home
RSS
May 18th, 2010
05:47 PM ET

How accurate is Google Flu Trends?

By Elizabeth Landau
CNN.com Health Writer/Producer

An Internet tool called Google Flu Trends launched in November 2008 with a lot of enthusiasm (although it was not called "Flugle" as I'd hoped). It promised to predict flu outbreaks based on the abundance of people searching for flu-related items on Google search engine.

But a new study questions its accuracy. Researchers at the University of Washington put Google Flu Trends to the ultimate test: comparing its estimates against the Centers for Disease Control and Prevention's national surveillance programs.

Google Flu Trends results have been shown to be mostly accurate in estimating influenza-like illness, but it had not been evaluated against laboratory tests for confirmed influenza virus, Dr. Justin Ortiz of the University of Washington, who led the study, said in a statement. He presented the findings at the American Thoracic Society meeting in New Orleans, Louisiana.

The Web tool is 25 percent less accurate than the CDC at estimating rates of influenza virus infection confirmed with laboratory testing, the research said. For flu-like illness it is robust, however; previous research showed a 92 percent correlation between Google and CDC for the 2008-2009 season, said Jamie Yood, spokesman for Google Flu Trends.

For Google, these findings are not surprising, Yood said. The system was modeled after flu-like illness data, not laboratory-confirmed cases.

That makes sense because flu-like illness isn't always caused by influenza - in fact, only 20 to 70 percent of flu-like illness cases during the flu season are actually influenza, Ortiz said. But the average Google users aren't likely to have a lab test before punching words like "aches" and "fever" into Google, perhaps to see what the diagnosis would be.

"For them, in a way, it doesn’t matter if it actually technically is influenza or not, it’s more or less the same ailment," Yood said.

Media attention to the flu may skew the results for Google Flu Trends, as evidenced by the deviation in Google vs. CDC data in the 2003-4 season, Ortiz said. That influenza season had early and intense flu activity, and substantial media coverage.

But Google Flu Trends does deliver information about flu activity in a fast and cheap way, and provides a good public health service, Ortiz noted. For individuals, knowing about flu activity may help remind people to get flu shots and take other simple precautions, Yood said.

Google Flu Trends is always trying to improve the model, but not because of this study, Yood said. The search engine gurus are working on expanding geographic areas for flu predictions and on improving granularity.

Editor's Note: Medical news is a popular but sensitive subject rooted in science. We receive many comments on this blog each day; not all are posted. Our hope is that much will be learned from the sharing of useful information and personal experiences based on the medical and health topics of the blog. We encourage you to focus your comments on those medical and health topics and we appreciate your input. Thank you for your participation.


May 18th, 2010
01:51 PM ET

BPA present in most canned food, groups allege

By Caleb Hellerman
Senior Producer, CNN Medical News

The chemical bisphenol-A, more commonly known as BPA, is present in virtually all canned foods, according to a report released today by environmental groups who say the omnipresent chemical poses a health risk, especially to infants and pregnant women. BPA is present in the thin plastic lining that protects the surface of metal food containers. A coalition calling itself “The National Workgroup for Safe Markets” conducted laboratory tests on 50 samples of canned food, purchased in stores or donated from home pantries in 19 U.S. states and Canada. Of the 50, 46 contained at least some BPA. The median level was 35 parts per billion, but some food had much more, as high as 1,140 parts per billion in a can of Del Monte green beans.

Bisphenol-A is what is known as an endocrine disruptor, meaning it has the potential to affect the hormones – chemical signals – that direct a range of processes in the body. In animal studies, researchers have linked BPA to various developmental problems, from behavior issues to an increased risk of cardiovascular disease.

Whether BPA is harmful to humans is unclear. Back in 1963, the FDA declared it safe, but more recently, there’s been a scientific reappraisal. The National Toxicology Program now says there is "some concern" for BPA's effects on the brain, behavior and prostate gland, in developing fetuses, infants and children. In January, the FDA posted guidelines urging parents to minimize infants’ exposure through bottles and feeding cups, but it stopped short of saying there is a definite risk of harm.

Pete Myers, a biologist who has studied the effects of BPA, says the level of in baby bottles that triggered alarm, was less than 30 parts per billion, lower than the numbers reported Tuesday about canned food. Myers is chief scientist at the privately-funded group Environmental Health Sciences, based in Charlottesville, Virginia, and has written several articles critical of BPA, including an editorial in the Journal of the American Medical Association.

The Environmental Protection Agency has also taken steps toward further study and possible restrictions of BPA, although the process is still under review. The National Institute of Environmental Health Sciences is funding $30 million worth of research on BPA, with the first round of findings expected as early as this summer, Myers said.

Myers says that research on animals, and other research involving human cell tissue, show that BPA may suppress the production of a hormone – adiponectin – that protects against heart disease. His biggest worry involves a pregnant woman who ingests BPA and passes it on to her developing fetus. “There are some indications it may concentrate in the fetus. It’s definitely not something the fetus is protected from,” says Myers. “There are several [health concerns about BPA], but for me the most worrisome relate to diabetes and heart disease, triggered in infancy or in the womb.”

The FDA declined to comment on Tuesday’s report.

The Grocery Manufacturer’s Association and several makers of canned goods said Tuesday that BPA is not a safety issue. However, both Del Monte and Conagra, whose Healthy Choice soups were tested by the coalition and found to contain BPA, said they are exploring alternatives to BPA in can liners. Conagra told CNN that “as part of our ongoing commitment to providing quality products that meet or exceed consumer expectations, we are constantly looking for new and better ways to package our products. That includes finding safe and effective replacement can liners that do not use BPA.”

Editor's Note: Medical news is a popular but sensitive subject rooted in science. We receive many comments on this blog each day; not all are posted. Our hope is that much will be learned from the sharing of useful information and personal experiences based on the medical and health topics of the blog. We encourage you to focus your comments on those medical and health topics and we appreciate your input. Thank you for your participation.


Advertisement
About this blog

Get a behind-the-scenes look at the latest stories from CNN Chief Medical Correspondent, Dr. Sanjay Gupta, Senior Medical Correspondent Elizabeth Cohen and the CNN Medical Unit producers. They'll share news and views on health and medical trends - info that will help you take better care of yourself and the people you love.

Advertisement
Advertisement