Scouter's Ebola Projection Update - 10/4/2014 - (Vanity)

Scouter | 10/4/2014 | Scouter

[scouter]

I have received a number of FReepmails asking for an update on my projections, in light of recent news. Here it is.

Please read and understand the following notes:

1. Because the rate of transmission differs in each country, the model now makes individual projections based on a Daily Transmission Rate (DTR) calculated for each country, and adds them together to produce the numbers presented below.
   
   
2. While the model can use any date range, I've used the the range from 9/1/14 to 10/1/14. In other words, the DTR for each country is based on the number of cases on September 1 vs. the number of cases on October 1. Previous calculations were made based on the rate of transmission from June 1 through August 31.
   
   
3. It appears that the DTR for each country has declined in recent weeks. This is good, but not good enough. The numbers are still scary.
   
   
4. The numbers presented below are based on reported cases and assume that 100% of the actual cases have been reported. We know this is not true. It is somewhat worse than this. Epidemiologists close to the situation believe the actual number of cases is 2-4 times higher. This would have a corresponding effect on the projections.
   
   
5. The model does not yet use a "governor" to limit the number of cases in a given area. In other words, the model projects that in Liberia alone, there would be over 174,000,000 cases by the end of October, 2015. Obviously that can't happen because Liberia doesn't have anywhere near that many people. Future versions of the model will include the "governor function" used by the CDC for this purpose. But according to epidemiologists, the "governor function" wouldn't have a substantial effect on the projections until the projections reach about 50% of the population of the country. So the shorter term projections are more accurate than the longer term. Obviously.
   
   

Scouter Ebola Projection Model Version 2.1 - Ebola Case Projections

*********************************************************
Projection Parameters
*********************************************************
Country: All Countries
Run Date/Time: 10/04/2014 at 12:32:03
Model: DTR Model
Start Date: 9/1/2014
End Date: 10/1/2014
Reported cases represent 100% of the true epidemic size
New Cases per Day at End of Period: 172
Rate of Increase per Day: 2.35%
*********************************************************

Weekly for the Next 8 Weeks

Date        Date                  Cases                 Deaths                Daily New Cases       
==========  ====================  ====================  ====================  ====================
10/01/2014                 7,491                 3,439                   175                    71
10/08/2014                 8,838                 4,058                   206                    95
10/15/2014                10,437                 4,794                   246                   113
10/22/2014                12,338                 5,671                   292                   135
10/29/2014                14,598                 6,716                   347                   161
11/05/2014                17,287                 7,961                   413                   192
11/12/2014                20,489                 9,448                   492                   229
11/19/2014                24,303                11,222                   587                   273


End of Month for the Next Year from the End Date

Date        Date                  Cases                 Deaths                Daily New Cases       
==========  ====================  ====================  ====================  ====================
10/31/2014                15,320                 7,051                   277                   110
11/30/2014                33,146                15,547                   844                   405
12/31/2014                73,924                35,305                 1,893                   923
01/31/2015               165,567                80,387                 4,261                 2,109
02/28/2015               344,116               169,366                 8,892                 4,453
03/31/2015               776,114               387,249                20,136                10,205
04/30/2015             1,710,309               863,619                44,533                22,805
05/31/2015             3,880,621             1,981,314               101,380                52,413
06/30/2015             8,596,752             4,431,396               225,240               117,407
07/31/2015            19,601,541            10,193,372               514,950               270,445
08/31/2015            44,788,229            23,473,373             1,179,439               623,547
09/30/2015            99,825,291            52,668,275             2,634,170             1,400,539
10/31/2015           228,894,782           121,502,640             6,051,399             3,234,010

[Veto!]

Do you know whether the vomiting passenger is African?

Someone who flew to Brussels from Liberia a week ago?

[JoeFromSidney]

What model are you using? Looks like exponential trend. I've done a lot of forecasting, and were I doing this I'd use a Logistic for the forecast. It takes into account the number of existing cases as well as the number of people not yet infected. That would handle the "governor" function, since the pool of uninfected is the difference between the infected and the total population of a country, and declines as the number of infected cases rises. The differential equation involve the product of existing cases and not-yet-infected population.

[Ray76]

Ebola Cases Surpass 8,000
October 8, 2014

http://www.freerepublic.com/focus/news/3213029/posts?page=1

[Brown Deer]

http://www.washingtonpost.com/national/health-science/the-ominous-math-of-the-ebola-epidemic/2014/10/09/3cad9e76-4fb2-11e4-8c24-487e92bc997b_story.html

[Brown Deer]

UN envoy: Ebola cases doubling every 3-4 weeks