FionaHuini/malaria_kenya_temporal

Spatial analysis of malaria risk change across Kenya counties 2010,2015 & 2019 # Spatial Analysis of Malaria Risk Change in Kenya (2010, 2015, 2019) ## Overview This project analyses spatial patterns of malaria transmission intensity across Kenya's 47 counties at three time points (2010, 2015, 2019) using PfPR (Plasmodium falciparum Parasite Rate) data. The analysis examines whether malaria risk is spatially clustered, identifies persistent hotspot counties, and tracks how the spatial pattern changed over time. ## Methods - **Global Moran's I**: tests for spatial autocorrelation in county-level PfPR across all three time points - **LISA (Local Indicators of Spatial Association)**: identifies specific High-High and Low-Low cluster counties - **Temporal change analysis**: arithmetic difference in PfPR between time points to identify improving and worsening counties - **Choropleth maps**: visualise geographic distribution for each year - **Interactive map**: county-level PfPR with popup details using leaflet ## Key Findings - Statistically significant spatial clustering confirmed in all three years (Moran's I: 0.563 in 2010, 0.564 in 2015, 0.661 in 2019) - Spatial clustering strengthened over time; malaria risk is becoming increasingly polarised between western and eastern Kenya - 7 persistent High-High cluster counties identified, all in the Lake Victoria basin: Bungoma, Busia, Homa Bay, Kakamega, Kisumu, Siaya, Vihiga - Malaria burden improved between 2010 and 2015, then rebounded by 2019 across most high-burden counties ## Public Health Relevance Spatial clustering of malaria risk indicates that transmission is not uniform across Kenya; seven Lake Victoria basin counties consistently bear a disproportionate burden. The strengthening of spatial clustering over time, combined with the post-2015 resurgence, suggests that current intervention strategies may be insufficient to sustain earlier gains. These findings directly support subnational tailoring of malaria interventions, with priority resource allocation to the identified High-High cl …