AUSTRIA

Current forecasts for case numbers and bed occupancies

current epidemiological analysis as of 19.10.2022 - nowcasting

Infection numbers, susceptible individuals, exogenous driver

The following figure shows the course of the active case numbers in the upper left-hand corner (random fluctuations, such as those caused by late reports, are reduced by suitable smoothing). Based on this, estimates of the susceptible proportion of the population (middle left) and the aggregated exogenous drivers, which significantly determine the epidemic events (bottom left), are produced using the differential flatness methodology. In addition, the diagram on the bottom left also shows the course of vaccination rates.

The diagram on the right shows the state progression in a phase diagram. Periods during which a hard lockdown was imposed are marked in purple, details of which can be found in our publication in the Journal of Nonlinear Dynamics.

 

Reproduction number

Für die Analyse der COVID-19-Epidemie kann die effektive Reproduktionszahl R_{\text{eff}} eines Landes als Maß zur Beschreibung der Infektionsaktivität im Zeitverlauf herangezogen werden. Dabei zeigt R_{\text{eff}}>1  ansteigende Aktivität bzw. Wachstum an. In der folgenden Abbildung wird die von uns berechnete Reproduktionszahl sowie zum Vergleich die amtlich gemeldete Reproduktionszahl dargestellt:

CURRENT PROGNOSIS of epidemiological conditions as of 19.10.2022 - infection figures.

The following figure shows the current epidemiological status analysis (nowcasting) and a three-week forecast for the total population.

The top diagram shows the development of the active case numbers including their forecast course, the middle diagram shows the flatness-based estimate of the proportion of susceptible individuals as well as its forecast course. Finally, the bottom diagram shows the course and prognosis of the exogenous input, which significantly determines the further course of the epidemic. For example, an increase in the exogenous input can be interpreted as an early indicator of rising case numbers.

Note: The effects of the opening steps of 5 March 2022 have not yet been taken into account in the current forecasts. This will be done as soon as sufficient data is available to be able to seriously assess the situation.

Forecast and analysis by age group

Die folgende Abbildung zeigt die aktuelle epidemiologische Analyse und Prognose für Österreich anhand zweier Altersgruppen (Gruppe 1: <65 Jahre, Gruppe 2: über 65 Jahre). Details dazu finden sich in unserer zweiten Publikation in Journal of Nonlinear Dynamics.

Note: The effects of the opening steps of 5 March 2022 have not yet been taken into account in the current forecasts. This will be done as soon as sufficient data is available to be able to seriously assess the situation.

Current hospital utilization projections as of 10/19/2022.

The following figure shows the current three-week forecasts for normal and ICU bed occupancy, respectively. These are based on forecasts of case numbers as well as real-time estimation of case-specific hospitalisation and ICU rates:

Note: The effects of the opening steps of 5 March 2022 are not yet taken into not yet taken into account in the current forecasts. This will be done as soon as sufficient data is available to assess the situation seriously. situation.

The figure above shows the actual occupancies of hospitals and intensive care units reported by the Austrian Agency for Health and Food Safety (coloured pink and blue) as well as occupancies smoothed over time. In addition, our forecasts (dashed line) and a forecast interval (grey) are shown.

Data preparation

The data sets used (raw data) are based on the data provided by the Austrian Agency for Health and Food Safety. They show certain country-specific anomalies, mainly due to inconsistencies in the reporting of positively tested and recovered persons. In some cases, inconsistencies were also found in the reported data from the electronic reporting system. In addition, there is a weekly pattern of under-reporting and over-reporting (as fewer tests are usually performed/entered on weekends). It should be noted that the infected persons are determined by cumulating the daily cases of the last fortnight. To reduce the impact of such anomalies, the data were first smoothed in an appropriate way. A window length corresponding to a seven-day multiplier was chosen. The smoothing is based on local regression using weighted linear least squares and a second order polynomial model.