Reflection on contextual factors associated with burden of infectious diseases in multi-country modelling approach on the example of Eastern Europe

Abstract

Thesis. The use of mathematical models to nowcast and forecast allows to improve predictive understanding of epidemiological targets and indexes among various populations during infectious disease epidemics. Such models enable trend predictions of various scenarios of the pandemic and guiding epidemic prevention and control.
 Concept. In order to avoid the influence of low-quality studies (for a given region) on the interpretation and decision making, a critical analysis based on experience in field epidemiology should be carried out. For the sake of transparency and quality of scientific discourse, such observations should indeed be collected and discussed by the scientific and medical community. Selected global modelling studies have been assessed according to their epidemiological outcomes such as cases and deaths.
 Results and Conclusions: I show that the discrepancy between reported and predicted epidemiological features varies more significantly than the order of magnitude between the countries in selected models. My findings highlight that models’ results; readers should prefer locally developed models over multi-country models, even those being published in prestigious journals. Thus, agent based models should be prioritised against system dynamics or machine learning models. I suggest that future epidemiological models should adopt healthcare access as a factor of so-called dark figure of infections, especially in Eastern Europe.