Flu Data

A few years back, I bought $300 worth of food and water to store in the basement in case some terrible bird flu hit and we couldn’t leave the house for a week or two. We’ve eaten the good stuff bit by bit and a lot of what is left is now well past its prime.

Even though you don’t hear much about the bird flu anymore, I still believe some sort of infectious pandemic is the single greatest threat to humans, so perhaps it is time to restock my basement.

In the meantime, for the hypochondriac who loves data, the C.D.C. and google.org have competing influenza-surveillance web sites. The C.D.C. has a very detailed weekly flu report, which even describes the antigenic breakdown of the strains of flu that are found. According to the C.D.C., there is essentially no flu activity at the present time in the U.S.

That doesn’t stop people from thinking they’ve got the flu, however, as evidenced by google.org’s Flu Trends page. Google measures flu activity by the number of flu-related searches that take place, and these searches are starting to rise (for more details, see yesterday’s Times article).

Historically, the C.D.C. data and Google data trend quite closely.

There hasn’t been a really bad flu season since 2004. Is that because more people are getting flu shots, or is it just chance?


"This blog is usually much more logical than this passage [that people who think they have the flu search more for flu] would suggest. — Posted by Dave"

Actually, if you read the paper that Google produced for Nature, you'll see that this hypothesis is borne out by the data.


@4: "The main factor of the bird flu is probably the fear of getting it. In 2007 the H5N1 bird flu killed 59 people worldwide, making it a small problem compared to other death causes."

One of the main concerns with H5N1 is whether it will mutate to become communicable from human to human. It has already mutated enough to transfer from birds to humans, so it seems to be very possible that this will happen. The number of deaths have been relatively constrained since people currently infected are not contagious to other humans.

The real question is whether the mortality rate is as bad as it appears in the current data. It is likely the true severity of this strain of influenza has been over exaggerated, since the number of cases that were not terminal are likely under-represented. I would guess that rural Asian chicken farmers are not very likely to go the hospital for mild cases of avian flu, and therefore not recorded.



I have to say, my father has recently got himself in a real panic that we are overdue for an epidemic that will wipe out thousands if not millions of usm, and he has bought tons of food which is stored in various cupboards and rooms throughout the house, so we can quarentine ourselves.

However I don't live at home any more, and I have been told that I won't be welcome in the event of an emergancy in case I "Kill us all"! (Although I maybe allowed to go in the shed in the garden for an incubation area to check I'm safe, depending on the nature of the virus)

Healthy paranioa, or slightly obsessive?


I had the famous flu of 2004 - and that was the year that the vaccine didn't include one particular strain, which left people open to catching it even if they got a flu shot. They updated the vaccine but by then there was a shortage if I remember right... it's all a blur...but not how I recommend spending a vacation.


IMHO, I don't think there is as strong a correlation between influenza epidemics and flu-related google searches as they'd like to think. I believe that there is still rampant confusion between life-threatening influenza and the normal gastroenteritis virus. The former scares me, the latter seems like an annual ritual.


I'm in disagreement with Al Marsh. I think you're pretty wise to have a small stash of food, water, etc. in the basement. We've grown up in a culture where the frequency of really bad things has diminished - but they're still possible (e.g., Katrina). If such an event does occur, you and your family will avoid panic.

Besides, in terms of decision analysis, it seems that the advantage of having such a stash in time of need (even considering the small probability of occurrence) is worth it when compared to the alternative.


The most interesting part of the google flutrends site is that they claim it's actually a better predictor and indicator than the CDC (the article was on wired yesterday). Even if there are a lot of hypochondriacs out there, the proportion of those to have the flu to those who don't is probably relatively constant so the trends are likely pretty accurate. What I'd be more interested in is the relation between decreasing medical coverage and increased usage of the internet for diagnosis/remedies...


I don't think your paranoid. I think it's very wise to have some sort of backup plan. Anybody ever read "The Road"? That's what I'm talking about.


No plug for the Iowa Influenza Market? For shame.


Whether or not the next disaster which (temporarily?) cripples our complex life and luxury support system is a disease, the Mormon custom (do they still observe it?) of maintaining and rotating 6 months' supplies in your household store has great attractions now. If we all decided to do just that, the fiscal stimulus would be more than sufficient to end the recession. And, our balance sheets as individuals (though not our liquidity) would not be harmed.

Nitin Rao

As this news was released, we must have seen a spike in flu searches.


"That doesn’t stop people from thinking they’ve got the flu . . . [as indicated] by the number of flu-related searches that take place"

It's a pretty big leap to suggest that flu-related searches indicate that people believe that they have the flu. Every year at around this time I get multiple emails (government, work, office building etc.) reminding me to get my annual flue vaccine. Perhaps the increase in flu-related searches is due to people (rationally) trying to avoid the flu, not because they think that they have it.

When SARS was in the news I doubt that a statistically significant amount of searches on the topic were from people who actually thought that they had SARS.

This blog is usually much more logical than this passage would suggest.

Sam R.

Won't the publicity surrounding Google's use of flu trends muck up their data (as everyone starts entering "flu" into Google just to see what comes up)?


I agree, the heighten attention given to the bird flu is not to be taken lightly. If not bird flu then some other communicable disease is going to hit us like something we’ve never seen before. Six of the 18 worldwide epidemics listed in Wikipeda are influenza so if you’re going to worry about some disease, that’s a good one to worry about.

People can’t imagine whole families or communities being wiped out. Of all the natural disasters that have befallen us - even before there was an us - a pandemic is probably more likely than the others, such as Yellowstone caldera erupting, a meteorite or comet striking the earth, that major landslide just waiting to happen in the Canary Islands causing a giant tsunami that wipes out all of the East Coast, New Madrid earthquake, or even the inevitable return to the Ice Age when Canada become uninhabitable and New York City get wiped off the face of the earth. Flu is the one to bet on.



Why would I google 'flu symptoms' this year when I obtained all the information last year? Do symptoms change that radically?

It would be interesting to see if numbers of searches over time decreases.


If memory serves, wasn't 2004 the year they didn't produce enough flu vaccine? If so, that would account for that particular spike ... meaning that, had the shortage not occurred, there would have been less flu that year.

(It's also odd that there was a spike in the flu that year, since when the shortage hit, the public was told that the flu vaccine wasn't all that necessary and if you had to forgo the shot, you would still be OK. If there truly was a spike in the flu because of the shortage, I'd say that their assurances turned out to have been quite wrong.)

Al Marsh

I have to say, I think you're pretty paranoid. Sorry.

Paul K

I think flu shots are a factor in stopping the spread. That is, if only a relatively small percentage get the shots, they break the normal vectors (e.g. kid gets flu, gives to parent, who takes to work, etc).

The other big factor is that various Asian countries have taken major steps to prevent outbreak and spread, from quick actions in chicken farms to stopping sick travelers (they monitor arrivals and departures at most of the international airports in China, HK, Japan, Singapore, etc).


I wonder if the reporting of this story will destroy its accuracy - I can't be the only person who's tempted to Google "flu symptoms" just to jack with the numbers.


Hm, I feel I constantly have a flu, however, that has no point.

The main factor of the bird flu is probably the fear of getting it. In 2007 the H5N1 bird flu killed 59 people worldwide, making it a small problem compared to other death causes.

This is the same as with terrorism. Although I do not have the numbers here, if I knew there was going to be a terrorist action in London during the next 24 hrs in the trams, I should probably be more scared to take my bike that day than travel underground.