Presentation Information
[ctkh3Px5-04]Spatiotemporal analysis for Evaluating the Impact of Climate Change on Malaria among Under-5 Children in Nigeria
*Kelechi Genevieve Ukeoma1, Saori KASHIMA1 (1. Environmental Health Science Laboratory, Graduate School of Advanced Sciences and Engineering, Hiroshima University. (Japan))
Keywords:
Climate change,Childhood malaria,Bayesian spatiotemporal modeling,Nigeria,INLA-SPDE
Climate change is expected to influence malaria transmission through temperature, rainfall, and humidity, yet few studies apply rigorous spatiotemporal methods that account for socioeconomic factors and spatial dependence. This study quantified the associations between climate variables and malaria prevalence among children under five years in Nigeria, while identifying spatiotemporal trends and high-risk areas.
We analyzed three waves of Nigeria Demographic and Health Surveys (2010, 2015, 2021), including1,114 cluster-years and approximately 45,000 children tested for malaria by rapid diagnostic test. Climate data were gotten from CHIRPS (rainfall) and ERA5 (temperature, humidity) for the May–October season, and from DHS spatial data for each survey year. Bayesian hierarchical models were fitted using the INLA-SPDE framework, incorporating spatial random effects, AR(1) temporal correlation, and second-order random walk (RW2) smooths for non-linear climate effects.
Models were adjusted for socioeconomic covariates including household wealth, maternal education, insecticide-treated net (ITN) use, and urban/rural residence. LISA was used to identify malaria hotspots. Malaria prevalence declined from 42.3% (2010) to 38.9% (2015) to 31.8% (2021). Spatial hotspots intensified 5.6-fold (13 to 73 clusters), concentrating in North-West and North-Central zones. In the fully adjusted model, Vegetation Index showed the strongest association with malaria (OR = 1.59, 95% CrI: 1.37–1.83), followed by temperature (OR = 1.33, 95% CrI: 1.00–1.78). Humidity showed a threshold effect, with protective effects below 65% and increased risk above. Household wealth was the strongest protective factor (OR = 0.55, 95% CrI: 0.47–0.64), followed by maternal education (OR = 0.68, 95% CrI: 0.58–0.80). The spatial correlation range was approximately 214 km, with moderate temporal persistence (AR(1) ρ = 0.58).
While malaria prevalence in Nigeria has declined over the past decade, it is increasingly concentrated geographically. Climate variables remain significant drivers after socioeconomic adjustment, underscoring the need for climate-responsive, geographically targeted control strategies.
We analyzed three waves of Nigeria Demographic and Health Surveys (2010, 2015, 2021), including1,114 cluster-years and approximately 45,000 children tested for malaria by rapid diagnostic test. Climate data were gotten from CHIRPS (rainfall) and ERA5 (temperature, humidity) for the May–October season, and from DHS spatial data for each survey year. Bayesian hierarchical models were fitted using the INLA-SPDE framework, incorporating spatial random effects, AR(1) temporal correlation, and second-order random walk (RW2) smooths for non-linear climate effects.
Models were adjusted for socioeconomic covariates including household wealth, maternal education, insecticide-treated net (ITN) use, and urban/rural residence. LISA was used to identify malaria hotspots. Malaria prevalence declined from 42.3% (2010) to 38.9% (2015) to 31.8% (2021). Spatial hotspots intensified 5.6-fold (13 to 73 clusters), concentrating in North-West and North-Central zones. In the fully adjusted model, Vegetation Index showed the strongest association with malaria (OR = 1.59, 95% CrI: 1.37–1.83), followed by temperature (OR = 1.33, 95% CrI: 1.00–1.78). Humidity showed a threshold effect, with protective effects below 65% and increased risk above. Household wealth was the strongest protective factor (OR = 0.55, 95% CrI: 0.47–0.64), followed by maternal education (OR = 0.68, 95% CrI: 0.58–0.80). The spatial correlation range was approximately 214 km, with moderate temporal persistence (AR(1) ρ = 0.58).
While malaria prevalence in Nigeria has declined over the past decade, it is increasingly concentrated geographically. Climate variables remain significant drivers after socioeconomic adjustment, underscoring the need for climate-responsive, geographically targeted control strategies.
Comment
To browse or post comments, you must log in.Log in