Flu outbreaks predicted with weather forecast techniques
BOULDER—Scientists at Columbia University and the National Center for Atmospheric Research have adapted techniques used in modern weather prediction to generate local forecasts of seasonal influenza outbreaks.
By predicting the timing and severity of the outbreaks, this pilot system can eventually help health officials and the general public better prepare for them.
The study, published this week in the Proceedings of the National Academy of Sciences, was funded by the National Institutes of Health and the Department of Homeland Security. NCAR’s sponsor is the National Science Foundation.
From year to year, and region to region, there is huge variability in the peak of flu season, which can arrive in temperate areas of the Northern Hemisphere as early as October or as late as April. The new forecast system can provide “a window into what can happen week to week as flu prevalence rises and falls,” says lead author Jeffrey Shaman, an assistant professor of Environmental Health Sciences at Columbia’s Mailman School of Public Health.
In previous work, Shaman and colleagues had found that wintertime U.S. flu epidemics tended to occur following very dry weather. Using a prediction model that incorporates this finding, Shaman and co-author Alicia Karspeck, an NCAR scientist, used Web-based estimates of flu-related sickness from the winters of 2003–04 to 2008–09 in New York City to retrospectively generate weekly flu forecasts. They found that the technique could predict the peak timing of the outbreak more than seven weeks in advance of the actual peak.
“Analogous to weather prediction, this system can potentially be used to estimate the probability of regional outbreaks of the flu several weeks in advance,” Karspeck says. “One exciting element of this work is that we’ve applied quantitative forecasting techniques developed within the geosciences community to the challenge of real-time infectious disease prediction. This has been a tremendously fruitful cross-disciplinary collaboration.”
Up next: your local flu forecast
In the future, such flu forecasts might conceivably be disseminated on the local television news along with the weather report, says Shaman. Like the weather, flu conditions vary from region to region; Atlanta might see its peak weeks ahead of Anchorage.
“Because we are all familiar with weather broadcasts, when we hear that there is an 80 percent chance of rain, we all have an intuitive sense of whether or not we should carry an umbrella,” Shaman says. “I expect we will develop a similar comfort level and confidence in flu forecasts and develop an intuition of what we should do to protect ourselves in response to different forecast outcomes.”
A flu forecast could prompt individuals to get a vaccine, exercise care around people sneezing and coughing, and better monitor how they feel. For health officials, it could inform decisions on how many vaccines and antiviral drugs to stockpile, and in the case of a virulent outbreak, whether other measures, like closing schools, is necessary.
“Flu forecasting has the potential to significantly improve our ability to prepare for and manage the seasonal flu outbreaks that strike each year,” says Irene Eckstrand of the National Institutes of Health’s National Institute of General Medical Sciences.
Worldwide, influenza kills an estimated 250,000 to 500,000 people each year. The U.S. annual death toll is about 35,000.
The seed of the new study was planted four years ago in a conversation between the two researchers, in which Shaman expressed an interest in using models to forecast influenza. Karspeck “recommended incorporating some of the data assimilation techniques used in weather forecasting to build a skillful prediction system,” remembers Shaman.
In weather forecasting, real-time observational data are used to nudge a numerical model to conform with reality, thus reducing error. Applying this method to flu forecasting, the researchers used near-real-time data from Google Flu Trends, which estimates outbreaks based on the number of flu-related search queries in a given region.
Going forward, Shaman will test the model in other localities across the country using up-to-date data.
“There is no guarantee that just because the method works in New York, it will work in Miami,” Shaman says.
About the article
Title: Forecasting seasonal outbreaks of influenza
Authors: Jeffrey Shaman and Alicia Karspeck
Journal: Proceedings of the National Academy of Sciences (Abstract)
Someone did a study in Baltimore, Maryland on how well Google Flu Trends correlates with actual flu cases. Their results:
However, another study concluded that “while Google Flu Trends is highly correlated with rates of ILI, it has a lower correlation with surveillance for laboratory-confirmed influenza.”
I found these with a search in Bing using “google flu trends” correlation as my search terms; there looked to be several other studies in the search results, but I didn’t have time to look at more than these two. Acronyms: GFT = Google Flu Trends; ILI = influenza-like illness (“ILI is defined as a fever ≥37.8°C and a cough and/or a sore throat without known etiology”); ED = Emergency Department.
