Cargando…

Bayesian variable selection in modelling geographical heterogeneity in malaria transmission from sparse data: an application to Nouna Health and Demographic Surveillance System (HDSS) data, Burkina Faso

BACKGROUND: Quantification of malaria heterogeneity is very challenging, partly because of the underlying characteristics of mosquitoes and also because malaria is an environmentally driven disease. Furthermore, in order to assess the spatial and seasonal variability in malaria transmission, vector...

Descripción completa

Detalles Bibliográficos
Autores principales: Diboulo, Eric, Sié, Ali, Diadier, Diallo A, Voules, Dimitrios A Karagiannis, Yé, Yazoume, Vounatsou, Penelope
Formato: Online Artículo Texto
Lenguaje:English
Publicado: BioMed Central 2015
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4365550/
https://www.ncbi.nlm.nih.gov/pubmed/25888970
http://dx.doi.org/10.1186/s13071-015-0679-7
Descripción
Sumario:BACKGROUND: Quantification of malaria heterogeneity is very challenging, partly because of the underlying characteristics of mosquitoes and also because malaria is an environmentally driven disease. Furthermore, in order to assess the spatial and seasonal variability in malaria transmission, vector data need to be collected repeatedly over time (at fixed geographical locations). Measurements collected at locations close to each other and over time tend to be correlated because of common exposures such as environmental or climatic conditions. Non- spatial statistical methods, when applied to analyze such data, may lead to biased estimates. We developed rigorous methods for analyzing sparse and spatially correlated data. We applied Bayesian variable selection to identify the most important predictors as well as the elapsing time between climate suitability and changes in entomological indices. METHODS: Bayesian geostatistical zero-inflated binomial and negative binomial models including harmonic seasonal terms, temporal trends and climatic remotely sensed proxies were applied to assess spatio-temporal variation of sporozoite rate and mosquito density in the study area. Bayesian variable selection was employed to determine the most important climatic predictors and elapsing (lag) time between climatic suitability and malaria transmission. Bayesian kriging was used to predict mosquito density and sporozoite rate at unsampled locations. These estimates were converted to covariate and season-adjusted maps of entomological inoculation rates. Models were fitted using Markov chain Monte Carlo simulation. RESULTS: The results show that Anophele. gambiae is the most predominant vector (79.29%) and is more rain-dependant than its sibling Anophele. funestus (20.71%). Variable selection suggests that the two species react differently to different climatic conditions. Prediction maps of entomological inoculation rate (EIR) depict a strong spatial and temporal heterogeneity in malaria transmission risk despite the relatively small geographical extend of the study area. CONCLUSION: Malaria transmission is very heterogeneous over the study area. The EIR maps clearly depict a strong spatial and temporal heterogeneity despite the relatively small geographical extend of the study area. Model based estimates of transmission can be used to identify high transmission areas in order to prioritise interventions and support research in malaria epidemiology. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1186/s13071-015-0679-7) contains supplementary material, which is available to authorized users.