Spatio-temporal heterogeneity of urban malaria in Libreville, Gabon (2012–2023): neighborhood-level hotspot analysis for micro-stratified control

Abstract Background In Libreville, Gabon, the Melen sentinel site has monitored malaria among febrile children for over a decade, offering a lens on intra-urban transmission. We aimed to characterize neighborhood-level spatial and temporal dynamics (2012–2023) and identify high-risk areas for micro-stratified control. Methods Geographic and parasitological data from 9,507 children (Jan 2012-Dec 2023) were analyzed. Diagnosis used standard microscopy (thick/thin smears). Plasmodium falciparum prevalence (PfPR) was calculated for 44 neighborhoods across four periods. Neighborhoods were classified by adapted WHO endemicity thresholds and grouped into socio-spatial categories (urban, semi-urban, shanty-town, semi-rural). Global spatial autocorrelation was tested with Moran’s I, then local Getis-Ord Gi* identified hotspots and coldspots. Results By category, PfPR in 2023 was highest in semi-rural areas (50.9%, 95% CI: 47.9–54.0) and semi-urban areas (46.5%, 95% CI: 43.8–49.3), followed by shanty-town areas (44.6%, 95% CI: 42.6–46.6) and urban areas (38.0%, 95% CI: 33.1–43.2). Substantial within-category heterogeneity was observed, with neighbourhoods within the same socio-spatial class exhibiting markedly different PfPR during the same period. The WHO classification showed a shift from mostly mesoendemic neighbourhoods in 2015 to 41.5% hyperendemic neighbourhoods in 2023. The Gi* hotspot analysis (k = 4) identified significant clusters that changed over time: hotspots were detected in semi-urban and semi-rural neighbourhoods in 2012–2013 and 2015–2016, shifted towards semi-rural neighbourhoods in 2018–2019 (with coldspots emerging in shanty-town settings), and re-emerged in both semi-urban and semi-rural strata in 2023, with a coldspot detected in a shanty-town neighbourhood. Conclusions Malaria transmission in Libreville is highly heterogeneous and dynamic, with predominant mesoendemic neighborhoods and recurrent hotspots. Findings support micro-stratified strategies prioritizing recurrent clusters and integrating fine-scale spatial analytics into national surveillance.