Ebola could kill 90,000 in Liberia by year-end

[tetano]

Efforts to control the Ebola virus in Liberia must be quickly and dramatically scaled up or tens of thousands of people will die in the coming months, said a study published on Friday.

Nearly 4,900 deaths have been recorded across West Africa since the virus was first detected in Guinea in March, according to the World Health Organization. Liberia has been the hardest hit of the countries most affected by the virus, with 2,705 deaths and 4,665 recorded cases.

Without accelerated efforts to control the disease, Ebola will kill 90,000 in Liberia and infect 171,000 by mid-December, a study in medical journal The Lancet Infectious Diseases said.

http://www.infowars.com/ebola-could-kill-90000-in-liberia-by-year-end/

 

[tetano]

Re: Ebola could kill 90,000 in Liberia by year-end

The original article

 

[Vibrant62]

Re: Ebola could kill 90,000 in Liberia by year-end

They need to find a way of executing safe burials that allows some ritual or incorporation of culture - perhaps taking a photo, wrapping the santised body bags into a shroud by family members, or something along these lines to get people back into the treatment centers, which are currently being avoided due to new cremation only rules. Social anthropologists could be very helpful here perhaps?

Otherwise this model may very well become an underestimate.

 

[AlaskaDenise]

Re: Ebola could kill 90,000 in Liberia by year-end

They need to find a way of executing safe burials that allows some ritual or incorporation of culture - perhaps taking a photo, wrapping the santised body bags into a shroud by family members, or something along these lines .......... Social anthropologists could be very helpful here perhaps?

Otherwise this model may very well become an underestimate.

Yes - I posted something similar earlier - not only could a PC ritual reduce the impact of this epidemic, but if it was widely accepted, future outbreaks might be more easily controlled. Social anthropologists????....ask the people what they would find acceptable for now and in the future.

.

 

[gsgs]

Re: Ebola could kill 90,000 in Liberia by year-end

> We estimated the basic reproductive number for EVD in
> Montserrado to be 2.49

I don't know how they "calculated" .
But this is pretty absurd, and not supported by the data, IMO.

So, you have to wonder whether there is some agenda behind this ...
NIH telling them they should estimate this or such.

 

[JJackson]

Re: Ebola could kill 90,000 in Liberia by year-end

> We estimated the basic reproductive number for EVD in
> Montserrado to be 2.49

I don't know how they "calculated" .
But this is pretty absurd, and not supported by the data, IMO.

So, you have to wonder whether there is some agenda behind this ...
NIH telling them they should estimate this or such.

gs
Check out the supplemental data. They give their rational and the equations used to derive the 2.49 (95% CI 2.38-2.60). It is not a native Ro or outbreak Ro but a current Montserrado Rt. It is significantly higher than the outbreak Ro derived in the NEJM paper but they discuss that.
No conspiracy theory required. The NIH had no input into the study design or the decision to publish. Any way you can recalculate on any Ro value you like the importance of this paper lies in the relative intervention effects. Which ever way you cut it the key take-away is that we need a maximum effort on every front simultaneously if we are to get control of the outbreak's doubling time.
I hope you are right about the local Ro, it would increase our chances of winning this race.

 

[gsgs]

Re: Ebola could kill 90,000 in Liberia by year-end

I can't find the supplemental data
Copyright ? 2014 Elsevier Ltd All rights reserved.

I'm not sure whether I understand the Elsevier terms and conditions correctly.
http://www.elsevier.com/legal/elsevier-registered-user-agreement
Do they allow me to quote from the article, to discuss their methods here ?
I'm also not familiar with the laws of the State of New York.
They may have a reason to "protect" the article ...

Anyway, JJ, when you argue about their methods, their reputation, the journal
peer reviewers or such, then please be aware that we are just talking about
mathematical facts here. You can't argue about that, it's decided by logic alone.
The data are available, it's in the Liberia situation reports, where they report case-counts.
From this you calculate the R0, there is no freedom of methodology or interpretation to do this.
It's a clear mathematical process.

Unless they ignore the more recent data (which I suspect) or unless they pick probable
cases only or such.The consensus was always to use suspected+probable+confirmed
cases for Liberia.

-----------------------------------------

there is a clear agenda here. WHO,UNO,NIH are seeking for government
money to stop the outbreak, which as we probably agree is positive and
should be done. But they are doing it by _lying_ about the R0, the
expected cases by Dec., the exponential nature. And this includes
Chan,Ban ki Moon, Fukuda, Aylward,...
And then we are seeing multiple papers about it. I think they are _ordered_
to conclude it's exponential, that's why they get the grant in the first place.

