Simulating market-oriented policy interventions for stimulating antibiotics development

The relative efficacy of intervention policies, aimed at stimulating development of antibiotics, can be estimated using Agent Based simulation. We propose that antibiotics development can be modeled as Markov Chains with time and cash loaded transitions, and that many intervention policies can be modeled as alterations to the stochastic distributions of said Markov Chains. Through the combination of these two models, Agent Based simulation can be used to estimate the relationship between interventions and the Expected Net Present Value of products. We apply this modeling to an intervention policy proposed by the EU-initiative DRIVE-AB, targeting the urgent need for antibiotics research and development due to increasing resistance. We focus on variants fully delinking profit from volume sales, and show that (1) implementation variations lead to differences in outcomes, and that (2) they exhibit diminishing returns.