AM Radio Waves May Be Harmful?

Klar writes "Wired News reports that: 'Korean scientists have found that regions near AM radio-broadcasting towers had 70 percent more leukemia deaths than those without.' The article continues: 'The study, to be published in an upcoming issue of the International Archives of Occupational and Environmental Health, also found that cancer deaths were 29 percent higher near such transmitters.' While 'their study did not prove a direct link between cancer and the transmitters', the FDA and the World Health Organization are urging more studies, especially of radio waves from cell phones."

Comments

Wi-Fi?

[caluml]

Wonder what this laptop, resting on my lap, cooking my legs with the battery, and my gonads with Wi-Fi is doing to me?

Re:Incomplete testing

[cytoman]

That was funny!! You wrote about DHMO and Magnetism and Gravity, and you got "Insightful"!!!!

ROFL

Re:Incomplete testing

[geomon]

Let's see:

First it was microwave towers, then power lines, then cell phones.

And every time, the National Academy of Sciences found NOTHING to warrant the claim of a causal link between elecromagnetics OF ANY FORM and cancer.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Wi-fi, Bluetooth and cancer

[otisg]

I wonder what Wi-fi will do to us, since all of us are going to be surrounded by it more and more. Here is what Google thinks about +wi-fi +cancer. And then there is Bluetooth...

Re:There's at least one Nobel Prize...

[zCyl]

... in medicine, and one in physics, and probably one in chemistry, waiting for anyone who can demonstrate a possible mechanism of action for health effects of non-ionizing radiation at athermal levels.

I used to agree with you, but a number of studies recently have shown that under these radiation wavelengths, some membranes in the body pass some molecules when they would otherwise block them.

Example here.

It turns out it's insufficient to just consider heating effects and ionization effects, since lipid membranes are composed of dipolar molecules which can be subject to other electromagnetic effects.

Epidemiologist's rule of thumb

[Beryllium Sphere(tm)]

When you have statistics as your only data and no matched control group, most of the correlations you can find will be coincidence or garbage.

Epidemiologists use the heuristic that they start paying attention when one group has three or more times the risk of another group.
>maybe we should be buying stock in Reynolds
Smoking is a good example: the risk of lung cancer among smokers is about thirty times higher than among nonsmokers.

>Find me a control group. You can't, not on this planet.
That's what lab studies are for. You can shield one group of rats from RF and microwave a genetically identical group. You can start from conception and have useful results in a year.

>Why are you all so reluctant to even entertain the notion that non-ionizing radiation might create a health risk?
After a hundred years of experience and a zillion negative lab studies skepticism is indicated. I'm willing to be surprised but I don't expect to be.

Get it over with

[Deathlizard]

Honestly, I'm just waiting for this statement to come out of a Scientist. It would get it over with and wouldn't spend millions of Dollars.

"If it is or uses either Electricy or a Chemical, and/or its not found in nature in any way, it will kill you slowly"

Re:what part of "needs further study" dont' you ge

[Idarubicin]

But don't worry, even if a study or three come out demonstrating a link between non-ionizing radiation and cancer risk, the EPA will sweep it under the rug when Infinity Broadcasting supresses the evidence under the Bush Administration's Data Quality Act.

Actually, a study or three demonstrating a statistically significant link between nonionizing radiation and cancer is exactly what I would expect, even in the absence of real harmful effects.

This is epidemiology--hardcore statistics. When determining the risk associated with some factor, you can never be entirely certain that the effects you see are 'real', and not just due to random clustering. Toss a coin ten times--you'd expect to get heads five or so times, but occasionally (1 time in about a thousand) you'll see ten heads in a row.

By making (generally reasonable) assumptions about the nature of the randomness in the data, scientists and epidemiologists tend to come up with one or more measures of how likely an apparent result is to be genuinely significant. Generally, a result is taken to be 'real' if there is less than a 5% chance that the result is the result of noise (a P value of less than 0.05). Alternately, a study may state an odds ratio and 95% confidence interval ("If you take drug foostatin you are 1.7 times more likely to have symptom bar (95% CI 1.4 to 1.95)") denoting that the relative risk is 95% likely to fall in the stated interval.

