absolute number of severe cases averted and per child protected, and has started to be implemented in 2024. Health zones were prioritised by availability of community health workers to deliver SMC. Mathematical modelling was a supportive tool to understand the relative impact of the different proposed SMC extensions and contributed to the decision-making process. Its integration significantly enhanced the utilisation of data for decision-making purposes. Rather than being used for forecasting, the model provided qualitative guidance that complemented other types of evidence.

Fig 1. Applied workflow to answer programmatic questions with country-specific mathematical modelling.

Schematics are for illustrative purpose only. Base layer of Benin map from https://data.humdata.org/dataset/cod-ab-ben.

Fig 2. Future scenarios for the period 2024 to 2026.

Fig 3. Comparison of observed, estimated and modelled malaria prevalence in children under 5.

(A) Estimated (green, Malaria Atlas Project [49]) and modelled (grey) yearly malaria prevalence in children under 5 aggregated at the national level. (B) Modelled malaria prevalence in children under 5 against prevalence measured during household surveys in each department [38, 39], matched by survey period. The 2011 and 2017 surveys [38] were used to produce the MAP prevalence estimtes. Ribbon (A) and error bars (B) represent uncertainty on intensity of transmission as well as model stochasticity, as described in the Methods section.

Fig 4. Trends in absolute numbers (A) and incidence (B) of modelled case indicators.

(A) Absolute numbers of modelled malaria episodes and treated cases in all ages, as well as cases reported by Benin and estimates from the World Malaria Report (WMR) 2024 [1]. (B) Incidence of modelled malaria episodes per 1000 persons at risk in all age groups and children under 5. Ribbons represent uncertainty on intensity of transmission as well as model stochasticity, as described in the Methods section.

Fig 5. Cases by scenario and age group between 2024 and 2026. Distribution of mean predicted malaria episodes (A) and severe cases (B) across age groups with planned interventions and adding SMC extensions in eligible zones. Bembereke-Sinende health zone has been excluded from Borgou since it is not eligible to the SMC extensions.

Fig 6. Cases averted per additionally targeted children between 2024 and 2026.

(A) All malaria episodes in all age groups averted by each SMC extension in eligible zones per thousand additionally targeted children (from 5 to 10 for demographic extension and under 5 for geographic extension). (B) Severe cases averted per thousand additionally targeted children; C: Malaria deaths averted per million additionally targeted children. Error bars represent uncertainty on intensity of transmission as well as model stochasticity, as described in the Methods section.

Fig 7. All malaria episodes and severe malaria cases averted by the geographical extension of SMC per 1000 children under 5 by health zone between 2024 and 2026.

Each acronym stands for one health zone (made of one or several communes): NKP: Nikki-Kalale-Perere; PN: Parakou-N’Dali; TCH: Tchaourou; DAGLA: Dassa-Zoume-Glazoue; SABA: Savalou-Bante; SAO: Save-Ouesse; BLA: Bassila; DCO: Djougou-Copargo-Ouake. Error bars represent uncertainty on intensity of transmission as well as model stochasticity, as described in the Methods section.