|
Post by cjm on Mar 23, 2020 7:44:17 GMT
This is but the second day of that happening, but let us hope.
|
|
|
Post by Trog on Mar 24, 2020 6:39:29 GMT
I think it is still way too early to derive meaningful conclusions from the data so far. There are too many variables.
Those in SA currently infected with the virus almost all got it from somewhere else, that is, they are people who travel internationally. This tells a lot of their demographics - they are reasonably well to do, they are healthy, they are probably fairly young, they live in wealthy suburbs. So the initial stats do not at all reflect, I think, the figures that will realize when this spreads to the more average South African demographic.
There are also many other aspects. The penetration of testing for instance - the more you test, the more people you are going to identify with the virus. At the moment, I do not think that testing procedures in South Africa are yet streamlined into a process where it delivers uniform data over time.
As it is, the figures are not reassuring. The simplest model will probably be that the amount of new infections is proportional to the number of current infections, and is constant. That makes sense and is probably close to what you have with a culture in a Petri dish. If you apply that to the figures so far, (24 March 2020, 08:33 - 402 total infections) then for today I expect another 170 new infections, and by Tuesday next week about 7000 total infections. The week thereafter 98000 with 1.3 million the week after that.
Of course, the infection rate is not constant, and will probably be much altered by the lock down procedures that are coming into effect. It will also change dramatically if this virus reaches a different demographic. This will have a profound effect on the above calculations, but the point I'm really making is that we have way too little data to speculate about this intelligently.
|
|
|
Post by Trog on Mar 24, 2020 8:03:37 GMT
|
|
|
Post by Trog on Mar 24, 2020 8:42:37 GMT
...If you apply that to the figures so far, (24 March 2020, 08:33 - 402 total infections) then for today I expect another 170 new infections, ... +152 in fact. South Africa is now the country with the 6th highest number of new cases in the world - more than China. (Although I'm sure that China are bullshitting their figures big-time).
|
|
|
Post by cjm on Mar 24, 2020 18:07:06 GMT
Some of the required data are straightforward. Is it however possible to ascertain from your proportion assumption the following:
Contact rate (number of people an infected person comes into contact with daily). My guess is that this is difficult/impossible.
Infection rate (how many (%) of those contact persons are infected). My guess is that this is also difficult and interrelated with the previous parameter.
The disease period one could probably leave at 6 days. I cannot get any concrete information about how long it is contagious.
Mortality rate. Perhaps 3% .
|
|
|
Post by Trog on Mar 25, 2020 8:02:58 GMT
I'm just at CT airport, at the moment. Came on last week Monday for 3 weeks, spent one day at work were we decided to close the office. So I've been camping out in the Guesthouse ever since, in case we needed to have some in-person meetings after all. So Monday I decided in the nick of time to change my flight back to today, so I'm going to be on just about the last local flight for some time. Looks as it I'm just about going to have the entire aeroplane for myself, too - there's hardly a soul about.
|
|
|
Post by cjm on Mar 25, 2020 8:12:51 GMT
I'm just at CT airport, at the moment. Came on last week Monday for 3 weeks, spent one day at work were we decided to close the office. So I've been camping out in the Guesthouse ever since, in case we needed to have some in-person meetings after all. So Monday I decided in the nick of time to change my flight back to today, so I'm going to be on just about the last local flight for some time. Looks as it I'm just about going to have the entire aeroplane for myself, too - there's hardly a soul about. Probably dead already.
|
|
|
Post by cjm on Mar 25, 2020 18:32:55 GMT
|
|
|
Post by Trog on Mar 27, 2020 4:25:31 GMT
Well, from the SA data so far I predict that by Thursday next week we'll be up to about 6,000 total cases, and to 40,000 the week after that. I do not expect the lock-down to have much (any) influence during the first week, since it takes about a week between infection and identification/diagnosis. If the total number of cases is less than 40,000 2 weeks from now, then one would be able to say that the lock-down has had an effect.
For South Africa, it seems to be spreading particularly fast - comparable to that of Spain. And we are not into winter yet. I'm not usually a pessimist, but I'm very weary of what might happen here. I'd not be surprised if SA becomes the hardest hit of all, by a huge margin. We have everything in place for a perfect storm - coming winter, widespread poverty, highest HIV-positive population in the world, one of the highest TB rates, non-existent public health care, non-existent public services generally, etc.
|
|
|
Post by cjm on Mar 27, 2020 6:43:08 GMT
The two worst areas are the Western Cape (Cape Town particularly) and Gauteng (I suspect JHB). Isolate those areas and you have progressed a great deal.
I think it is worth noting that there are no fatalities yet and that less than a 1000 cases have been unearthed by more than 15 000 tests. Also interesting is the infection rate per age group, with 41-50 being the highest. I assume they move about most.
|
|
|
Post by Trog on Mar 27, 2020 6:59:48 GMT
The two worst areas are the Western Cape (Cape Town particularly) and Gauteng (I suspect JHB). Isolate those areas and you have progressed a great deal.
I think it is worth noting that there are no fatalities yet and that less than a 1000 cases have been unearthed by more than 15 000 tests. Also interesting is the infection rate per age group, with 41-50 being the highest. I assume they move about most.
I think the reason that there are no fatalities so far is because those infected were probably youngish, healthy, and from upper-income groups, and precisely because there are less than a 1000 cases so far. By the time you get to 40,000 cases, whatever remains of health services will be crippled, and a much larger percentage of vulnerable people will be infected. In South Africa, statistical tracking will break down, there is no hope in hell that we will be able to do testing to the extent that will deliver accurate results.
|
|
|
Post by cjm on Mar 27, 2020 7:49:00 GMT
It suggests at least one parameter although it seems to roll the contact rate and the infection rate in the model into one. How could one obtain separate figures from the R0 parameter?
|
|
|
Post by Trog on Mar 27, 2020 8:40:10 GMT
So South Africa has had its first 2 fatalities.
|
|
|
Post by Trog on Mar 27, 2020 8:48:46 GMT
It suggests at least one parameter although it seems to roll the contact rate and the infection rate in the model into one. How could one obtain separate figures from the R0 parameter? I think that the Mathematics post actually have some explanation of how to panelbeat the various parameters widely reported into values one can plug into the 'model' webpage. If I can find time this weekend I'll try to see if I can come up with some realistic values.
|
|
|
Post by Trog on Mar 28, 2020 4:56:43 GMT
So South Africa has had its first 2 fatalities. Actually it did not. The one case was subsequently discovered to have not been infected with the virus at all, whereas the other was treated for the corona virus, but actually died of a pulmonary embolism - so that it is a bit of a stretch to attribute her death to covid-19.
|
|