And all the press repeats it, I haven't found anyone except me,
who questions the exponential nature, while it's sooo obvious.

 

[GardenSpider]

Re: Ebola could kill 90,000 in Liberia by year-end

After logging years and years in the field in Africa, recording oral histories and organizing "consciousness raising" teams that went out to villages in effort to ban Forced Marriage and FGM, I was left extremely frustrated. It was clear the older generation was very attached to enforcing FM and FGM. In one of our all women's sessions between the youth and elders tensions rose. It was hard to penetrate the conviction of the elders.

In one village where I had spent a great deal of time over the years in Mali, I offered to provide funding for primary and secondary education in exchange for the ban on FGM and FM. The elders didn't like it, but they accepted it because they understood education was valuable even if they didn't fully grasp how.

Only through financial support; hiring a project manager who checked up on the girls, building a women's health clinic and some providing perks for the chief and his family, were these practices finally abandoned. I maintained the program with my daughter for 10 years. Now that my funding dried up, I don't know whether or not the practices have resumed, but I assume they have even though Mali outlawed them a few years ago.

In the Ebola zones, I would also assume money talks- not introducing some new mythology. Introductions to new practices don't work. It has to be something grandparents pass on the their grandchildren.

Their traditions and practices run deep. And as I mentioned in a previous thread, in the most remote areas where education is lacking as well as a connection to their original way of life, villagers are vulnerable to profound ignorance and their ability to comprehend consequences is deeply compromised.

Studies have shown that children who do not learn to structure sentences using "if/then" can not grasp consequences. This is true in some of the poorest places in America where the study was conducted.

It's going to be a hard road in Africa.

 

[Ronan Kelly]

Re: Ebola could kill 90,000 in Liberia by year-end

Studies have shown that children who do not learn to structure sentences using "if/then" can not grasp consequences. This is true in some of the poorest places in America where the study was conducted.

I'm interested in that study - do you have a link or name?

 

[GardenSpider]

Re: Ebola could kill 90,000 in Liberia by year-end

I'm interested in that study - do you have a link or name?

I will hunt for the study. I read it a long time ago. It was conducted on American middle school children. The study showed that children who did not know conjunctive adverbs (i.e. subsequently, however, thereafter) could not understand the concepts they illustrate. Because these words have no synonyms, it was found that people who do not understand them can not access the type of thinking that they produce. Once the middle schoolers were taught these words their abilities to deduce, write papers and think critically went from one of the worst in America to a much higher percentile.

The ability to reason is not innate to human beings. It is an acquired skill and without the correct tools in language people can not unlock even basic reasoning.

 

[AlaskaDenise]

Re: Ebola could kill 90,000 in Liberia by year-end

This google book may contain references to the study you read, & it also discusses the subject. See: The Mental Models Theory of Reasoning: Refinements and Extensions

.

 

[gsgs]

Re: Ebola could kill 90,000 in Liberia by year-end

here is the curve that they assume is exponential,
that they approximate by an exponential function
and from which then they are trying to predict future values

it can't fit

cumulative ebola cases on Mondays in Montserrado, all and HCW, weeks 24,...,42
10,1,12,2,29,5,41,5,57,6,74,7,102,11,137,20,170,33,256,38,405,45,541,57,707,69,1059,71,
1282,73,1490,74,1659,84,1814,89,2136,96

--------edit----------
> The Director of Health Communication, John Sumo, in an Ebola update issued in Monrovia,
> the nation's capital city said between Oct. 17 and Oct. 18, 98 Ebola patients were
> discharged after testing negative for the virus.

that may explain the reported 91 suspected cases on Oct.17 and missing 7 suspected cases on Oct.18
and just the peak in above curve in the last week.

 

[gsgs]

Re: Ebola could kill 90,000 in Liberia by year-end

this study is now all over the headlines, just search google-news for montserrado :

Ebola Outbreak: 90000 Deaths Predicted in Liberia's Montserrado ...
Response to Ebola epidemic in west Africa too slow, say scientists
Ebola control measures in Montserrado, Liberia 'not enough to ...
Ebola death toll likely to exceed 90000 by December in Montserrado ...
Study: Current Aid Promises Won't Contain Liberia's Ebola Outbreak
Ebola epidemic will explode by mid-December and kill tens of ...
Ebola epidemic likely to explode by mid-December: Study
Yale researchers forebode "Ebola epidemic explosion may take ...
'Catastrophic' W. Africa Ebola Epidemic Possible
Investors Should Pay Attention To Ebola - It Is Getting Worse
[just the first 10 headlines out of ~200 from major newspapers and magazines]


is it that what NIH wanted, why they did it ?
Of course, those reporters don't understand that it's highly speculative,
that the study was not about estimating cases or deaths, but rather
they _assumed_ some exponential growth and then considered how certain
measures might change the numbers.
There is no discussion, how well the assumed growth fits the reality,
they just approximate the weekly data since June by some exponential-like
curve, ignoring the non-matching data points since mid-September.

