Predictability of epidemic malaria under non-stationary conditions with process-based models combining epidemiological updates and climate variability.

摘要: Previous studies have demonstrated the feasibility of early-warning systems for epidemic malaria informed by climate variability. Whereas modelling approaches typically assume stationary conditions, epidemiological are characterized changes in intervention measures over time, at scales longer than inter-epidemic periods. These trends control efforts preclude simple application validated retrospective surveillance data; their effects also difficult to distinguish from those variability itself. Rainfall-driven transmission models falciparum and vivax fitted long-term data four districts northwest India. Maximum-likelihood estimates (MLEs) model parameters obtained each district via a recently introduced iterated filtering method partially observed Markov processes. The resulting MLE is then used generate simulated yearly forecasts two different ways, these compared with more recent (out-of-fit) data. In first approach, initial conditions generating predictions repeatedly updated on basis, based new inference that naturally lends itself this purpose, given its time-sequential application. second themselves refitting moving window time. Application examine predictability reveals differences effectiveness parasites, illustrates how ‘failure’ can be informative evaluate quantify effect context approach performs adequately, sometimes even better one, when remains major driver dynamics, as found Plasmodium which an effective clinical lacking. offers skillful dynamics shift case years declining incidence under improved control. Predictive infectious diseases such malaria, process-based variables, applicable non-stationary conditions.