From: Mathematical modelling for antibiotic resistance control policy: do we know enough?

(1) Explaining population level resistance trends: testing and combining current model structures with diverse multi-level datasets. While there exists a suite of plausible mechanisms that may drive trends in resistance evolution, we currently lack the empirical data to evaluate the relative importance of these mechanisms. To resolve this difficulty, we will have to collect these data and systematically calibrate the suite of models to these data. Doing so will allow us to not only distinguish the underlying mechanism(s), but also to quantify other key parameters such as the strength of selection and competition for a particular bacteria and drug. | |

(2) Disentangling transmission routes: fitting models to data to generate a standardised modelling framework that shows the pathways of ABR will help to improve intervention targeting as well as to predict future burden. | |

(3) Translating model predictions to economic outcomes: evaluating the cost-effectiveness of competing ABR control strategies. Although the framework for integrating mathematical model predictions into economic frameworks exists in principle, more work is needed to develop methods specific to antibiotic resistance, such as calculating the short- and long-term costs of antibiotic resistance across priority pathogens and correctly identifying counterfactual scenarios that are contingent on the epidemiology of the pathogen and its setting. Adopting a standardized approach for evaluating the efficiency and optimality of strategies would be invaluable across hospital settings, where arguably the need is greatest. |