I read the article on the upcoming cold snap in the UK; temps in December may get to -20 C (4 degrees F), not 20 C (68 degrees F); it was misquoted early in the article and corrected later on. As for the humidity-flu link: How long does the dry spell need to last in order to trigger a flu outbreak? I think the incidence of flu still has a lot to do with contagion, crowds, people going to work sick or sending children to school sick, and the like; most cannot avoid going to work or school simply because others are likely to be there who have gone there not feeling well. I can see it now: Worker calls boss and says, “The flu forecast predicts a high contagion level today, so I’m staying home.” Boss: “Your last paycheck will be mailed to you; with a high unemployment rate, I’m sure I can replace you right away. Thanks for the warning!”
Flu and infectious diseases are merely Mother Nature’s right hand man at work. We need harsh weather to cull out the schtumpig. Better, live on a cold, poor island of old fisherfolk, they know how to live and when to die.
“Might” “eventually” “could” “potential” blah blah blah
“tended to occur following very dry weather” Dew point plunges here mid to late November. Without moisture in the air to retain heat, temperature dives when the sun goes down. This extremely dry situation lasts til about March. Meteorology, it’s called “winter.”
I find it hard to believe that any one in the country would do anything differently if they were told that the seasonal flu outbreak was going to occur in 7 weeks. Even if it were believed to be 100% accurate, it wouldn’t change anything.
Lastly, hygiene is important, regardless of time of year or specific threats. Though it would be just like Stephanie Abrams to tell us to start washing our hands because a flu outbreak is pending.
How about working on the cause of the flu?
Logan in AZ says:November 27, 2012 at 7:58 pm
Perhaps you can point me to an information source. I am interested in the affects of D3 on sarcoidosis granulomous. I have scavenged PubMed, learning that in the past decade or so, Williams et al 2001, seemed to first mention that D3 exacerbates sarcoisosis. It seems the D3 is used by the cells that differentiate between self and not self, I’ve not found explicitly how. Later abstracts appear to confirm the D3 connection. Avoiding D3 seems to alleviate or minimize the sarcoidosis. Are there any further information sources?
This was my reaction. Data mining produces all types of low p-value results. Remember there’s a 5% chance that any correlation with a p-value of less than 0.05 will be a coincidence. If 20 correlations are explored, then it should be expected that one will be a coincidence.
It’s impossible to back test something when all of the back data has been used to find a high correlation. I, too, am curious about the predictive ability of this model.
Fixed
It’s a pity the paper was submitted in May 2012 and approved in October, otherwise we could have all had the benefit of their actual predictions for peak flu season this winter.
Still, I sure we can expect to see these predictions, and monitoring of their accuracy and usefulness, in the mainstream media. Probably about the same time as Foster and Rahmstorf publish some notable, verifiable temperature predictions with all appropriate definitions and dates.
The question is, given the same environmental circumstances, why do some people get the flu and others not? A robust immune system is key. For that, people have to ask, what is good health? I for one am very skeptical about the claims of the efficacy of high-doseage D supplements (i.e. 5,000 IU’s). I do believe good dietary and exercise (particularly outdoors) practices to be keys. Some vitamin supplementation may be beneficial as well. I do take a Vitamin A (10,000 IU) and D (400 IU) capsule, but not every day, perhaps three times a week. I haven’t had the flu in perhaps 40 years or more, nor have I ever had a flu shot.
@nerd – right on!!
@alexwade – IMnsHO, you’re playing the fool.
@Sam Grove – It’s UVb, not UVa
@Logan in AZ and @kimyo – THANKS!
D3 (25-hydroxy-D3 at >50 ng/ml) protects against B strains of flu but apparently not the A strains.
BAC
– Dr Lisa Jackson’s out of season influenza vaccine research
http://ije.oxfordjournals.org/content/35/2/337.short
Great bit of research from Dr Jackson on the “healthy user” problem and seasonal influenza. Turns out that the vast majority of the benefits claimed for the seasonal influenza vaccine are due to the healthy user effect. So you are way better off looking at other ways like hand washing and vitamin D3 supplements.
