Sunday, September 7, 2014

Ebola 3

Roz just made this comment on the previous post:

          hey Greenpa - is this chart a little like what you were looking for? It's a bit scary, it is.

         BBC- Ebola-How bad can it get?


  Good catch, Roz; I was just getting ready to make a post here about that exact article.

The graph is not what I want to see; but this is the presentation that made "that other guy" go all wacky and start saying "OMG, Ebola has gone asymptotic!"

This looks so scary because they are adding all deaths in the epidemic together to make the next point on the graph.  That's not an entirely illegitimate way to look at the numbers; but from the epidemiology standpoint, it is not as informative as a graph tracking "number of new cases this week" or "number of new deaths this week" would be.  

Those graphs give an appearance that is much less accessible to the general public - they're very jagged, and the "trend" is harder to see; and most of the jaggedness is purely accidental; caused by differences in reporting, not differences in actual infection/mortality.

Last I really looked at the numbers, the data for "new cases" looked something like "43, 17, 30, 59, 28, 37, 19, 67"  etc.   The question is "is the disease accelerating",  not "totals".  Still have not seen the data presented that way.  There are standard ways to draw a line for "best fit" that evens out the jags.  If you have a straight, but up-slanted line; that means the epidemic is still speeding up, and that's not great.  It's when that  line goes asymptotic- that it's time to be very, very scared.  Not happening yet, I think.  The WHO was reporting "new" cases in mid August; but they're not, now.  Not sure why.

That BBC article - is notable for 3 things; basically all good, actually.

 1)   It's the first "scary" article in the major press I've seen - telling the truth about the epidemic.  Yes, it's potentially dangerous, to the entire world.

  2) The author did a good job of talking to researchers, and translating for the public.  That's been very uncommon on the BBC in the past few years, and getting worse; but this one is excellent.

  3)  Some of the researchers raised good points I had not thought of in particular, which could easily lead to epidemic outcomes that are less than apocalyptic.  That would be nice.  In particular, the scenario where the virus mutates to "more transmissible"; which could then lead to evolving to "less lethal" more quickly than if it stays as it is.  Reaching a point where it never goes away, but subsides to normal "bad disease" status, where it is all around the world; it still makes people very sick during outbreaks, but maybe only 5% of infected people die.  

That's actually a fairly high probability in the evolution of pathogens scenarios.  I was focusing on the vastly increased chances for it to become easier to catch; now that there are uncountably more virions available for evolution to act on.  That's very scary.  But yes; the next evolutionary tactic is usually to: stop killing your host.

I recommend folks read that article.  Twice; once today, and again in a couple days.  Lots of information to absorb there, and it's far more honest than other stuff still being circulated.

There are several other hopeful developments too; like the preliminary finding that there may be many people in the region who are already immune to Ebola; for reasons they're guessing at.  Fewer susceptible people would be a huge help.  And; those fighting the disease think they may be able to use blood from those who managed to survive the infection to treat active patients.  That would sure help.

All in all; keep paying attention.  But total imminent world collapse is looking a little less likely - from Ebola, anyway.


Roz said...

ok, thanks for explaining the graph better! I think the potential immunity factor is probably likely and good news, but I did also notice they have stopped reporting new cases now hmmm. I'll keep watching

Anonymous said...

About graphs... Over on the Wiki page they have all of the WHO data along with graphs of the total cases and deaths. They also have the log version. The point of doing the log is that the slope is a "growth rate." The other thing about exponentials is that their derivatives (new cases)look just like the function itself. Derivatives are naturally noisy, so I see no reason NOT to look at the totals because they perform the "averaging" function. I hate to say it, but the growth rate looks pretty constant since the middle of May. Of course at some point it will have to break from the constant growth trend. Right now it seems about x10 every three months. At that rate, it could infect everyone in West Africa in 18 months. So we are all watching to see when that curve starts to break. It is hard to see why we should expect it to stop at 20,000 at this point,however. With the medical teams on the ground reporting chaos and insuffecient resources, it seems that transmission rates are unlikely to go down any time soon.

Greenpa said...

Squash - yep. Good to find someone with a functional sense of statistics. I hadn't looked at the Wikipedia page since my first Ebola post - wow; it's exploded; and their information is far more inclusive than any other source I can find. I continue to be impressed with Wikipedia.

They do indeed have the graph of new cases; and it does indeed look worrisome; with a distinct up-curve in recent days, which is what I said would worry me... See my next post...