USA: The Wall Street Journal over de statistiek bij het Interphone onderzoek.

In onderstaand artikel geeft de Wall Street Journal een interessante analyse van de statistische gegevens uit het Interphone onderzoek:

Bron: The Wall Street Journal 29 mei 2010

Static in Cellphone Study

Auteur: Carl Bialik

The final results of a major international study of the potential link between cellphone use and cancer were published last week. The finding: Using a cellphone seems to protect against two types of brain tumors.

Even the researchers didn't quite believe it.

The apparent shield of cellphone radiation, most likely fictitious, illustrates how hard it is to analyze, let alone quantify, the potential for a small elevated risk in a rare disease from a widespread, mundane activity.

''They found that ever having used a cellphone appeared to be protective against brain cancer,'' says David O. Carpenter, director of the University at Albany's Institute for Health and the Environment, in Albany, N.Y. ''And that just simply makes no sense.''

The researchers conducting the study, which was called Interphone, were flummoxed at nearly every turn. They tried to find a control group that matched participants who had suffered a brain tumor, but potential subjects were reluctant to participate, for various reasons. Then there were subtle behavioral differences between individuals with and without brain tumors. Internal squabbling over how to interpret the results delayed publication for so long that usage patterns of study participants didn't match those of mobile users today.

The Interphone researchers acknowledged in their resulting paper, published online last week by the International Journal of Epidemiology, that something had probably gone wrong with the controls.

The study tracked cellphone use across 13 countries. It looked at a group of adults 30 to 59 years old who had been diagnosed with glioma or meningioma, types of brain tumors that can be either benign or malignant, between 2000 and 2004. They were compared with control subjects, people selected to match the individuals with tumors in terms of age, gender and place of residence.

Then both groups were interviewed extensively about their cellphone use. If the two groups matched in other ways, and the group with brain tumors used cellphones more frequently, that would suggest that cellphone use might have caused the tumors.

But they didn't really match. For one thing, just 53% of people selected to participate as controls agreed, and a survey of those who declined showed that they were less likely to use cellphones than those who participated. That may have artificially raised cellphone use in the tumor-free control group and made mobile phones seem less dangerous than they are.

The result is a strange set of numbers. Many levels of cellphone use appeared to reduce the chance of developing a tumor. Only the people who talked on cellphones the most had a significantly greater chance of developing glioma—40% greater—than those who didn't use cellphones.

Yet, as some of the study's authors themselves pointed out, if those who never used cellphones—who were more prevalent among those with tumors—were excluded, and the lightest users were contrasted with the more avid ones, then the bizarre protective effect of cellphone use mostly disappeared, and the risk among the heaviest users was 82% greater.

Even in this analysis, the risk doesn't steadily increase with use, which is what epidemiologists typically look for—a discernible dose-response relationship. ''It's certainly less compelling than if you saw some kind of graded response,'' says David A. Savitz, director of the Disease Prevention and Public Health Institute at the Mount Sinai School of Medicine in New York.

Disputes about how to interpret these numbers held up publication of the research, says Christopher Wild, director of the World Health Organization's International Agency for Research on Cancer, in Lyon, which coordinated the study. The study was published more than six years after its conclusion, by which time cellphone use had both surged and changed.

''Interphone made more effort than most other studies to identify and quantify its own flaws,'' says co-author Martine Vrijheid, a researcher at the agency. ''It has thereby also attracted more attention to these flaws.''

A U.K. study under way will take a different approach, tracking cellphone users over time to see if heavy use is tied to a greater incidence of cancer. But the study will still need to enlist hundreds of thousands of volunteers to yield useful results, and it could take decades to spot any divergence in cancer rates.

Epidemiologists say such research may be difficult and expensive but is important.

''Even if you think it's very, very unlikely that it's a problem,'' Dr. Savitz says, ''it's always worth some effort to make sure you haven't done something really terrible'' as a society by enabling widespread cellphone use. Such open questions, and the difficulty of solving them, he says, ''keep epidemiologists in business for a long time.''

Learn more about this topic at blogs.wsj.com/numbersguy/ .
Email: numbersguy@wsj.com.

Voor de bij het artikel behorende grafiek zie het bijgevoegde PDF bestand.

Voor het originele artikel zie:
online.wsj.com/article/SB10001424052748704269204575270482215602218.html .

Bekijk het PDF bestand


Lees verder in de categorie Berichten Internationaal | Terug naar homepage | Lees de introductie