As to vitamin D (really a hormone, not a vitamin) it is vital to get enough as some very good research is showing. The sun needs to be above 50 degrees for the UVB to get through to you at ground level. That is both seasonally and daily. So if you live above the 35th parallel north you are looking at May to September and 11 AM to 2 PM. If you are going to supplement then make sure you are getting the D3 form not the D2. The D2 form is more stable so it is prescribed and in milk etc but the D3 is better utilized by your body.
I like google flu trends and think if they could tie in a few more feedback loops it would be great. Problem is the old correlation vs causation one. People scouring the web doesn’t mean they have it. Same with the weather but interesting approaches to a problem.
It is not a medicine blog here. But anyway: Vitamin D plays a decisive roll. For those who want some earnest medicine-stuff try this article (it is real science):http://www.virologyj.com/content/5/1/29
Believe it (or better not): I just attended an examination in a german Backward-Cimate-School near Berlin with the Professors Sourcreamtorf and Bellhuber. The pupil Felix K. Calculus was asked to tell the temperature if it was 27° C and increased 11° C. His result: 49° C. Prof. Sourcreamtorf remarked: That’ s completely wrong, my dear, it’s much too high. But Felix insisted. He declared, that his result was higher than expected and thus it was much better. Prof. Belllhuber agreed immediately. Yes! he exclaimed, it is indeed much higher than predicted in the lessons. Therefore it’s much better too! It is that way that climate-mathematics work. Prof. Sourcreamtorf acknowledged his mistake. So Felix became best of the class.
W. Sander says:
November 28, 2012 at 12:04 pm
It is not a medicine blog here….
________________________________
No it is not but it is amazing the expertise, such as yours, that comes crawling out of the woodwork no matter what the subject.
Thanks for the pointer to the paper.
“…(9) Why has influenza mortality of the aged not declined as their vaccination rates increased? Is enough to make anyone over the age of 45 keep reading.
b a cullen, you’re correct, UVB.
This presentation by Dr. Holick is quite entertaining. It’s the keynote address he gave at the 34th European Symposium on Calcified Tissues in Copenhagen.
http://www.uvadvantage.org/portals/0/pres/
Interesting that the time when people can benefit most from the sun to produce Vitamin D is the time the skin cancer scare lobby have virtually banned schoolchildren in Australia from going outside without full sunscreen or clothing protection. And, adults that work outdoors are increasingly being made to cover up because their employers are told they could be sued if an employee gets skin cancer.
Gail wrote, “it is amazing the expertise … that comes crawling out of the woodwork no matter what the subject.” True, but we should remember that Anthony’s asked us to generally remain on climate topics. Hmmm… although I guess we could argue that all those fluey people running high temperatures might contribute to global warming…
;>
MJM
@ur momisugly Steve Keohane, who asks about D3 and sarcoid.
Yes, there are extensive but controversial sources on this and related topics. The background remark is that chronic infections that our immune response finds difficult to control (TB, leprosy) can produce a granuloma that contains cells (activated macrophages) which convert too much of the calcidiol pool to the active calcitriol, producing a toxic state (hypercalcemia). The macrophages try to ‘wall off’ the bacteria that are hard to kill.
The vitamindwiki page is at —
http://www.vitamindwiki.com/tiki-index.php?page=Vitamin+D+causes+sarcoidosis+in+1+in+5000+people
The critique by Dr. Cannell is at —
http://www.vitamindcouncil.org/news-archive/2008/professor-marshalls-recent-discovery/
The Cannell critique is not the last word, and the Marshall Protocol is more extensive than a concern for calcidiol conversion to calcitriol. Trevor Marshall has an elaborate theory about the role of cell wall deficient (CWD) bacteria and autoimmune conditions. CWD organisms are hard to detect and often difficult to culture. The big names on CWD or ‘L-form’ bacteria are the late L. H. Mattman and the retired Gerald J. Domingue. There are short wikipedia articles on Mattman and Domingue. Medline papers by a certain author can be found by entering, say,
domingue gj [au]
There is a book by Lida Mattman — Cell Wall Deficient Forms, Stealth Pathogens. You might find it in a large university library — the third edition was the last. It can also be purchased from Amazon for a high price.
