This is my second post about Ebola, in which I continue to pontificate about things I don’t really know anything about. Yesterday I suggested Ebola is the Kardashian of diseases and we ought to be a little less worried about Ebola and talk more about HIV, malaria, and TB.
There are some excellent graphs and discussion from Kim Yi Dionne and Stéphane Helleringer, who do know what they are talking about.
Africa is a continent – not a country. If a health problem is only prevalent and problematic in one country rather than in many of Africa’s 54 countries, does that make it as irrelevant as a Kardashian? Continent-wide metrics can mask dramatic impacts of disease outbreaks in countries or even sub-regions. If we consider Ebola in the context where it’s unfolding, it matters a great deal. By the end of 2014, it may matter even more in these countries than the other infectious diseases mentioned by Blattman.
…So, how many deaths from Ebola can we expect by year’s end under these conditions? To try and answer this question, we did some simple arithmetic: at a rate of 18 deaths per day (i.e., the average number reported for August so far in Liberia), it will only be 10 days until Ebola has killed as many people as road-traffic accidents usually kill in the country in an entire year. It will be 20 days until Ebola reaches the yearly level of maternal deaths; 70 days until it reaches the number of deaths from HIV/AIDS deaths and 125 days (i.e., before the end of 2014) until it reaches the estimated annual number of malaria deaths in the country. And of course, these “projections” rest on the very optimistic assumption that the public health response will be able to maintain the number of deaths due to Ebola at 18 per day. In recent days however, the daily number of deaths seems to have been rising quite sharply over time.
This is a good response. The full post is worth reading. A few thoughts.
- I care unusually about Liberia, having worked there for the past six years and seeing the toll this disease is taking on the country. But it seems to me the fearful and overblown coverage will do more damage in the long run as businesses and NGOs pull out, or deals in the future never get done. I’d venture a guess that shaving a percentage point off GDP for the next few years will lead to more preventable deaths than the disease will in the end. This is disastrous for the country and it doesn’t help when organizations like MSF say it is “spiraling out of control“.
- Unless it is actually spiraling out of control. Donne and Helleringer tell us what happens if they project a linear trend. But diseases also expand exponentially. I think the discussion ought focus mainly on the realistic potential for a linear or exponential increase for a long period of time. If agencies can get this under control in a month, the hype will have done much more lasting damage than the actual disease.
- I’m conflicted as to whether history suggests the disease has this potential. If this blog is correct, there have been less than about 3000 confirmed deaths from Ebola in human history in every country in the world. This does not sound like a world ravaging killer.
- At the same time, the 2014 outbreak probably accounts for about half those deaths. So maybe this time is different. 99% of the time “this time is different” is wrong. Until it is not.
- I am waiting for science to weigh in on the trend. Anyone?
68 Responses
RT @engagedscholar: Does Chicken Little have Ebola? http://t.co/BOGAIozAsL
Does Chicken Little have Ebola? http://t.co/BOGAIozAsL
“Does Chicken Little have Ebola?” http://t.co/TBMjj2xETP
RT @cblatts: The $1m question on Ebola is not the level but trend. And I haven’t seen a scientific discussion. Anyone? http://t.co/vKhnZQjs…
Interesting debate about how big a crisis Ebola is: http://t.co/kxtVaM7Kfn
@maiamajumder @bsmorse @cblatts I’d be really delighted to see smart people like you guys play with this approach, debug it, improve it.
@maiamajumder @bsmorse @cblatts I just aggregate up by “generations”. But we’re getting the same results regardless.
@bsmorse @cblatts @DavidFisman (I wanted to use a daily timescale for the data to provide flexibility when picking the serial interval.)
@bsmorse @cblatts In using @DavidFisman’s method, I personally tried to get around lags in reporting by optimizing N for “missing” dates.
