Changing malaria infection prevalence in Ghana 2011-2025: a multi-diagnostic spatiotemporal analysis

Abstract Background Malaria infection prevalence in Ghana has seen remarkable decline due to upscale of intervention since 2011. While subnational variations are notable, differences in diagnostic results and its implications on policy and resource allocation have not been explored. In this study we predict and characterize differences in malaria prevalence for Ghana, stratify malaria risk for both methods at district level and discuss its implication on policy. Methods We obtained data from major population-based surveys conducted in Ghana between 2011 and 2023 and extracted the number of children tested for malaria by microscopy and RDT, number tested positive and cluster location coordinates. We fitted a spatio-temporal Bayesian model that incorporated multiple likelihoods. Prediction maps were generated for observed survey years, interpolated for unobserved years in between surveys and extrapolated for 2024 and 2025. We computed the exceedance probability (EP) of prevalence greater than 10% for 2011 and 2025, stratified the 2025 prevalence into categories and assessed its implications on subnational tailoring of interventions. Results Malaria prevalence, measured using either diagnostic, has reduced steadily from a national average of 43% in 2011 to 6.9% in 2025 when measured by RDT. Metrics of urbanicity were the strongest predictors of decreased malaria prevalence. While overall trends in prevalence were similar across both diagnostic methods, there was a 84% decrease in RDT prevalence compared with 81% for microscopy. Subnational stratification by risk category varied cubistically depending on the diagnostic method used: microscopy prevalence classified 92% (240/261) of districts as low (< 10%) compared with just 60% (156/261) for RDT. Conclusion We have described the changing endemicity of malaria in Ghana over fifteen years of intensive control efforts. While patterns of decline were similar regardless of diagnostic method used, stratification by prevalence threshold yielded very different classifications of risk. There is the need for further discussion by National Malaria Programmes and funding partners to discuss standardized measures for risk in the era of dwindling funding for malaria control and elimination.