Posted on 09/19/2014 8:46:26 AM PDT by scouter
Before starting, let me say that I do NOT believe the Ebola epidemic will get as bad as these projections indicate. I believe it will get pretty bad, especially for those in Africa, but not as bad as the current numbers say. I think there will be a number of factors that will significantly mitigate these numbers. How much? I don't know. But no one else does, either. I am not predicting the end of humanity. I'm simply showing where the current numbers lead, should nothing change. But of course, things will change. In any case, I do believe we need to take action now to prepare for the worst, hope for the best, and pray for the those affected.
The following information is presented for the reader's information and discussion. It it not a prediction of future events. As I mentioned when I originally posted my projections, my motivation is out of professional curiosity, the fact that my daughter will be on the front lines if Ebola does appear in the U.S., and because I think the numbers indicate that it is time to take prudent prepatory measures.
Another FReeper, DannyTN, has convinced me that version 1.0 of my model is too complicated, and that one relying more on the concepts of simple compound interest is likely to be more accurate. I've created the Scouter Ebola Projection Model Version 2.0. It has changed in the following ways:
1. Rather than calculating an exponent to apply to the number of cases on a given date, I now use a simple "compound interest" approach in which the Daily Transmission Rate (DTR) is determined using the Microsoft Excel Rate function, as follows:
DTR = RATE(NumDays, 0, NumStartCases, -NumEndCases)
where NumDays is the number of days from the start of the period being examined to the end of the period being examined, NumStartCases is the number of cases on the first day of the period, and NumEndCases is the number of cases on the last day of the period.
The DTR is then applied to the future, using the EndDate of the period selected as the first day of the "future", using the following formula:
ProjectedCases = StartingCases * (1 + DTR)^NumDays
where StartingCases is the number of cases on the End Date of the period being examined, DTR is the Daily Transmission Rate calculated above, and NumDays is the number of days from the End Date.
This allows me to examine how accurate it is by using past performance to project into the "future" and compare that to what actually happened. Using June's numbers, for example, projects that there will be 1,604 cases on August 1. In fact, there were 1,603. Pretty darn good. Using July's numbers projects that there will be 2,971 cases on September 1. There were 3,707. So it underestimated by 736 cases, or about 20% of the actual number. This is probably due to improved reporting. In other words, the number of reported cases in July was probably low.
2. The original model used the case report data as published on the Wikipedia article Ebola virus epidemic in West Africa. I have verified and changed the more recent values, and added additional values, based on my own research. The references for the data will be provided in the spreadsheet when I publish it.
3. I now calculate a separate Daily Transmission Rate for each date for which there is published data. It is calculated for the period of time between that date and the previous date for which there is data.
4. I've included some graphs, displayed below.
5. I now calculate the number of new cases and deaths on the last day of the period being projected.
Many readers have commented on several points that must be kept in mind when interpreting these numbers, and they need to be enumerated here.
1. Obviously the epidemic will not kill trillions of people. But if you project far enough into the future, that's what it will tell you. That's because I have not yet included anything to limit the number of cases. I'm working on that. But in the near case (out until 6 months or so), I don't see that as being a significant factor. But clearly, if the epidemic develops as these numbers suggest, there will come a point when the number of uninfected people in any given area will become significantly smaller, due to death and possible immunity, which will cause the rate of transmission to slow naturally.
2. Some credible epidemiologists and virologists are now saying that it may, in fact, be possible for Ebola to be transmitted through the air, without direct contact with the bodily fluids of an infected person. See COMMENTARY: Health workers need optimal respiratory protection for Ebola at the Center for Infectious Disease Research and Policy.
3. Besides the issues mentioned above, there are many, many factors that can and will affect the rate at which the epidemic is spreading. Some of these will increase the rate, and some will decrease the rate. We don't know how it is going to play out. This is why I'm trying to find a good way to incorporate the change in the Daily Transmission Rate (DTR) in future projections, and why I'm including graphs showing the change in the DTR over time, including trending lines.
4. There are various arguments for which time period to use for projecting into the future. One can argue that the longest interval for which we have data should be used because that smooths out the variability in the Daily Transmission Rate the most. Another argument is that a shorter interval is best because it doesn't matter how we got to the current numbers. Rather, what's important is how fast the epidemic is spreading now. For the numbers presented below, I've chosen a path between those two extremes. The period of time on which I based the DTR for the numbers below is August 1 through September 14.
5. Experts with front line knowledge of the current epidemic have testified before Congress that the reported numbers only represent 25% to 50% of the true size of the epidemic.
6. I do NOT account for variability in the accuracy or completeness of reporting, or for the possibility of bias in reporting. Garbage in, garbage out.
7. I do NOT include the cases of Ebola in a separate, unrelated outbreak in Congo.
I hope to publish the spreadsheet so you can make your own projections based on the parameters you are interested in. I've worked out how to sanitize it and host it. But I'll have to pretty it up first.
So here are the revised numbers, based on the most recent data and my revisions. It assumes that all actual cases have been reported. Graphs follow the numbers.
