Abstract
BackgroundMulti-model ensembles could overcome challenges resulting from uncertainties in models’ initial conditions, parameterization and structural imperfections. They could also quantify in a probabilistic way uncertainties in future climatic conditions and their impacts.MethodsA four-malaria-model ensemble was implemented to assess the impact of long-term changes in climatic conditions on Plasmodium falciparum malaria morbidity observed in Kericho, in the highlands of Western Kenya, over the period 1979–2009. Input data included quality controlled temperature and rainfall records gathered at a nearby weather station over the historical periods 1979–2009 and 1980–2009, respectively. Simulations included models’ sensitivities to changes in sets of parameters and analysis of non-linear changes in the mean duration of host’s infectivity to vectors due to increased resistance to anti-malarial drugs.ResultsThe ensemble explained from 32 to 38% of the variance of the observed P. falciparum malaria incidence. Obtained R2-values were above the results achieved with individual model simulation outputs. Up to 18.6% of the variance of malaria incidence could be attributed to the +0.19 to +0.25°C per decade significant long-term linear trend in near-surface air temperatures. On top of this 18.6%, at least 6% of the variance of malaria incidence could be related to the increased resistance to anti-malarial drugs. Ensemble simulations also suggest that climatic conditions have likely been less favourable to malaria transmission in Kericho in recent years.ConclusionsLong-term changes in climatic conditions and non-linear changes in the mean duration of host’s infectivity are synergistically driving the increasing incidence of P. falciparum malaria in the Kenyan highlands. User-friendly, online-downloadable, open source mathematical tools, such as the one presented here, could improve decision-making processes of local and regional health authorities.
Highlights
Multi-model ensembles could overcome challenges resulting from uncertainties in models’ initial conditions, parameterization and structural imperfections
Kericho provides a good scenario for modelling the timing and severity of P. falciparum malaria outbreaks and the potential impact of changes in climatic conditions on malaria morbidity profiles
This paper described the results of the implementation of a multi-malaria-model ensemble framework to assess temporal changes in malaria morbidity profiles in Kericho, in the highlands of Western Kenya
Summary
Multi-model ensembles could overcome challenges resulting from uncertainties in models’ initial conditions, parameterization and structural imperfections. Recent advances in the theory of mosquito-borne pathogen transmission seeks to better understand uncertainty in the traditional malaria modelling framework by realistically acknowledging spatial heterogeneity of transmission in complex epidemiological landscapes [3]. Another important approach to reducing uncertainty in model results is improvements in the quality and quantity of appropriate data used to both drive and test the model outputs. After identifying the most appropriate model(s) with least assumptions and using the best available data, two other sources of uncertainty must be taken into account These are the starting conditions used to initialize the model and the specific parameterization of the model itself. Choices made in model structure are significant sources of model uncertainty
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