@bsmorse @cblatts @DavidFisman R0 & d can be used to determine magnitude & longevity by the following method: http://t.co/LI8rgVQq7M
@maiamajumder @bsmorse @cblatts easy epidemic to model, unfortunately. Especially Liberia. Just straightforward exponential growth.
@maiamajumder @bsmorse @cblatts if anything suspect case reporting will fall off as things fall apart.
@maiamajumder @bsmorse @cblatts posit exponential increase in case reporting over time.
@maiamajumder @bsmorse @cblatts hard to see how this is derived from time-varying reporting effects. You’d have to
@bsmorse @cblatts @DavidFisman Here’s a preview of the model to give an idea of how tightly it fits the data so far. http://t.co/wqeai6K7ul
@maiamajumder @bsmorse @cblatts for sure, happy to connect via email.
@bsmorse @cblatts @DavidFisman While the modeling method has many caveats, it can project final magnitude & longevity of the outbreak.
@bsmorse @cblatts I’d be happy to address your more specific questions over email if you’d like, and I’m sure @DavidFisman would, too!
@bsmorse @cblatts In short, there’s a fabulous modeling method I recently learned from @DavidFisman that gives us a quantitative projection.
@cblatts Our analysis of the reproduction number of #Ebola gives some hints and we are also working on projections. http://t.co/1KHM26NVEE
@bsmorse @cblatts From a brief review of the attached blog post, re: projections, I think it would be best to talk offline.
@bsmorse @cblatts …But I can assure you there’s been plenty of scientific discussion on & offline about most components of the outbreak.
@bsmorse @cblatts I may be able to help with this if you could specify what you mean by “trend” vs. “level”.
Read MT “@cblatts: The $1m question on Ebola isn’t the level but trend. And I haven’t seen a scientific discussion http://t.co/UBCqJ3ZpGL”
I’m with Ed. To take another example, it is very rare for kids to be abducted and raped by strangers, and many Americans fear it more than they should, but it is also disrespectful to the small number of victims to trivialize it.
Here’s an alternative that expresses your point: “Ebola is the al Qaeda of diseases. Terrifying but overrated.”
@cblatts @dadakim re “chicken little”: IMO calculations using empirical data with clearly stated assumptions ≠ fear mongering. Refute away!
@cblatts @dadakim several detailed projections (including hypotheses about control) under way, e.g., work by @DavidFisman
@cblatts @dadakim Thanks! quick re exp growth: best fit for Liberia, but requires lot more detailed explanations than a blog post can hold
@cblatts for outbreak, progression looks like this. Exp growth in Liberia. All details on @MackayIM’s essential page http://t.co/S5FxGbPgQ5
RT @dadakim: in which Kim Yi Donne and Stephane Helleringer say the sky is falling. http://t.co/qeZ4VGMe2K #Ebola
@docregi @soniashah @cblatts @WHO 1 more! Also, Zaire ebolavirus virus outbreaks have varied in proportion of deaths. Expect variability!
@MackayIM @soniashah @cblatts @WHO Understand. Thanks so much for engaging. I thoroughly enjoyed the discussion.
@docregi @soniashah @cblatts @WHO I does but that might be more that some species (not strains in your context) have been found only rarely
@MackayIM @soniashah @cblatts @WHO Was asking regarding truth. I thought initially it was 90% mortality, but now know differently. Thanks
@MackayIM @soniashah @cblatts @WHO As Asking b/c I read the mortality rate differs in certain strains. Was wondering about more mutations.
@cblatts The Embassy in Liberia is floating this around, which is dangerous: http://t.co/yaCQBy495a
@docregi @soniashah @cblatts @WHO Simply hasn’t been info on whether those subtle changes have an impact on virus or disease
@docregi @soniashah @cblatts @WHO Are few subtle genetic diffs, nothing major-all fall within being ‘variant” of a virus of Zaire ebolavirus
@MackayIM @soniashah @cblatts @WHO Have there been any recorded mutations recently?
@MackayIM @soniashah @cblatts @WHO That’s true.
