7

This article claims that a computer algorithm predicted the recent (2014) Ebola outbreak 9 days before the World Health Organization announcement. Is this true?

As you can read in the article the algorithm didn't identify the disease, only collected the data before the official announcement. Did it really have more information before the health organizations?

  • Am I missing the part where they say the algorithm is called Big Data? Looks to me like the correct name is HealthMap – nico Aug 13 '14 at 11:20
  • There is no algorithm called "Big Data." Also the algorithm and website is described in a PLOS Medicine paper linked in the article. – shadowtalker Aug 13 '14 at 11:26
  • "Big Data" is essentially just a buzzword for statistics/data-mining performed on the sort of very large databases that are now collected (e.g. Twitter feeds). en.wikipedia.org/wiki/Big_data – Dikran Marsupial Aug 13 '14 at 11:29
  • 2
    I have edited it to focus on the Ebola outbreak, while trying to retain as much of your original language as possible. I hope it suits you. – Flimzy Aug 13 '14 at 13:42
  • 1
    To the answerers This is a very specific claim: only use references which address the claim. If they made this prediction in any official capacity there has to be a document trail of it. Showing that it's possible, or that predictions are not their core business does not prove anything. – Sklivvz Aug 15 '14 at 12:30
7

Did Healthmap issue an alert about Ebola before the WHO announced an outbreak had been confirmed

(question heavily paraphrased)

Yes (probably).

Their alert was probably issued on the basis that one or more healthcare workers had blogged about their concerns that the illness they were confronting might be Ebola.

It seems the WHO wait for reports and confirmation before issuing announcements of an outbreak.

(reference - see text of article linked in question)

Did a computer algorithm predict the 2014 Ebola outbreak

No.

The linked article is entitled "How A Computer Algorithm Predicted West Africa’s Ebola Outbreak Before It Was Announced" and goes on to discuss Healthmap.

The claim appears to be that Healthmap computer algorithms predicted an Ebola outbreak 9 days before the WHO announced it.

However Healthmap made no prediction saying something like "there will be an Ebola outbreak in West Africa in 9 days time"

Healthmap issues alerts about recent historical events, it does not issue predictions about future events.

So the claim that Healthmap computer algorithms made a prediction is false.

The article in question appears to be an entry on a personal blog written by "a doctoral student working on my PhD in Social and Behavioral Health"

The body of the article doesn't attempt to support its headline. The claim in the headline may not be a serious or notable claim.

The body of the article supports a claim that Healthmap announced possible cases of Ebola in West Africa before the WHO did. The article claims that Healthmap did this because Healthmap picks up on mentions of diseases in the personal blogs of healthworkers rather than only relying on official formal reports via national health agencies.

So how did a computer algorithm pick up on the start of the outbreak before the WHO? As it turns out, some of the first health care workers to see Ebola in Guinea regularly blog about their work. As they began to write about treating patients with Ebola-like symptoms, a few people on social media mentioned the blog posts.

About Healthmap

HealthMap, a team of researchers, epidemiologists and software developers at Boston Children's Hospital founded in 2006, is an established global leader in utilizing online informal sources for disease outbreak monitoring and real-time surveillance of emerging public health threats. The freely available Web site 'healthmap.org' and mobile app 'Outbreaks Near Me' deliver real-time intelligence on a broad range of emerging infectious diseases for a diverse audience including libraries, local health departments, governments, and international travelers.

So its subject is near real-time monitoring of reports of health problems.

You can sign up for email alerts. These are shown graphically on it's front page. A summary on that page counts numbers of alerts under various categories including "Ebola(1543)"

At the Healthmap website I can find no predictions of future events.

  • Can you provide references? The only link you provide doesn't seem to talk about this at all. Also you are not addressing the claim. – Sklivvz Aug 15 '14 at 12:25
  • @Sklivvz: I've added some references, clarified the specifics of the claim my answer addresses and removed extraneous material. – RedGrittyBrick Aug 15 '14 at 13:53
  • 2
    I get the point: if the copywriter of the headline had used "detected" rather than "predicted", it would have matched the journalist's article (and the facts) better, but this result seems a teensy bit pedantic about word-usage. I would have been tempted to edit the question to use "detected" rather than answer it this way. We still don't know if the tool DID detect the outbreak (in a specific manner). – Oddthinking Aug 15 '14 at 15:07
  • 1
    @Oddthinking: Detection seems a much less dramatic claim than prediction. Maybe someone can write a better answer. I don't know if Healthmap detected Ebola, the infection, in a meaningful sense or just detected a healthcare worker blogging about the possibility of Ebola. The latter is still useful, I assume it takes multiple independent reports to trigger an "alert". Maybe someone can dig up info about their false-alert rates. – RedGrittyBrick Aug 15 '14 at 15:45
  • FWIW, this abuse of predict to mean recognize, detect, estimate is commonplace in medicine. The metric "positive predictive value" of a test is a shining example. The only "future" aspect is that the test predicts the conclusion a human would eventually reach, but what the test measures is already present. But it IS the accepted definition of predict in medical contexts, and therefore your dismissal of it as incorrect is itself incorrect. If you use the intended meaning, then yes the paper is correct (and so is your final sentence that no future events are concerned). – Ben Voigt Apr 7 '18 at 23:29
3

This sort of thing is an active research topic (a friend of mine works on it), for example social media can be mined to detect outbreaks of disease before it would normally be reported by health organisations. Here is an example of a workshop held on this sort of thing (although in this case it was motivated via bioterrorism, rather than naturally ocurring events).

Social media and its dual use in biopreparedness: communication and visualization tools in an animal bioterrorism incident.

Sjöberg, Elisabeth; Barker, Gary C; Landgren, Jonas; Griberg, Isaac; Skiby, Jeffrey E; Tubbin, Anna; von Stapelmohr, Anne; Härenstam, Malin; Jansson, Mikael; Knutsson, Rickard Biosecurity and bioterrorism : biodefense strategy, practice, and science. 11 Suppl 1: S264-75; 2013. (English).

This article focuses on social media and interactive challenges for emergency organizations during a bioterrorism or agroterrorism incident, and it outlines the dual-use dilemma of social media. Attackers or terrorists can use social media as their modus operandi, and defenders, including emergency organizations in law enforcement and public and animal health, can use it for peaceful purposes. To get a better understanding of the uses of social media in these situations, a workshop was arranged in Stockholm, Sweden, to raise awareness about social media and animal bioterrorism threats. Fifty-six experts and crisis communicators from international and national organizations participated. As a result of the workshop, it was concluded that emergency organizations can collect valuable information and monitor social media before, during, and after an outbreak. In order to make use of interactive communication to obtain collective intelligence from the public, emergency organizations must adapt to social networking technologies, requiring multidisciplinary knowledge in the fields of information, communication, IT, and biopreparedness. Social network messaging during a disease outbreak can be visualized in stream graphs and networks showing clusters of Twitter and Facebook users. The visualization of social media can be an important preparedness tool in the response to bioterrorism and agroterrorism. (source)

See also this article on Influenza surveillance.

The particular tool involved in this question is HealthMap, and they provide a timeline of the Ebola outbreak here, which allows you to see what it was saying at different points in time. HealthMap doesn't specifically predict outbreaks of disease, but what it does do is to find and collate relevant information from press reports and social media that allow potential outbreaks to be identified earlier than could be reasonable done manually. It would be more accurate to say that HealthMap alerted us to a potential outbreak of some hemorrhagic disease 9 days earlier than the WHO.

  • Can you find anything specific to this claim? As is, this is an extremely partial answer. – Sklivvz Aug 15 '14 at 12:26

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .