Public health decision making with stochastic individual-based models: a translational framework driven by advances in health economics, model inference and reinforcement learning

This project proposes a methodological framework in the context of
respiratory pathogens with pandemic potential, based on historical
data of SARS-CoV-2. Clustered social contact patterns have been
pivotal in combination with stochasticity to explain disease spread
and heterogeneous behaviour. Therefore, we focus on mathematical
models that accommodate heterogeneity in infection acquisition and
additional randomness at the individual level. However, estimation of
key epidemiological parameters based on stochastic and
computationally intensive individual-based models is challenging.
Especially when we focus on multiple outcomes, which is required
when evaluating the health economic impact of preventive measures.
A coarse-grained cost-effectiveness analysis is possible through
individual-based modelling, yet complicated by a cascade of
uncertainties and stochasticity in the underlying disease process. The
availability of options to define the most (cost-)effective scenario
requires multi-criteria selection techniques. Machine learning
methods have been proven useful for this, however, this complex
modelling context requires progressive algorithms. Informing the
decision making process is particularly challenging in an epidemic
setting with unexpected events such as the emergence of new
variants of concern. We aim to accelerate decision making in the next
pandemic with a public health framework grounded in advanced
statistics, health economics and computer sciences