@MackayIM @soniashah @cblatts @WHO Possibly. 1 country outlawed families burying the dead and not reporting last wk. We just hv estimates
@docregi @soniashah @cblatts @WHO I just try and plot the data that we have to hand. Prob a good guide to the *trends*
@docregi @soniashah @cblatts @WHO Agree! But all infection data suffer from same “tip of the iceberg” issues http://t.co/jyv0Y7R94y
Does Chicken Little have Ebola? http://t.co/B6BKRhtIS9 via @feedly
@soniashah @cblatts @MackayIM @WHO Likewise. I don’t trust the old numbers b/c they don’t reflect the number of unreported cases.
@cblatts U could try http://t.co/KNVfutFi9i via @maiamajumder, but tricky w/ time-varying reporting biases of unknown (surely lrg) magnitude
@cblatts Have you seen this http://t.co/E5DiAqakRI from @c_althaus ? Cheers @NyawiraNjoroge
@cblatts public health conundrum: over reaction is often a correct way to help stem an epidemic.
@cblatts have you seen this? RT @MackayIM The rise and rise of #Ebola. Awaiting new @WHO numbers.
http://t.co/rl9Ox5TiQq …
@cblatts That calc completely misses the point. Outbreak has literally shut down care in Liberia. Reported deaths from EBV are 1/x of total
@cblatts you should check out @MackayIM’s TL
The $1m question on Ebola is not the level but trend. And I haven’t seen a scientific discussion. Anyone? http://t.co/vKhnZQjsdh
Does Chicken Little have Ebola? http://t.co/QiwDEfHheC via @cblatts
*Dionne. #typos
@cblatts @helleringer143 that’s not me. My last name is French-Canadian.
RT @cblatts: Some thoughts on @dadakim and @helleringer143 on Ebola in Liberia http://t.co/l1oExLAyh4
Some thoughts on @dadakim and @helleringer143 on Ebola in Liberia http://t.co/l1oExLAyh4
RT @jay_ulfelder: “99% of the time ‘this time is different’ is wrong. Until it is not.” @cblatts on Ebola, but quote is evergreen. http://t…
“99% of the time ‘this time is different’ is wrong. Until it is not.” @cblatts on Ebola, but quote is evergreen. http://t.co/EN8HaYiT8y
Does Chicken Little have Ebola? http://t.co/eGBgsqjKf4
RT @dadakim: in which Kim Yi Donne and Stephane Helleringer say the sky is falling. http://t.co/qeZ4VGMe2K #Ebola
Does Chicken Little have Ebola? http://t.co/vtcGsGTgiP
in which Kim Yi Donne and Stephane Helleringer say the sky is falling. http://t.co/qeZ4VGMe2K #Ebola
Does Chicken Little have Ebola?: This is my second post about Ebola, in which I continue to pontificate about … http://t.co/P1bvwAJyLi
Chris, I’m with you on the need to push back on glamorized, hyped up threats and focus on ones that are less sensational but more serious. However, this is not just about quantifying the impact of Ebola in comparison with other diseases. For the purpose of allocating resources towards public health problems, some perspective is of course in order. But using a phrase like the “the Kardashian of diseases” is pretty insulting to the thousands of people who hemmorage to death with this disease. Once again, I understand that HIV, malaria, TB etc are more pressing matters. But to use a catchy and flippant phrase like that defeats the purpose of trying to shed light on the sensationalization of the Ebola epidemic. Hope I made this point clearly. Big fan of the blog, btw, just thought I’d share my concern on this.
Answer: yes. @cblatts
Is Chicken Little crying Ebola? http://t.co/Ec15fPmg4A …
Ebola is only the Kardashian of diseases if you think Africa is a country http://t.co/yoIbY8ka8B + answer by @cblatts http://t.co/fTJnAKuZFY
Is Chicken Little crying Ebola? http://t.co/niYGcmOTQZ