Now, as we are going into November, will they please surveille and adapt their
"estimates", as new data is coming in ? Or do they want us to keep with
these headlines, that are then being quoted again and again although
they will soon (and are already since at least Oct.05) be old and no longer valid ?

 

[JJackson]

Re: Ebola could kill 90,000 in Liberia by year-end

gs
I think we need to separate out the NIH, the paper and the press coverage.
I have no knowledge of the thinking of the parties involved but suspect the NIH wanted some basis to understand the relative efficacy of the measures being undertaken to slow transmission.
As the measures are only now being instituted the data is not yet available to determine what works so they funded a modelling team to help in guiding the focusing of interventions.
The modelers did what modelers do and ran lots of simulations to try and get a feel for which interventions, or combination of interventions, were most effective and where there were synergies.
The press did what they do and, sadly in many cases, picked up on some of the big numbers without grasping the purpose of the exercise.
I am not sure that the exact shape of the current, relatively lightly controlled, growth curve is particularly important to this exercise. If the aim is to see what works best against a growing case count then the findings should continue to be valid regardless of the exact Ro.
The numbers are a mess and I would be delighted if the true numbers showed a slowing in the doubling time but I think that would be a dangerous assumption to make at this point. If it turned out to be wrong, and caused a slackening of containment efforts, then gaining control of the situation would be made even more difficult.
There are lots of wild guesses in the paper which are more troubling than the Ro estimate. For instance they modeled the effectiveness of the home PPE kits with values between 10 & 50% but this was based on two previous uses of home PPE in Uganda and the Congo one of 12% & one at 88% and both had confidence limits that rendered the results close to meaningless. Hopefully there will be epidemiologists following the effects of the interventions so we can place more meaningful bounds on some of these numbers and so improve the usefulness of this type of model.
Again just my personal take on the paper.

 

[gsgs]

Re: Ebola could kill 90,000 in Liberia by year-end

jj, the exact shape of the curve is what makes the numbers that are being quoted.
When they use exactly the same methods, these numbers will change
from week to week quickly. Will they show those new updates ?

You could at least expect that they discuss, why the shape should be exponential
or even clearly mention to the press that they just _assume_ it's exponential.
It may have looked quite exponential until mid-Sep, when several papers
came out about this, the most prominent one, widely quoted as you may remember,
"predicted" 1.4 million cases by January.
None of these had been updated since then, and the new papers quote the old
conclusions to support their new conclusions.

> The numbers are a mess and I would be delighted if the true numbers showed a slowing
> in the doubling time

the "true numbers" are from the same source as those that they had in the paper.
And they _clearly_ show a decline since mid-Sep, not just a "slowing in the doubling time"

> but I think that would be a dangerous assumption to make at this point.
> If it turned out to be wrong, and caused a slackening of containment efforts, then gaining
> control of the situation would be made even more difficult.

you mean, we should _deliberately_ use worse data than the real one (without telling the audience...)
because underestimates are more dangerous than overestimates ?
You'd lose credibility quickly.
That credibility that might be needed urgently, if things get as bad as they "predict".

 

[JJackson]

Re: Ebola could kill 90,000 in Liberia by year-end

jj, the exact shape of the curve is what makes the numbers that are being quoted.
When they use exactly the same methods, these numbers will change
from week to week quickly. Will they show those new updates ?

Will they show updates, I doubt it as it was a one off paper.
The difference between their worst and best case scenarios (both of which are very unrealistic) is over 130,000 but it is difficult to extract case numbers as the whole paper looks at cases averted rather than case.

You could at least expect that they discuss, why the shape should be exponential or even clearly mention to the press that they just _assume_ it's exponential. It may have looked quite exponential until mid-Sep, when several papers came out about this, the most prominent one, widely quoted as you may remember, "predicted" 1.4 million cases by January.
None of these had been updated since then, and the new papers quote the old conclusions to support their new conclusions.