Scouter Ebola Projection Model Version 2.0 - Ebola Case Projections
*********************************************************
Projection Parameters
*********************************************************
Run Date/Time: 09/18/2014 at 23:02:18
Model: DTR Model
Start Date: 8/1/2014
End Date: 9/14/2014
Reported cases represent 100% of the true epidemic size
Rate of Increase per Day: 2.81%
*********************************************************
Weekly for the Next 8 Weeks
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
09/14/2014 5,418 2,589 148 71
09/21/2014 6,576 3,143 180 86
09/28/2014 7,982 3,814 218 104
10/05/2014 9,689 4,630 264 126
10/12/2014 11,760 5,620 321 153
10/19/2014 14,274 6,821 390 186
10/26/2014 17,326 8,279 473 226
11/02/2014 21,030 10,049 574 274
End of Month for the Next 2 Years from the End Date
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
09/30/2014 8,437 4,031 230 110
10/31/2014 19,898 9,508 543 260
11/30/2014 45,648 21,813 1,246 595
12/31/2014 107,662 51,446 2,939 1,404
01/31/2015 253,922 121,337 6,932 3,312
02/28/2015 551,157 263,371 15,046 7,190
03/31/2015 1,299,911 621,165 35,486 16,957
04/30/2015 2,982,161 1,425,030 81,410 38,902
05/31/2015 7,033,459 3,360,950 192,005 91,750
06/30/2015 16,135,646 7,710,444 440,485 210,486
Note: The negative "increase" in the above chart is due to reporting corrections made in the data by the reporting agencies. The straight lines indicate the trends in Daily Transmission Rate and Fatality Rate.
To me the key is Logistical Support. If the infection rate and death rate rises to a point in which it over whelms the the systems Medical/Logistical support capabilities then the ‘genie’ is truly out of the bottle.
A truly serious as well as unknown factor will be people reactions and how much of government services, manpower and logistical support might have to be diverted to handle panic arising from such.
We are seeing this in Africa right now if it reaches a point in which it becomes unsafe or impossible for medical teams securely preform their duties the spread of the disease could quick start spreading virtually unchecked.
Most of the homes in America are probably superior to some of the primitive field clinics run by Africans.
Some of us forget that a typical American home is full of porcelain and stainless steel surfaces, a laundry room, tubs and showers, screened windows, scalding hot water, bleach, and antiseptics, air conditioning and communications, and people who can understand and follow basic medical instructions, and that will eagerly seek more knowledge and solutions to such medical threats to their families.
Our people eat vitamins and walk through drugstores and pharmacy and cleanser and disinfectant sections of grocery stores that are extraordinary, yet just a part of our daily shopping for bread and milk.
We are many stages ahead of where the Africans are. Even in mass catastrophes we have huge modern buildings to move people, to warehouse them and still have modern plumbing, and electricity and easy access and such, we even have vast amounts of plastic and construction goods and Home Depot materials and bedding to set up makeshift clinics in warehouses and other buildings, that again, would be the envy of African medical people. We even have the ability to move resources vast distances almost instantly, for example from Seattle to Chicago if needed.
None of this cures Ebola, but it sure is a different reality than what those poor saps in Africa face.
You mean like this?
Just a reminder: Soap breaks down viruses, but the last time I checked, anti-bacterial hand sanitizers did not break down viruses.
More bad news: The natural rate of epidemic growth varies from country to country. If this model separated the countries, the numbers would be even worse - FAR worse!
It was kind of a “joke post” meant to point out that our hygene standards are a bit higher and it seems to matter.
Yup.
If you can plot the projections on a semi-log scale, it will be easier to see where you see the disease upshifting to a higher gear.
It would not take long, for instance, for the sanitation workers to figure out they could be getting exposed to the disease and just not show up for work. We have seen how well a garbage worker's strike works in an urban area.
That, however, is just one of many sectors where employees might decide the paycheck is no longer worth the risk, and the absence of or undermanning of those sectors could lead to a breakdown of common services and a loss of civility which has the potential to spread far beyond areas directly affected by the disease, especially with the current social climate in the US having been whipped to a frenzy by recent events and the media.
I think the same quandry applies here as there: at what point do those who enforce a quarantine kill those seeking to escape it out of fear for their lives?
The very real possibilities get very ugly very fast as food, fuel, and other shipments are delayed or stop due to travel restrictions, hazards, or no-go areas.
The system is more fragile than many believe.
I have the data. I'll work on it.
I'll work on it.
There was a Model T and an MG rally going through the same time I was. The winding roads kept me awake. And when I went down a dirt road through that area I saw one of the largest bucks standing in the road, along side a big turkey.
I think I will come to your house instead.....
“The system is more fragile than many believe.”
What Retail Might Look Like at TEOTWAWKI
http://freerepublic.com/focus/f-bloggers/3205100/posts
It is a just in time world!
We sometimes get both deer and turkeys milling around in the morning outside our bedroom window. Never at the same time though.
What happens if people do place themselves in self imposed quarantine? Who operates the system then?
Your model at this point is confined to rural West Africa.
Volunteers at best, it falls apart at worst.