Under those circumstances, if the scientists do everything correctly, and account for every possible confounding factor, and do all their math correctly...that still leaves as many as one study in every twenty potentially reaching the incorrect conclusion.

The journal in question here--The International Archives of Occupational and Environmental Health--isn't exactly a top-flight journal, either. I'm not at work at the moment so I can't check their archives, but their impact factor is fairly low. (Down to 0.924 in 2002, declining steadily since 1997.) Yes, impact factor is by no means the only criterion by which a journal should be judged--but Nature they are not. Unfortunately, the Wired article refers to an 'upcoming' paper, so I can't get at the publication cited.

Looking at the other paper mentioned in the Wired article demonstrates that Wired can't be trusted to accurately report the findings of scientific papers, either. Wired says:

Two years ago an Italian study found death rates from leukemia increased dramatically for residents living within two miles of Vatican Radio's powerful array of transmitters in Rome.

The abstract of the original paper in the American Journal of Epidemiology says: (in part, emphasis added)

...In the 10-km area around the station, with 49,656 residents (in 1991), leukemia mortality among adults (aged >14 years; 40 cases) in 1987-1998 and childhood leukemia incidence (

eight cases) in 1987-1999 were evaluated. The risk of childhood leukemia was higher than expected for the distance up to 6 km from the radio station (standardized incidence rate = 2.2, 95% confidence interval: 1.0, 4.1), and there was a significant decline in risk with increasing distance both for male mortality (p = 0.03) and for childhood leukemia (p = 0.036). The study has limitations because of the small number of cases and the lack of exposure data. Although the study adds evidence of an excess of leukemia in a population living near high-power radio transmitters, no causal implication can be drawn. There is still insufficient scientific knowledge, and new epidemiologic studies are needed to c

Re:what part of "needs further study" dont' you ge

[Idarubicin]

But is statistics the only way? Can every ill health effect be demonstrated via the appropriate confidence interval and a large enough sample size? (Godel's Incompleteness Theorem?)

Well, we're talking biochemistry here, so there's really no cause or need to invoke the Incompleteness Theorem.

Further, no--it's not possible to demonstrate every ill health effect. A thought experiment, if you will...

Consider the very rare but highly feared disease X, which affects one of every million people. Consider also potentially toxic compound Y. It is present in the drinking water of every person in Los Angeles (population approximately ten million), and nowhere else in the state.

One would expect approximately ten cases of disease X in the city population, but there will be some deviation due to random clustering. One expects the number of cases to follow a Poisson distribution, giving a standard deviation of about three cases.

Under those circumstances, there's a 95% chance that the number of cases observed in the city will fall between 5 and 15. To have any hope of discerning a risk associated with compound Y, you need to see more than fifteen cases. Realistically, you probably need to get out to about twenty cases observed before you can start saying anything about the 'dangers' of Y. In other words, for this compound and this population, if chemical Y increases your risk of disease X by less than about a factor of two, you're not going to be able to clearly see it.

If Wired saw thirteen cases in LA, they'd say that compound Y causes a dramatic (thirty percent!) increase in disease X. If a scientist saw thirteen cases in LA, they'd say that's interesting, but easily attributable to noise.

Clearly a jump from 4 to 8 leukemia cases means practically nothing -- statistically. But I don't think it's always good science, esp. when dealing in real-world non-controlled systems with intangible variables, to rely on statistical analysis as the impetus for public policy decisions.

If there is sound evidence (good animal or at least biochemical models) that particular conditions are harmful, then by all means such evidence should be considered. Controlled trials in the laboratory are very useful for sorting out cause and effect. In the absence of demonstrated mechanisms for harm in the lab, epidemiological data are all that we have. If sound statistical analysis reveals a significant correlation--that cannot be reasonably explained by other means or attributed to confounding factors--then it may be a fair basis for policy decisions.

I suppose the problem arises when one asks what constitutes a 'sound' analysis...and in some cases that's a difficult question.