[ad_1]
“If you want to know what the weather is like, look out the window,” writer Max Goldt once said. And who wants to know what the weather will be tomorrow? He “can look out the window tomorrow.” That may be true for the weather. It is not easy to find out exactly where Germany is in the crown pandemic. A look at today’s data in no way reflects the current situation. Instead, you look back, sometimes several weeks.
Tomorrow’s vision is full of even more uncertainties. This is a problem for politicians who have to decide on relaxation. How should I continue? Meanwhile, many stores have reopened, school operations are beginning, and at least daycare concepts are being developed. In other areas, for example, for large events or for regulations on visits to hospitals or nursing homes, the restrictions should persist for a long time.
Several indicators play a role in the discussion on what measures should be maintained against the spread of the pathogen Sars-CoV-2. One of them is the number of new infections (read more about the number of reproductions, the proportion of people who have recovered from the sick and other parameters). The problem with this: The Robert Koch Institute (RKI) reporting data for new infections is extremely incomplete. And this is not only due to the large number of unreported cases, which can remain symptom-free, and therefore tests for the coronavirus are never done.
Recording officially known cases is also not that easy. The health authority at the place of residence of a sick person finds out about the case only after a delay. Time is wasted again until locally collected data is centrally processed at RKI headquarters in Berlin. The problem of reporting delay should now be solved with a calculation model.
The wave of infection was not what was expected at the beginning of the outbreak.
In its daily management report, the RKI now shows a different course of the pandemic, in addition to the reports received on newly infected people thanks to this model. It is no longer based on the date of registration, but on the supposed onset of the disease. If you look at these numbers and their development, there are some interesting effects:
Tracking the course of the disease after the date of the report, in the diagram above, this is the blue line, brings with it several problems. It is inevitable that you will fall behind the infection due to the delay in reporting. In particular, the most recent days in the statistics are based on very incomplete data. And some health authorities do not send any cases on weekends, so the number of reports drops regularly on Saturdays and Sundays.
No prediction of the future, but of the present.
A more realistic image shows the course of time after the suspected onset of illness, shown in the diagram above by the red line. The onset of the disease is known in a large part of the registered cases. For patients without a known onset of disease, time is statistically calculated using empirical values. Thus, the cases of which the RKI may not have been aware until the second half of March are already assigned to the alleged days of onset of the disease in the first half of the month.
In addition to the known and calculated onset of disease, the new time course also contains prognoses for new infections, the reports of which have not yet been received by the RKI. The so-called “nowcast” method, which is also based on empirical values and statistical estimates, is used to compensate for delays in reporting. It is not a prediction of the future, how to use it in the weather, but one of the present that cannot be better described. This can still help political decision makers.
The “NDR” reports that the RKI has been providing “Nowcast” data to the government since April 9. Previously, the model had to be adapted to the circumstances of the pandemic and fed with data. The tool was not yet available for political action against Corona in March.
But what about the now-recognizable effect that the number of new infections had already decreased before federal and state governments established contact blocks? Virologist Christian Drosten de Charité in Berlin justified this, among other things, by the prior cancellation of major events: “That prevented much basic transmission activity and the spread of large outbreaks.” School closings, which had also gone into effect earlier, also ensured that transfers between households were “very severely limited” even before contact restrictions.
Nowcast data can now be useful for possible future waves of infection. Also in these cases, just looking at the log data would distort the events again. If more people are infected, this is reflected in the officially reported figures just a few days later. If the number of new infections per day remains the same, the latest log data always suggests a decrease in the past few days. And if the number of new infections decreases, the delay in reporting increases the impression of the decrease.
One problem persists, of course: A person infected with Covid 19 only appears in the statistics if they have also tested positive as such. The question of how widespread the disease is in the population cannot be answered with new data processing.
