Quantification of cure for pharmacodynamic models of antimalarial drugs: Deterministic versus stochastic approaches

In silico pharmacokinetic-pharmacodynamic (PK-PD) models are used to inform dose optimisation in antimalarial drug development. Here the PK-PD models capture the change in parasite count over time because of exposure to antimalarial drug(s). For current deterministic PD models, simulated parasite numbers decline proportionally during treatment asymptotically approaching zero, requiring a cure threshold to be selected. We implemented an alternative stochastic model, where parasite-time profiles were simulated using a binomial function with an hourly parasite survival probability. This probabilistic model enables a time of exact cure to be calculated, when zero parasites remain. The 28-day cure rates predicted by the deterministic and stochastic models for multiple antimalarial treatment strategies were equivalent, with an absolute difference ranging from -0.8% to 3.1% (mean of 0.48%, positive values indicate higher stochastic model cure rates). This confirms that the standard deterministic method is an adequate proxy for the underlying stochastic process of parasite death during antimalarial treatment.