I understood the joke (and I was amused). The image just reminded me of something worth pointing out.
However....it can’t be dismissed that only twice before has the Security Council met to discuss the security implications of a public health issue – both times on the AIDS epidemic...and now Ebola....all relating to Africa.
I've broken the data down by country for those countries with intense outbreaks.
Relative to the other countries, Guinea isn't in that bad of shape. I was able to project out until June 30, 2015 without the numbers getting astronomical.
Scouter Ebola Projection Model Version 2.0 - Ebola Case Projections
*********************************************************
Projection Parameters
*********************************************************
Country: Guinea
Run Date/Time: 09/19/2014 at 14:46:33
Model: DTR Model
Start Date: 8/1/2014
End Date: 9/14/2014
Reported cases represent 100% of the true epidemic size
New Cases per Day at End of Period: 14
Rate of Increase per Day: 1.52%
*********************************************************
Weekly for the Next 8 Weeks
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
09/14/2014 942 467 14 7
09/21/2014 1,047 519 16 8
09/28/2014 1,164 577 17 9
10/05/2014 1,293 641 19 10
10/12/2014 1,437 713 22 11
10/19/2014 1,597 792 24 12
10/26/2014 1,775 880 27 13
11/02/2014 1,973 978 30 15
End of Month for the Next 2 Years from the End Date
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
09/30/2014 1,199 595 18 9
10/31/2014 1,914 949 29 14
11/30/2014 3,010 1,492 45 22
12/31/2014 4,805 2,382 72 36
01/31/2015 7,671 3,803 115 57
02/28/2015 11,703 5,802 175 87
03/31/2015 18,683 9,262 280 139
04/30/2015 29,378 14,564 440 218
05/31/2015 46,897 23,249 702 348
06/30/2015 73,743 36,558 1,104 547
Liberia's got problems. The numbers get so astronomical so quickly that it only made sense to graph the projections out to the end of this year, although I provide the numbers until the end of February. The problem is that the Daily Transmission Rate (DTR) is currently 4.072%. Clearly that's unsustainable, but that's what it is. Something's gotta give soon there.
Scouter Ebola Projection Model Version 2.0 - Ebola Case Projections
*********************************************************
Projection Parameters
*********************************************************
Country: Liberia
Run Date/Time: 09/19/2014 at 14:38:12
Model: DTR Model
Start Date: 8/1/2014
End Date: 9/14/2014
Reported cases represent 100% of the true epidemic size
New Cases per Day at End of Period: 106
Rate of Increase per Day: 4.07%
*********************************************************
Weekly for the Next 8 Weeks
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
09/14/2014 2,710 467 106 18
09/21/2014 3,584 618 140 24
09/28/2014 4,739 817 185 32
10/05/2014 6,266 1,080 245 42
10/12/2014 8,286 1,428 324 56
10/19/2014 10,957 1,888 429 74
10/26/2014 14,489 2,497 567 98
11/02/2014 19,159 3,302 750 129
End of Month for the Next 2 Years from the End Date
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
09/30/2014 5,132 884 201 35
10/31/2014 17,689 3,048 692 119
11/30/2014 58,578 10,094 2,292 395
12/31/2014 201,887 34,790 7,900 1,361
01/31/2015 695,796 119,903 27,225 4,692
02/28/2015 2,127,411 366,606 83,242 14,345
Sierra Leone is worse than Guinea, but not as bad as Liberia. Still, it only mad sent to graph the projections out to the end of April. I provide the numbers, though, through June 30, 2015.
Scouter Ebola Projection Model Version 2.0 - Ebola Case Projections
*********************************************************
Projection Parameters
*********************************************************
Country: Sierra Leone
Run Date/Time: 09/19/2014 at 14:50:15
Model: DTR Model
Start Date: 8/1/2014
End Date: 9/14/2014
Reported cases represent 100% of the true epidemic size
New Cases per Day at End of Period: 36
Rate of Increase per Day: 2.19%
*********************************************************
Weekly for the Next 8 Weeks
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
09/14/2014 1,673 467 36 10
09/21/2014 1,946 543 42 12
09/28/2014 2,265 632 48 14
10/05/2014 2,635 735 56 16
10/12/2014 3,065 856 66 18
10/19/2014 3,566 996 76 21
10/26/2014 4,149 1,158 89 25
11/02/2014 4,827 1,348 103 29
End of Month for the Next 2 Years from the End Date
Date Cases Deaths Daily New Cases Daily New Deaths
========== ==================== ==================== ==================== ====================
09/30/2014 2,365 660 51 14
10/31/2014 4,623 1,290 99 28
11/30/2014 8,845 2,469 189 53
12/31/2014 17,293 4,827 370 103
01/31/2015 33,809 9,437 723 202
02/28/2015 61,947 17,292 1,325 370
03/31/2015 121,110 33,807 2,591 723
04/30/2015 231,714 64,680 4,957 1,384
05/31/2015 453,017 126,455 9,692 2,705
06/30/2015 866,733 241,939 18,543 5,176
Scouter’s revised Ebola projection bookmark
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.