Didn't the 1.4 million come from a CDC sheet that just plotted a best fit growth curve and extrapolated based on no additional mitigation effort? Like you I have been trying to make some sense of the published numbers but once bed capacity became saturated they became increasingly suspect as too many people 'seemed' to be outside the system. More recently the big MSF elwa facility has not been full but anecdotal evidence points to it not being due to a lack of patients but possibly fear of enforced cremation rather than burial. Time should tell if we really do have more beds than patients but it seems unlikely at the moment.
I know you listened to the Twiv with Jeff Sharman as a guest as you posted a very helpful segment break-down. Did you have a look at his model. It is interesting as it has a 'tick box' which allows you to run the plot with the predictions at each time point and compare them with what actually happened. The 'un-ticked' graph shows all the latest data, and what the model predicts going forward, but the 'ticked' data shows each of the predicted curves as they were prior to the new data becoming available.
Could the modelers have explained their thinking to the press, yes but I doubt many would have bother to ask or printed any answer. Nothing is likely to turn of your readers faster than a article discussing the relative merits of different modelling scenarios.

you mean, we should _deliberately_ use worse data than the real one (without telling the audience...)
because underestimates are more dangerous than overestimates ?
You'd lose credibility quickly.
That credibility that might be needed urgently, if things get as bad as they "predict".

No the numbers have to be a best guess as to the real situation and their limitations should be explained as far as they are known. In this case most of the authors were from Yale with 2 from the MoH in Monrovia so I can only trust they did the best they could with what ever data they had. For this outbreak I would like to see modelling done on the best available data and containment efforts should be based on the worst case scenario end of model spectrum on the grounds it is better to deploy too many resources, even if they get shipped back unused, than find things were worse than we thought and go back to playing catch-up.
We underestimated the opposition for the first half of 2014 and now have a lot of catching up to do. We can not afford to make the same error again or we will lose geographic containment.

 

[gsgs]

Re: Ebola could kill 90,000 in Liberia by year-end

here comes the next modeling study:

http://www.sciencemag.org/content/early/2014/10/29/science.1260612.full

Under intervention efforts implemented through September 19, our
model predicts that the epidemic will continue to spread, generating a
predicted 224 (134?358) daily cases by December 1; 280 (184?441) by
Dec 15 and 348 (249?545) by Dec 30.

... To calibrate our model, we obtained data from Situation Reports provided by the Liberian
Ministry of Health and Social Welfare (12) (table S1 in SOM
http://www.sciencemag.org/content/suppl/2014/10/29/science.1260612.DC1/Pandey-SM.pdf
[and there they write:]
We fit our model to data from the Liberian Ministry of Health and Social Wel-
fare situation reports from 8 June to 7 August, 2014 inclusive

[Aug07 only !, the model should be fit to all available data !]

'The close agreement between the epidemiological data and the temporal
trajectory of our model between August 8 and September 19 suggests
that the cordon sanitaire and curfews implemented in that period (13)
have had little impact on the exponential increase of Ebola prior to Sep-
tember 19 (Fig. 1)

[but it started to decrease since then]

Although recent cumulative case data
over October 2014 from Liberia suggest that the exponential growth that
has characterized the epidemic for both communities and healthcare
workers may be slowing down

[indeed, it looks no longer exponential]

 

[Sharky]

Re: Ebola could kill 90,000 in Liberia by year-end

Although recent cumulative case data
over October 2014 from Liberia suggest that the exponential growth that
has characterized the epidemic for both communities and healthcare
workers may be slowing down

[indeed, it looks no longer exponential]

Although the reported data no longer looks exponential, there are many reasons to believe there is a large and growing disconnect between that data and reality. (field reports, for example).

 

[gsgs]

Re: Ebola could kill 90,000 in Liberia by year-end

field reports are only local.
The total numbers describe the whole problem.
And in all the locations.

And with growing numbers the data becomes more reliable.

In what region is it still growing exponentially ?
Or can there be a conspiracy, everyone systematically
increasing underreporting ...

And why the agenda, trying to paint a dark picture ?

Using tricks like cumulative data instead of weekly numbers,
why all the papers with old data, the "exponential" statements
by Ban ki Moon and Chan etc. and _noone_ trying to
do an actual analysis ? Not even discussing that that the trend
could be no longer exponential since mid-Sep (I found no newspaper,magazine to do this)

Instead these apocalyptic horror prediction numbers are apparently widely accepted.
But somehow noone cares. Instead any one new case in USA gives a
big bruhaha.

 

[MHSC]

Re: Ebola could kill 90,000 in Liberia by year-end

The reporters in Africa are not being allowed to report freely anymore so we really don't know what is happening there. As far as the United States, we have a lot more reporters and newspapers following each case, so it may appear to be a big bruhahaha but it's really not, it's just the sheer number of reporters we have.