Translated newspaper article about the study discussed here earlier about viral loads in children by Drosten:
Christian Drosten, the top virologist who stumbled on statistics
As a world authority, he is an important face of science in the corona crisis. Now the German professor Christian Drosten rolls over the pavers with the boulevard magazine Bild about a study on children. 'As colleagues, we warned Drosten early on that his statistics are no good. He should have been extra careful. "
There were doubts from the start. Doubts about the probative value of the data, doubts about the conclusion, doubts about the policy advice that the scientists linked to their study: 'Judging from these results, we have to be very careful in the unlimited reopening of schools and nurseries in the current situation. Children could be as contagious as adults. " Despite all doubts, the study, including policy advice,
picked up internationally . The study was therefore not from any stable. It happened under the watchful eye of Professor Christian Drosten, the German virologist from the Berlin University Hospital Charité. Drosten heads the virology department. In 2003, he co-discovered the first Sars virus. So yes, feel free to call Drosten a world authority when it comes to corona. But now, four weeks after the publication of the manuscript, it appears that the first doubts about the study were not unfounded (
DS April 30 ) . Indeed, there seems to be something wrong with the statistical analysis. This study stands or falls with sound statistics. She compares how much virus people with sars-CoV-2 have in their throats. The "viral load" in the throat could be a measure of how contagious the person is. And so it's a good idea to compare the viral load of children and adults. Statisticians, however, have questioned whether these comparative analyzes between age groups have been properly performed. A correct analysis might even lead to a completely new conclusion: that children have less virus in their throat than adults. This, of course, sheds a different light on the reopening of the schools. A compelling discussion, actually, and very relevant in clarifying the riddle surrounding the role of children in this epidemic. But how the discussion emerges and how it is conducted is not a good story.
Campaign against Drosten
The
German tabloid Bild on Monday joined the criticism of some statisticians on the study by quotes from their more extensive analysis of the Drosten study put online. The statisticians criticize the applied statistical tests and come to different conclusions with the same data and other more accurate tests. Important detail: the statisticians in question never wanted Drosten to lose weight with their critical analyzes. They emphasize that they wanted to contribute constructively to the scientific debate through the usual channels - and thus transparent and accessible to everyone - just as they would under 'normal' circumstances. They did not want their criticism to become Fressen for Bild . But it just became. This article in Bild is the provisional culmination of the campaign against Drosten. The Bild- critical blog
BILDblog has drawn up a list of articles at the beginning of this month with which the newspaper tries to show how Drosten would contradict and that he would be at odds with colleagues and politicians like Chancellor Angela Merkel. There is always little real about it. Drosten, the defender of drastic measures to stop the spread of the coronavirus, rather acts as a lightning rod for the frustration with the German corona approach. It should come as no surprise that Drosten does not accept that Bild writes that his study contains 'gross mistakes', that he uses 'questionable methods' and that the newspaper asks 'how long has the top virologist known about it'. The professor posted a screenshot of an email on Twitter on Monday afternoon. In that email, a Bild journalist asked for a response. 'Interesting. Bild is preparing a biased message about us manuscript around the virus load ', added Drosten. "I have to respond within an hour. I have better things to do. "
Weak methodology
Does it really not make sense then, what does Bild write about Drosten again? Is it really all nonsense? Drosten, who did not respond to a request for a response from De Standaard yesterday , says in the latest episode of his podcast on
German public broadcaster NDR that he consciously used crude statistical methodologies, but that the medical significance would not change with finer techniques . However, the criticism of the statisticians goes to the core of the Drosten study. In reports that have multiple statisticians online they fillet his statistics. This includes the 'Omnibus test', with which Drosten and his colleagues tested whether the viral load is the same for all age groups. They also critically examine the 'log transformation' with which Drosten and his colleagues present their results. Statisticians are not quick to get rid of it.
In their own analyzes, statisticians even come to very different conclusions than Drosten and his team. "In this study, children have on average 67 to 85 percent less virus load than adults," Christoph Rothe, a professor at the University of Mannheim specializing in statistics, summarized the findings of a fellow statistician on Twitter last weekend. "That such large differences were dismissed as" insignificant " by the authors ( Drosten's team, ed. ) Is due to the fact that the statistical methods used are very weak." That Rothe colleague, Dominik Liebl (professor at the Universität Bonn), immediately responded on Twitter: 'Thank you, that's exactly how I wanted my report to be understood!' Jörg Stoye, who teaches statistics as a professor at Cornell University in the US, even writes in his own report: 'There are many good arguments against a quick restart of the schools, but the study of the Charité does not help . " In an interview released on
Der Spiegel 's
website yesterday , Stoye says that like his colleagues who are quoted by Bild , he is very unhappy about how their criticisms are getting to the press now. "I don't want to be part of a Bild campaign." But Stoye confirms to Der Spiegel that he does not agree with the study of Drosten. The children in the study carry fewer viruses than the adults. The question then is: is that just a coincidence? Or is there a pattern behind that? I think it is the latter, unlike the study. We do indeed disagree there. ” In his report, and again in the interview, stresses Stoye that he suspects no malicious intent at Drosten.
Correct and continue
Herman Goossens, professor of microbiology at UZ Antwerp, also does not suspect his colleague - the two are part of a European network of researchers - of ill will. "He's a brilliant virologist. The virological part of the study is also all right. But the statistics… ' Two days after the publication of the manuscript, on Friday 1 May, Goossens and Drosten held their weekly teleconference with all other colleagues in their network. "We told him clearly then:" Christian, your statistic is no good, you go too fast in your conclusions, this has enormous consequences for policy. " Especially in his role as a scientist who is asked for advice, he has to be extra careful. ' 'As a scientist, you must remain open to the fact that it all works differently than you first assumed. The jump from research results to a policy recommendation on schools has left him very vulnerable. Christian was suddenly too much a policymaker and too little a scientist. " To protect each other from similar errors in papers, Goossens and his European colleagues have agreed to critically discuss new manuscripts among themselves before swirling them online. Earlier Goossens denounced the scientific review articles in corona times has been postponed too many times (
DS May 4 ) . According to Goossens, he did not deserve that Drosten is now finished by Bild . "This is one extension. Under normal circumstances I would say: so what , your statistics are not correct, correct it and continue. '
Facebook
Twitter
Instagram