Ascertainment rate estimate from hospital data used in modelling COVID-19 epidemics
|Year of publication
|Appeared in Conference without Proceedings
|MU Faculty or unit
|We based our approach on a mechanistic compartmental SEIR model with additional undetected cohort A (stands for absent infected). To estimate the size of compartment A, we use a novel concept, a moving ascertainment rate estimate computed from data of hospitalized subjects. We estimate the probability of detection from the proportion of cases not detected before hospital admission using a conditional probability. We have developed an extended ZSEIAR model that also includes unknown dynamics in the affected clusters. We optimize the size of affected clusters in the model since the effects as seasonality or government measures cannot be easily distinguished. We submit our predictions to European Covid-19 Forecast Hub https://covid19forecasthub.eu/ and the web Czech Monitoring, Analysis and Management of Epidemic Situations https://webstudio.shinyapps.io/MAMES/.