SteveFx

Interesting article about germany’s Coronavirus expert Christian Drosten https://www.theguardian.com/world/2020/apr/26/virologist-christian-drosten-germany-coronavirus-expert-interview?fbclid=IwAR18_-gYuO8gOuwG-jDS4Eam0SD6nNplRBajrO1LPHaHGJqTK6-TZUa0370

Here is an update on Sweden as I think it instructive to see whether a more relaxed approach can work. This is daily reported deaths compared with UK - smoothed over 7 days. https://ourworldindata.org/grapher/daily-covid-deaths-per-million-7-day-average?time=2020-03-06..&country=SWE+GBR I would suggest that shows a steady decline in the UK since 10th April and for Sweden the trend is still upwards. However one could perhaps make the case that the 4 points for Sweden show a possible flattening. The reported deaths figure is in both cases the number of deaths formally registered - the actual date of death may be several days previously. It is possible to find a graph which shows the number of deaths that occurred on a date. So if you look here you can find an alternative view (This is for England which is 75% of the UK) https://www.cebm.net/covid-19/covid-19-death-data-in-england-update-24th-april/ Here we can see that the actual peak was on 8th April and there has been a steady decline since. The grey area at the right are data points that are incomplete as it is likely that more deaths will be logged against those dates in subsequent daily updates. Most of the changes will be to the last 4 data points which will probably rise to match the line of steady decline. The BBC has a lengthy article on Sweden and if you look this includes a similar graph mapping reports to the date of death. This again has a grey area of incomplete data points https://www.bbc.co.uk/news/world-europe-52395866 So this suggests that Sweden has reached a possible levelling out or is continuing to rise won't be clear until next week. As an example of how hard it is to rely on the daily reports, the UK reported a big figure today (813) but many of them are for deaths weeks ago. Sweden reported 131 yesterday but just 40 today.

The link I posted was for confirmed deaths per million not cases, I put *cases* in the text by mistake - you are right it would be the wrong thing to compare which is why I didn’t. So it’s exactly the same graph as the other one, but this time continued to today and not ending when a quirk in the Swedish reporting made it look as if they had started to dip when in fact they then experienced a steep rise. it isnt certain what will happen next to Sweden. Their figure of 175 two days ago would be equivalent to 4000 deaths in a day in the USA, but if I read their dashboard correctly ICU beds may be dropping which may be a good sign.

Try adding Belgium they are in the 20s - also New York would have been in the mid 20s at their peak. I have been watching Sweden as the Imperial model believes that their R is still around 2. Their daily reporting jumps around a lot more than other countries. So yesterday Sweden had a huge 170 odd and today 84 - the UK is 6.5 x the size so that would be like 1100 and 550 for the UK - the UK was 800 ish yesterday and 616 today. The next week will tell whether the more relaxed Swedish social distancing will work or not.

Here is the current graph normalised for deaths per million showing UK and Sweden - which show the Sweden looks to have a problem... https://ourworldindata.org/grapher/daily-covid-deaths-per-million-7-day-average?country=SWE+GBR

turn on smart grid super sampling in nvidia inspector

Dave mentioned this study which at the time I (and possibly Alan) thought was the same as the Santa Clara study. but it isn't - LA County is an actual place in LA, Santa Clara county is somewher else in California. http://www.publichealth.lacounty.gov/phcommon/public/media/mediapubdetail.cfm?unit=media&ou=ph&prog=media&cur=cur&prid=2328&row=25&start=1 Reading the press release it sounds better than the Santa Clara study. It seems to suggest that participants were recruited using a database to get a statistically correct population sample. No further details are given so we don't know what the response rate was like and so whether there could still be a possible bias with the people agreeing being more likely to include people who thought that they had been infected. They tested on April 10th and 11th and estimated that 4% of the population had been exposed to covid-19 i.e around 300K. Following the logic in the Santa Clara study we should treat that as being the proportion 4 days earlier on 6th April. We need to then look 3 weeks forwards from that date to allow for the progression of the disease which gives 27th April as the date to use for the death figure. To date there have been a total of 660 deaths and its going up by around 40 a day. So that gives us 860 as an approximate figure to use. That suggests a fatality rate of 0.3% which is good news as its less than half the values being currently used (0.7-0.9%).

On the french aircraft carrier we know that almost half the crew tested positive with the antigen test. The Austrian population sample used the antigen tests on a statistically selected group from which they inferred the population infected to date. So the % infected cannot be more than twice what they calculated. New York also sets a minimum IFR of 0.1% with new infections still occurring. The number of people one person infects (R) can be calculated from the rate at which the numbers change so it’s at least 3. The % of the population that needs to be infected can be calculated from R So if R is 3 then it’s just over 66%. If you think about it if a person passes it on to 3 others and 66% are immune then 2 of the 3 it would have been passed onto would be immune and it only gets passed on to 1 other person. If slightly more than 66% are immune then it dies out as each person passes it on to less than one other.

