Modeling schistosomiasis spatial risk dynamics over time in Rwanda using zero‑inflated Poisson regression

Elias Nyandwi, F.B. Osei*, S. Amer, A. Veldkamp

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

3 Citations (Scopus)
47 Downloads (Pure)


The recorded clinical cases of S. mansoni at primary health facility level contain an excessive number of zero records. This could mean that no S. mansoni infection occurred (a true zero) in the health facility service area but it could also that at least one infection occurred but none were reported or diagnosed (a false zero). Standard statistical analysis, using exploratory or confirmatory spatial regression, fail to account for this type of data insufficiency. This study developed a zero-inflated Poisson model to explore the spatiotemporal variation in schistosomiasis risk at a fine spatial scale. We used environmental data generated at primary health facility service area level as explanatory variables affecting transmission risk. Identified risk factors were subsequently used to project the spatial variability of S. mansoni infection risk for 2050. The zero-inflated Poisson model shows a considerable increase of relative risk of the schistosomiasis over one decade. Furthermore, the changes between the risk in 2009 and forecasted risk by 2050 indicated both persistent and emerging areas with high relative risk of schistosomiasis infection. The risk of schistosomiasis transmission is 69%, 29%, and 50% higher in areas with rice cultivation, proximity to rice farms, and proximity to a water body respectively. The prediction and forecasting maps provide a valuable tool for monitoring schistosomiasis risk in Rwanda and planning future disease control initiatives in wetland ecosystem development context.
Original languageEnglish
Article number19276
Pages (from-to)1-9
Number of pages9
JournalScientific reports
Issue number1
Publication statusPublished - 6 Nov 2020




Dive into the research topics of 'Modeling schistosomiasis spatial risk dynamics over time in Rwanda using zero‑inflated Poisson regression'. Together they form a unique fingerprint.

Cite this