Smoke Signals

Smoke Signals

Mobile phone health alarmists bereft of credible arguments

Age-adjusted brain cancer rates have flatlined over nearly 30 years. Giuseppe Milo/Flickr, CC BY-NC

In May this year, I led a paper published in Cancer Epidemiology, which looked at the incidence of brain cancer in Australia between 1982 and 2012.

The first mobile phone call was made in Australia in 1987 and today their use is all but universal.

Cancer is a notifiable disease: all newly diagnosed cases are gathered from doctors by state cancer registries and nationally aggregated by the Australian Institute of Health and Welfare in publicly available data.

I summarised our study in this column, which to date has had more than 44,700 readers.


New study: no increase in brain cancer across 29 years of mobile use in Australia


We found that with extremely high proportions of the population having used mobile phones across some 20-plus years (from about 9% in 1993 to about 90% today), age-adjusted brain cancer rates have flatlined over nearly 30 years.

There were significant increases in brain cancer incidence only in those aged 70 years or more. But the increase in this age group began from 1982, before the introduction of mobile phones in 1987 and so cannot be explained by it.

The most likely explanation of the rise in this older age group was improved diagnosis that happened with the introduction of imaging machines that (for example) could more accurately diagnose some strokes as brain cancers.

In the days and weeks after publication, our paper received massive global news and social media attention, achieving an Altmetric score of 835. On the basis of the most media-covered research in all fields in 2015, this would have put it just outside the 100 highest Altmetric scores if we’d published it last year (2016 figures are published early next year).

It also drew the ire of the close-knit international network of mobile phone and wifi alarmists, who are utterly convinced mobile phones are deadly and won’t hear otherwise. Their opening salvo was to accuse me of being an undeclared phone industry stooge.

In 1997 I had been given a small grant by AMTA, the Australian Mobile Telephone Association, to conduct a national survey of how many mobile phone users had ever used their phone to call emergency services such as ambulance, police and fire. Large proportions of people had done so, probably saving many lives by alerting these services far more quickly than when having to find a landline.

I didn’t report this because I got the one-off grant 19 years ago, and all reputable journals and research agencies rule that competing interests are not lifetime but extinguish typically between one and three years after such support has expired. The grant also had nothing to do with cancer.

I also got a series of mostly verbally incontinent email. One from an excitable correspondent in Swaziland, insisted that I answer his many eureka moment insights into why what we had published was wrong in every respect. We should withdraw our paper, he demanded and tell the world we were wrong.

Predictably, several wrote to Cancer Epidemiology, setting out a litany of our egregious errors and failures to understand that an epidemic of brain cancer, comparable to the deluge of smoking-caused cancers, was just around the corner. Three of these were published this week with our response (open access until October 20, 2016).

The three letters were written by five individuals, three of whom are affiliated with a non-accredited Environmental Health Trust, headed by Dr Devra Davis, the alarmist doomsayer who featured in the much-criticised ABC Catalyst program which has now been withdrawn.

Assuming they got their heads together to rain blows on our heretical findings, it was amusing to see the barely audible blanks they decided to fire.

Their main arguments were:

‘It’s too soon to see an epidemic of brain cancer’

One argued several decades of widespread phone use were needed before increases in cancer might be seen. She seemed intent on diminishing the number of years that large numbers of Australians have used mobile phones, in order to preserve her argument. She argued that only the last nine years of data since 2001 when mobile subscriptions reached 50% of the population ought to be considered in any analysis. And nine years was not nearly enough.

But by 1996, some 20% of Australian adults (some 2.9 million) were using mobile phones. Apparently we ought to have joined her in seeing this as a trivial exposed population, unworthy of consideration. Quite obviously, there’s no alleged carcinogen where 20% of the population is exposed where any credible scientist would seriously maintain such widespread exposure should be ignored in assessing population attributable risk.

Further, in one of the studies cited in a review published by our critics, excess risks of brain cancer from mobile phone use are argued as occurring following exposures of as little as between five and ten years of mobile phone use. These critics even suggested in the same paper that the international INTERPHONE study may suggest a cancer “promotion effect”, with use as few as one to four years being dangerous.

We concluded that:

This therefore looks like an argument trying to walk on both sides of the street: if a short latency period show excess risks they are deemed to be credible, while if they show no excess (as with our study) they are to be dismissed.

‘Various case-control studies show evidence of increased risk’

Case-control studies in this field have been criticised because they rely on users’ recall of the extent of phone use going back many years. Just try recall your own mobile phone use in, for example, 2003 and you will immediately understand how data obtained this way are hugely problematic.

Moreover, people with brain cancer often have memory loss. And if you have brain cancer, are part of a study considering its cause, and have been exposed to frequent claims about the hypothesis that mobile phone use causing brain cancer, the likelihood of recall bias resulting in recall of high mobile phone use is probably going to increase.

The strength of our study was the ability to look at all cases of brain cancer in Australia in the 29 years since the first call was made here. The inconvenient fact for the alarmists is that there has been no significant increase in brain cancer in either men or women compatible with the mobile phone hypothesis.

‘Decreased use of X-rays is masking an increase in cancers caused by mobiles’

Perhaps the silliest argument thrown at us was an unreferenced hypothesis that “discontinued or reduced use of established carcinogens such as X-rays” may have reduced the incidence of brain cancer from such exposures while, simultaneously, the rise of mobile phone use would have replaced those cases, thereby explaining the largely flat line incidence across our data period.

This hypothesis would need to account for how reductions in a very uncommon radiation exposure (full head X-rays) could ever possibly produce the exact same decreased incidence of brain cancer that they claim arise in daily exposure to an alleged carcinogen by most of the entire population would add to that incidence.

Our Swaziland critic finished one of his missives writing that “it behooves you, as a scientist, to take note of fatal errors in your work.” It would “behoove” mobile phone alarmists to stop unnecessarily alarming people with their weak arguments.