@ birdman All the studies to date suggest that only a small percentage of most populations have currently had covid19. Three in a hundred in Netherlands (4000 deaths) one in three hundred in Austria ( 500 deaths) and according to the Stamford study four in a hundred in Santa Clara ( 90 deaths however there are some flaws in that last study). As one in a thousand people in New York have died we can conclude that a great deal more people there have been exposed. When talking about lockdowns you keep coming back to your risk which is the wrong way of looking at it. Consider voting. Logically there is no point in any one person voting as their vote will never be the deciding vote and so the outcome will be the same if they hadn’t bothered to vote. However society functions better if we all ignore that and behave as if our vote matters and go and vote as the majority then elect the candidate that they want. In the same way the lockdown isn’t to protect you. There are only currently a very small number of people in Roswell who are currently symptomless with the virus and you are therefore unlikely to either be one of them or meet one of them. But there are almost certainly some people in Roswell who are infected and because this disease is very infectious each one will infect 3 others and each of those will infect 3 others and very quickly a lot of people will have it. The idea of the lockdown is that if everyone follows the rules then the people who do have it (whoever they are) will also follow the rules and so the virus doesn’t get passed on. Now if everyone decides that it is silly and doesn’t do it then the virus will get passed on. If everyone follows the rules then society as a whole benefits.

Back to the coronavirus... Here is paper from the Tony Blair institute considering exit options https://institute.global/sites/default/files/inline-files/A Sustainable Exit Strategy%2C Managing Uncertainty Minimising Harm.pdf The graph of Rt by country by intervention is interesting - it overuses blue but the blue points that are low down are actually Greece. Looking at that and cross referencing with the Imperial data it seemed that most European countries started with Rt of around 3.8-4.5 experimented with lots of things which got it down to around 2.5- 3.0 and then went for full lockdown which got it down below 1- typically around 0.7. But there are some outliers Belgium started at 6.8 and with full lockdown are above 1 still. Greece started at 1.8 and almost got it down to just above 1 without a lockdown and now seem to be at 0.3 with a lockdown. Why do Greeks pass the virus between themselves at such an astonishingly low rate? and why is Belgium so different?

So next week it will be tele-dentistry?

Delux included some additional content over standard - e.g garmin glass cockpit plus Maule and also included the SDK. Gold is Delux + Acceleration and includes F18 Hornet, P51 Mustang EH101 helicopter

That’s correct. The dx10 controller has its own fsx.cfg editor which you can use from the top menu which will do this for you if you tick the right option.

Going back to the concept of weekly population testing a group of academics have written to the uk government and suggested that we trial it on a city of 200k population and if it’s successful then build up a capability to test 10M people a day and then roll out to the whole country. https://marlin-prod.literatumonline.com/pb-assets/Lancet/pdfs/S0140673620309363.pdf For such a trial I would have thought that the Isle of Wight would be a great candidate. It has a population of 140k and is not connected to the mainland other than by ferries. This would clearly be a massive undertaking but at least its a way forward to restart normal life in a few months.

This is rather depressing.. this is the imperial model in particular the effect of social distancing etc. https://mrc-ide.github.io/covid19estimates/#/interventions If you then use the menu to look at a country like Germany you can see how all the interventions before lockdown serve to reduce Rt a little but it’s only lockdown that makes a big difference and gets Rt below 1. that being the case I don’t see how we can relax anything other than say letting schools go back,

The Imperial model used by the UK uses 0.9% if I recall correctly. The calculation that the Lancet paper came to using Diamond Princess and the sampling of returnees from Wuhan was around 0.7%. You can read the Santa Clara paper on medrxiv. It recruited by Facebook ads. Its therefore possible that people who thought they had been exposed might have have been more receptive to the appeal (which they acknowledge as a potential bias in the paper). Note that they actually had 50 positives out of 3300 tests (1.4%) . It gets scaled up to 2.5% - 4% because the test population disproportionally consisted of white women (soccer moms?!) The testing was carried out on the 3rd - 4th April. They explain that they assumed that antibodies could be detected 3 days after infection and thus took the figure to represent the number infected up to 1st April. They then projected forward deaths to 21st April to allow for the time lag from infection to death. Their calculated IFR is 0.12 -0.2% which is a lower than the concensus but not by 50 times. The 50-85 times claim is misleading its the ratio of estimated infections to positive tests. The Dutch study was antibody based but used blood donations and the profile of a blood donor might not match the population. They suggest that 3% of their population were infected at the beginning of April so roughly 500K. Following the same logic as the Santa Clara paper if we look at deaths today its 3600. So projected forward to 21st it will be approx 4K which gives an IFR of around 0.8%

