LONDON (Reuters) - The frustratingly inconclusive results from the world’s biggest study so far into possible links between mobile phone use and cancer are symptomatic of problems that can dog scientific research like this.
Campaigners, cancer doctors and mobile phone manufacturers have been waiting for a decade to see if the findings of the study led by the respected International Agency for Research into Cancer (IARC) would finally provide an answer.
It didn’t and experts say the biases and potential errors that rendered the study unreliable are difficult to avoid, yet very hard to adjust for.
“This was a very complex study, and results were very difficult to interpret because of a number of methodological issues,” said Elisabeth Cardis at the Centre for Research in Environmental Epidemiology in Barcelona, who led the group of 21 international scientists conducting the study.
Jack Siemiatycki, an epidemiologist at the University of Montreal Hospital Research Centre in Canada, described the outcome as “ambiguous, surprising and puzzling.”
The problem, say the researchers, is how the study was done.
The Interphone study was an epidemiological case-control study that began with cases -- people with brain tumours -- and controls -- people with no cancer -- and asked them to remember how much they had used mobile phones in previous years.
Experts say that case-control studies can be useful in establishing whether a disease is associated with a certain exposure or lifestyle, but they are also susceptible to a number of possible biases that mean results can be unreliable.
One, known as “selection bias” or “participation bias,” comes about because of the voluntary nature of taking part in a study.
Almost 13,000 people were covered in the Interphone study and one potential problem may be that some of those who took part because they have a brain tumour did so because they already believed the disease was caused by using a mobile phone.
This could skew their estimations of how much they used a cell phone and how often they held it on one side of their head where their tumour appeared rather than the other.
The criteria for asking people to participate were also likely to vary from country to country, scientists said, and adjusting for this without knowing details is almost impossible.
And awareness of media coverage of campaigners voicing fears about links between cell phones and cancer may distort things even more.
“It is possible that the participants did not provide an accurate portrait of cell phone usage,” said Siemiatycki, who was one of the scientists who worked on the study.
Another possible flaw is so-called “recall bias” which comes about when participants in a study may misremember the length of time they spent on the phone because the study is not being done in real time.
“All these things are estimates, and sometimes people don’t estimate very well,” said Cardis. “We also know that heavy users tend to overestimate, and light users tend to underestimate.”
But the Interphone study is not alone in showing weaknesses.
MANY SUCH STUDIES HAVE BIASES AND FLAWS
Scientists analysing the accuracy of research into mobile phones and cancer -- which in many years has always failed establish a link -- reported last year that studies vary widely in quality.
In an assessment of 23 published studies involving more than 37,000 people, they found that results often depended on who conducted the study and how well they were able to adjust findings for bias and other errors.
According to Mireille Toledano, an epidemiologist at Imperial College London, case-control studies are particularly problematic when they are used for something like mobile phones.
If the study were looking at another exposure, say asbestos at work, she explained, scientists would have other documents such as employer records, clocking-in times and other pieces of data to help build up a more objective picture.
“The problem with mobile phones is that when you’re trying to do a retrospective study like this, you’re almost totally dependent on people recalling. You haven’t got anything else on which to base your data,” she said.
Toledano is helping lead a newly launched study on mobile phones and health by European scientists, which aims to track around a quarter of a million of people in five countries for up to 30 years.
She hopes this study, called a prospective study because it is forward-looking, will be more accurate because it tracks healthy people in real time to assess whether their mobile phone use is linked to health developments as years go by.
Until then, and despite Interphone’s inconclusive results, scientists say people shouldn’t panic about using mobile phones.
“If there are risks, they are probably pretty small,” said Siemiatycki. “Should anyone be concerned about potential dangers of cell phones, they can remedy the issue by using hands-free devices and avoid exposure to radio frequencies around their head.”
Editing by Maggie Fox and Reed Stevenson
Our Standards: The Thomson Reuters Trust Principles.