Marshall founded an organization to advance his views and treatments.
http://autoimmunityresearch.org/
And, there is such a thing as a ‘Marshall Protocol DataBase’
http://mpkb.org/
I don’t know if there is an impartial evaluation of Marshall’s therapy. I only see two articles by Marshall on Medline, but he has written much more that has not been indexed. You will have to do some homework to judge such a controversial issue.
Best Regards,
Logan
[snip . . please just post links to articles rather than clip and paste the entire article as there can be copyright issues. Thank you . . . mod]
kwik says:
November 28, 2012 at 2:15 am
Marc says:
November 27, 2012 at 8:24 pm
“Flu is more common in the winter because the gel-like coating that protects the inner virus from dying on a surface stays intact at lower temperatures and breaks down at warmer temperatures allowing the live virus to die from exposure, put in overly simple terms.”
Interesting ! Do you have a link? It would explain the funny situation that we all know the flu is a virus infection, and yet you tend to get it when it’s cold. When you mention the virus, there are all kinds of ecplanations, like, people tend to cluster “inside” when it is cold, and so on.
But here we have a scientific explanation/theory that sounds reasonable. Again; Link?
http://www.reuters.com/article/2008/03/02/us-flu-winter-idUSN0228175320080302
Here is another link with some further elaboration on the virus temperature relationship that also addresses the humidity/dryness issue.
It is all about the gel-like viral envelope and its temperature and humidity sensitivity.
http://www.nytimes.com/2007/12/05/health/research/05flu.html?_r=0
@Nerd:
Spouse has S.A.D. and oral Vit-D doesn’t fix it. We got a “Lizard Lamp” and during the winter she spends about 10 minutes a day under it. All better….
I was feeling “unenthusiastic” this last week (as it’s been all-indoor all-the-time due to rain and schedules). Spend 20 minutes under the lamp. Loads of energy and ‘feel good’…
@Sam Grove:
I’m pretty sure it’s UVB that makes Vit-D. Production goes up as wavelength goes down…
http://www.uvguide.co.uk/phototherapyphosphor-tests.htm#photoactivity
FWIW, the Zilla 25 (a 13 Watt CFL) runs about $22 on Amazon. Yes, I know, it’s placarded as ‘only for lizards and reptiles and don’t even think about it if you are a human’. So call me a lizard. It works for me… then again, it’s not used for 12 hours a day as on a lizard and we use eye protection – even though likely not needed. Plain old glasses absorb UV, or just sit with your back to it. You can get a sunburn in about 40 minutes to an hour with your back less than 2 feet from the bulb; provided you have pasty white transparent skin like me ;-). I deliberately did that test, in increments, to establish guidelines for the spouse… 4 feet and less than 20 minutes…
Maybe we’ll go blind or die of skin cancer or ‘whatever’, but at least we’ll be happy 😉
(The spouse has the doctor do Vit-D assay every so often to monitor how she’s doing, so it’s not quite like we’re flying by the seat of our pants here… Taking the oral form didn’t get her assay up enough. Yes, there are commercial S.A.D. lamps. They are not significantly different, near as I can tell. Some more visible / blue and often less effectively controlled for enough UV or sometimes too much UV. It’s still an ‘evolving science’. Personally, I think the lizard folks have done more, in more detail, longer duration… The Zilla lamps now have a detailed spectrum chart on the package with power ratings by spectrum. Nice.)
Oh, and your LED lamps often put out too much blue in just the spectrum that resets your biological clock. So if you put in LED lamps and now find you stay awake late at night…
http://chiefio.wordpress.com/2012/07/22/superchiasmatic-led-light-insomnia/
We now use the LED bulbs in the morning and swap to ‘yellower bulbs’ in the evening and all is well… ( I have a stash of incandescent bulbs… now that ‘curly bulbs’ are mandated in California. Expect as LEDs spread and more “daylight type’ CFLs enter use we’ll have a lot more sleepless irritable people. Note that we discovered this AFTER converting to LEDs… so it’s not like I was resisting them…)
Logan in AZ says:November 28, 2012 at 5:25 pm
Thank you for the information and links. It is difficult to find anyone with much, if any, experience with sarcoid.