We’ll start providing unofficial English summaries of the NDR coronavirus update podcast (see notion linked). The podcast provides weekdaily interviews with Christian Drosten.
Christian Drosten is leading the virology department at the German university hospital “Charite” in Berlin. He co-developed the PCR test that Germany is using to detect SARS-CoV2 (the virus causing Covid-19), is continuing to research Covid-19 and consulting the German government. He has been researching Coronaviruses, including the old SARS, for many years.
Important: We are no virologists, epidemologist, medics or have any other relevant expertise. We are lay people trying to write up Christian Drosten’s evaluations as good as possible without verifying anything he says and we might misrepresent some things. Christian Drosten himself is not the relevant expert for some of the niche topics he talks about and gives informed guesses which he sometimes revises at later times. None of this should be taken at face value without double checking.
German transcripts (time-lag of one day) are available here.
Here are the summaries from today and yesterday, written by Chau Nguyen with some editing from me:
Episode 30: 2020/04/03 - Scientists hope for “data donations” The most promising tool to safely exit lock-down is a mobile contact tracing app and Germany releases symptom tracker app
The German health institute (RKI) published a “data donating” app to report fitness tracking data from smart watches etc. and build a “fever” map in Germany
Estimates fevers but no contact tracing
Users enter postcode and demographic data
Concept for potential contact tracing app (most promising tool to exit lock-down):
Symptoms are reported to the app, which orders tests automatically, informs contacts, estimates the risk of infection and gives isolation instructions
Use user tokens and bluetooth (no GPS) to track contacts which would protect data privacy
From a previous episode: a recent model suggests R0 < 1 is possible, if
60% of all citizen use an app
60% of these follow instructions for self-isolation
This is assuming a start R0=2 with 0.9 transmitted pre-symptomatically, 0.8 symptomatically, 0.2 via surfaces, and 0.1 asymptomatically
Episode 29: 2020/04/07 - We need to target testing better
We don’t know how most infections occur
Too early to find a good answer
Lockdown means most infections currently occur in private households which gives us no indication on how infections would happen in “real” daily life.
Early stage of the pandemic means we don’t have enough data, yet
Ceiling for scaling up testing makes targeted testing and tracing apps more important
Germany
Last week around 350k tests were reported even though the RKI (Germany’s public health institute) predicted for 500k
Possible explanation 1: Some tests might not be reported yet and will be reported in the upcoming days
Possible explanation 2: Shortage of certain resources such as reagents or swabs might be hitting the test centres as all European countries have an increased demands for these resources—there is also anecdotal evidence for this
Politicians started reaching out to companies asking why their country’s demand is not being met
Global
Conclusion: Even if test centres are able to increase efficiency, the necessary resources (reagents, swabs) are a bottleneck, so we can’t increase testing as necessary
Upshot: We need to target testing better
Catch hospital employee infections early and who already had it
Prioritize according to who’s likely to be hospitalized to prevent them being hospitalized too late: Check-in by phone every two days with positive cases
This is an argument for importance of tracking via apps
It is unlikely that there is a high dark figure of infected people, and there are false positive antibody tests
Antibody tests have some cross reactions with other corona viruses: Someone recently sick with another corona cold virus might test positive in a Sars-CoV-2 antibody test
Germany and other countries (e.g. China) started doing antibody tests: No evidence for a surprisingly high percentage of people that have already been infected but have not noticed
Antibody tests can track silent infections via seroconversion number: People who have been tested in one week without antibodies and then tested in the next week with antibodies
Tests have to be rolled out country-wide and not just regionally to factor out coincidence
Potential prospects of tracking via sewage water and tab water
English summaries of German Covid-19 expert podcast
Link post
This is a cross-post from LessWrong.
We’ll start providing unofficial English summaries of the NDR coronavirus update podcast (see notion linked). The podcast provides weekdaily interviews with Christian Drosten.
Christian Drosten is leading the virology department at the German university hospital “Charite” in Berlin. He co-developed the PCR test that Germany is using to detect SARS-CoV2 (the virus causing Covid-19), is continuing to research Covid-19 and consulting the German government. He has been researching Coronaviruses, including the old SARS, for many years.
Important: We are no virologists, epidemologist, medics or have any other relevant expertise. We are lay people trying to write up Christian Drosten’s evaluations as good as possible without verifying anything he says and we might misrepresent some things. Christian Drosten himself is not the relevant expert for some of the niche topics he talks about and gives informed guesses which he sometimes revises at later times. None of this should be taken at face value without double checking.
German transcripts (time-lag of one day) are available here.
Here are the summaries from today and yesterday, written by Chau Nguyen with some editing from me:
Episode 30: 2020/04/03 - Scientists hope for “data donations”
The most promising tool to safely exit lock-down is a mobile contact tracing app and Germany releases symptom tracker app
The German health institute (RKI) published a “data donating” app to report fitness tracking data from smart watches etc. and build a “fever” map in Germany
Estimates fevers but no contact tracing
Users enter postcode and demographic data
Concept for potential contact tracing app (most promising tool to exit lock-down):
Symptoms are reported to the app, which orders tests automatically, informs contacts, estimates the risk of infection and gives isolation instructions
Use user tokens and bluetooth (no GPS) to track contacts which would protect data privacy
From a previous episode: a recent model suggests R0 < 1 is possible, if
60% of all citizen use an app
60% of these follow instructions for self-isolation
This is assuming a start R0=2 with 0.9 transmitted pre-symptomatically, 0.8 symptomatically, 0.2 via surfaces, and 0.1 asymptomatically
Episode 29: 2020/04/07 - We need to target testing better
We don’t know how most infections occur
Too early to find a good answer
Lockdown means most infections currently occur in private households which gives us no indication on how infections would happen in “real” daily life.
Early stage of the pandemic means we don’t have enough data, yet
Ceiling for scaling up testing makes targeted testing and tracing apps more important
Germany
Last week around 350k tests were reported even though the RKI (Germany’s public health institute) predicted for 500k
Possible explanation 1: Some tests might not be reported yet and will be reported in the upcoming days
Possible explanation 2: Shortage of certain resources such as reagents or swabs might be hitting the test centres as all European countries have an increased demands for these resources—there is also anecdotal evidence for this
Politicians started reaching out to companies asking why their country’s demand is not being met
Global
Conclusion: Even if test centres are able to increase efficiency, the necessary resources (reagents, swabs) are a bottleneck, so we can’t increase testing as necessary
Upshot: We need to target testing better
Catch hospital employee infections early and who already had it
Prioritize according to who’s likely to be hospitalized to prevent them being hospitalized too late: Check-in by phone every two days with positive cases
This is an argument for importance of tracking via apps
It is unlikely that there is a high dark figure of infected people, and there are false positive antibody tests
Antibody tests have some cross reactions with other corona viruses: Someone recently sick with another corona cold virus might test positive in a Sars-CoV-2 antibody test
Germany and other countries (e.g. China) started doing antibody tests: No evidence for a surprisingly high percentage of people that have already been infected but have not noticed
Antibody tests can track silent infections via seroconversion number: People who have been tested in one week without antibodies and then tested in the next week with antibodies
Tests have to be rolled out country-wide and not just regionally to factor out coincidence
Potential prospects of tracking via sewage water and tab water
Group in the Netherlands tries to track Sars-CoV-2 in sewage water
First positive evidence, group did the same for Polio before
Note: No evidence that virus load found in sewage water would be infectious, just a potential tracker