Covid-19 dynamics in Albania: first estimates and projections

The SARS-CoV-2 epidemic is one of the biggest challenges healthcare systems worldwide have ever had to face. To curb transmission many countries have adopted social distancing measures and travel restrictions. Estimating the effect of these measures in each context is challenging and requires mathematical models of the transmission dynamics. Projections for the future course of the epidemic strongly rely on model predictions and accurate representation of real-time data as they accumulate. Here I develop an SEIR modeling framework for Covid-19, to evaluate reported cases and fatalities, and to enable forecasting using evidence-based Bayesian parameter estimation. This Bayesian framework offers a tool to parametrize real-time dynamics of Covid-19 cases, and explore the effect of control as it unfolds in any setting. I apply the model to Covid-19 data from Albania, where drastic control measures were put in place already on the day of the first confirmed case. Evaluating the dynamics of reported cases 9-31 March 2020, I estimate parameters and make preliminary projections. Three weeks into the measures, Albanian data already indicate a strong signature of more than 40% transmission reduction, and lend support to a progressively increasing effect of control measures rather than a static one. In the Albanian setting, the model and data match well, projecting the peak of the outbreak may be around 5-15 April, and be contained within 300 active confirmed cases if control continues with the same trend. This framework can be used to understand the quantitative effects of different control measures in real-time, and inform adaptive intervention for success in other settings.