There are starting to be some population sampling studies. These are early days and should be treated carefully. Netherlands 3% ( antibody of blood donations) https://www.reuters.com/article/us-health-coronavirus-netherlands-study/dutch-study-suggests-3-of-population-may-have-coronavirus-antibodies-idUSKCN21Y102 Austria 0.3% ( sampling using antigen) https://www.theguardian.com/world/2020/apr/10/less-than-1-of-austria-infected-with-coronavirus-new-study-shows Santa Clara County 3-4% ( antibody) https://edition.cnn.com/2020/04/17/health/santa-clara-coronavirus-infections-study/index.html It isn’t entirely trivial to map these to the deaths that had occurred, as you have to consider the timings involved, e.g when in the disease progression do antibodies occur in mild patients (that would have tested) against when deaths occur. They all suggest somewhat lower IFRs than the 0.7% which I think has been the figure used in the Imperial model. This is an animation showing how the daily cause of deaths has changed in the US over the last month or so. https://public.flourish.studio/visualisation/1712761/

Sorry to hear that Alan. I had to deal with exactly the same thing with my father last year, and yes that’s what it’s like. He is currently in a care home.

And I agree entirely. I have friends who own businesses who face ruin through no fault of their own. This is a very good article in the Guardian which discuses the idea of really scaling up testing to 10s millions a day as a possible way out. https://www.theguardian.com/commentisfree/2020/apr/17/political-imagination-end-lockdown-mass-testing-contact-tracing

If you see a car accident you drive extra carefully for what 25 minutes and then push it out of your mind. The same will be true people want to socialise - its who we are. After the big vaccination exercise when the hospitals report no cases we will be cautious for a few weeks and then all go out for a drink! Unless you get to Venice on the first day you will never see it quiet. I would hope that one of the treatments (or a combination) makes a big difference before we get there though.

To add to my last post, I do think that it is conceivable that there is some truth that the virus is passed on more in big cities but we need to look for evidence that that is the case. Perhaps the South Koreans can infer such things from their tracking apps, do people catch it on the subway or in the office or in bars?

Your principal argument is that "Its different here" - more specifically "people pass the virus on less in NM than in NYC" Now when Bergamo came to light in Italy in February we in the UK thought "Its different here" - more specifically "Italians socialise in big family groups", "there are lots of old Italians", "Italians hug and kiss all the time" If you look on OurWorldInData the day by day progress of deaths in the UK has followed exactly the same curve as Italy. Countries have tried to halt the virus with social distancing, hand washing, advice to avoid groups of 500, 100, 10, 3 etc and ultimately decided that the only way to stop it was a lockdown One of the benefits of the big reset we are all doing is to have a pause and try and learn how the virus spreads and to what extent we can relax yet ensure than on average an infected person doesn't pass it on to more than one other person. We need to boost testing such that we can test everyone with a cold. In NM I see on your dashboard that you have sadly had 44 deaths. Two weeks ago you had tested and found 400 positives. With an IFR of 0.5% there must have in fact been around 9000 infected people in NM at that time. Without a lockdown that would have grown to 130K by now if the infection rate is around the typical value of 3. Germany has started relaxing restrictions along with Austria and Denmark so that may give us all some clues. I worry that keeping a virus with a natural R0 of 3 down to 1.0 is going to be very difficult.

I think you are quoting me somewhat out of context. I have consistently argued the opposite to Noel. But I don’t think it’s fair to suggest that he is blatantly being word not allowed as was posted ( I know that wasn’t you), he is describing the world as he sees it and searching out new information and he is always friendly and respectful of other’s viewpoints and is pretty much following the state guidelines whilst querying why it is necessary. When reflecting on Noel’s viewpoint let’s imagine that the Coronavirus behaved differently. 0.5% of the population (mostly elderly) became ill and quietly passed away at home with a heart attack. Were that the case then I suspect that our governments would view that with sadness and provide advice for the elderly to isolate but that aside the world would carry on. The problem that that we face with the actual virus is that it hospitalises a much greater number maybe 2.5% In a distressed state and the majority of those can be saved with medical care ( e.g like Boris Johnson). The high Rt of around 3 means that numbers double every 3-4 days if left unchecked. To achieve herd immunity we would need 2/3rds of the population to have been exposed and even the best healthcare system in the world in Germany will be overwhelmed. When that occurs all the saveable cases will die as well and the exponential growth curve means that the vast majority of cases will occur when the healthcare system is overloaded. Hence the overall fatality rate will be around 2% of the population which is huge. When shown these numbers every world leader has concluded that this cannot be allowed to happen. It should be noted that one part of their decision may be that in the peak weeks there will be terrible scenes as people try to take loved ones to hospitals which are too full to accept them.

I think you are being rather unfair on Noel. He has done lots of research and is trying to understand from his own perspective whether the sacrifice being asked is worth it.

the model derives it from the change in cases a week ahead which all makes sense. The imperial group thinks Rt is 0.7 in the uk although this model disagrees but that may be because we are changing our testing scope as more capacity is available.