Read on Twitter
Our statistical tool to analyze the epidemic trends of #SARSCoV2 in Switzerland is now online. (1/n) @ISPMBern @unibern @ETH @SwissScience_TF https://ibz-shiny.ethz.ch/covidDashboard/
In the current situation, it is critically important to follow the epidemic trends of #SARSCoV2 across age groups and cantons. Changes in the number of daily confirmed cases, hospitalizations, ICU occupancy and deaths can all provide meaningful insights. (2/n)
One can fit a negative binomial generalized linear model (glm.nb in R) to the data with reported numbers as a response variable and date and weekend as predictors. This allows to estimate the exponential increase or decrease of the different indicators of the epidemic. (3/n)
The trends are typically best observed over a time period of two weeks. Due to reporting delays, we also need to remove the last 3 and 5 days of confirmed cases and hospitalizations/deaths, respectively. (4/n)
On a national level, we see that confirmed cases and hospitalizations decreased during the first two weeks of November with a half-life of 19 and 21 days, respectively. (5/n)
In contrast, ICU occupancy and deaths were still slowly increasing but might now follow the trend indicated by the other two indicator variables. (6/n)
Rather than just looking at national trends, it is important to follow the situation across different cantons and age groups. Confirmed cases declined in all age groups, but to a lesser extent in the youngest and oldest. (7/n)
Given that various cantons have different measures in place, it is important to analyze how they perform in reducing the numbers of confirmed cases. (8/n)
The cantons in the French-speaking part of Switzerland (GE, VD, VS, NE, FR, JU) have all implemented additional control measures and could reduce the number of confirmed cases by 50% in less than two weeks. (9/n)
Most other cantons are above the national average and see a slower decline in confirmed cases or even an increase (BS and BL). The coming days and weeks will show whether they need to implement additional measures as well. (10/n)
