Mobile phones: where's the uncertainty?

As of the 23rd May 2022 this website is archived and will receive no further updates. was produced by the Winton programme for the public understanding of risk based in the Statistical Laboratory in the University of Cambridge. The aim was to help improve the way that uncertainty and risk are discussed in society, and show how probability and statistics can be both useful and entertaining.

Many of the animations were produced using Flash and will no longer work.

The media in the UK and many other countries have been full this week of stories about mobile phones and brain cancer. Some were really pretty scary - the Daily Express gave a SHOCK WARNING: MOBILE PHONES CAN GIVE YOU CANCER. But others were much more cautious. The BBC reported that a link between phones and cancer was "not clearly established" and that "the evidence was too weak to draw strong conclusions from." What's going on?

Possible effects of mobile phone use on health have been discussed, often vigorously, for many years. The whole story of how the media have dealt with the controversy is a fascinating one, that I'm planning to write more about later. But the fact that there's controversy indicates that there must be some uncertainty. I want to examine just one aspect of that uncertainty.

The cause of this week's media fuss is that the International Agency for Research on Cancer (IARC), part of the World Health Organization (WHO), issued a press release announcing that they have just decided to classify the radiation that comes from wireless phones as "possibly carcinogenic to humans".

It's important to realise that this classification isn't based on original research on groups of people using phones (or not); instead, a group of 31 experts spent a week at the IARC offices in Lyon, France, assessing evidence from existing studies. Now some of the evidence they looked at has not quite been published yet, so there is something new here apart from the classification - but so far all we have is the press release. It promises that a proper report on their conclusions will be published very soon, but that hasn't actually appeared yet.

So nothing in the press release, that was behind the media fuss, actually said how big the risk actually might be. All the IARC have reported so far is a classification of the strength of evidence for a risk, not on the size of the risk, if there is one.

So there are two kinds of uncertainty here. It may (or may not) be true that mobile phone use causes some brain cancers - if it does, then we're not talking about all phone users getting a brain tumour, we're talking about an increased risk of a brain tumour. Even if we knew exactly what the size of the risk was, it wouldn't tell us exactly who would get a tumour and who wouldn't. There would still be uncertainty.

But that's not the kind of uncertainty being discussed in the IARC press release. It is classifying the uncertainty in the evidence. They classify all the possible cancer-causing agents they have considered into five categories, described here.

These range from Category 1, where they consider that the agent definitely causes cancer, through 2A ("probably carcinogenic to humans") and 2B ("possibly carcinogenic to humans"), through to Group 4.

Group 4 ("probably not carcinogenic to humans") contains only one substance (caprolactam, a substance used in making artificial fibres). In a way that's not surprising - the IARC would not investigate something unless there was some kind of reason to believe it might cause cancer. Things that nobody has ever dreamed of causing cancer will simply not be on their lists at all But it's interesting to note that the biggest group is 3, "Not classifiable as to its carcinogenicity to humans". That is, very often the IARC concludes there's just not enough evidence to say either way. (The actual lists of possible carcinogens are here.)

However, they did put mobile phone use in Group 2B. That means that there is a considerable degree of uncertainty about whether there is any risk at all - if that doubt was smaller, they would have put it in Group 2A or possibly even 1. Despite the headlines in some of the newspapers, we're talking here about considerable uncertainty on whether there is any risk, because (in the IARC's own words) the evidence for a causal link is in a few cases "limited" and in other cases "inadequate". Nothing is established very firmly - and that's why the IARC is calling for more research. It's not a SHOCK, and it's not a U-turn as the Sun headline said. We had limited or inadequate evidence of a risk before, and that's still the position. Nothing has really changed.

Most of the media studies said nothing at all about how big the risk of mobile phone use might be, if it exists. The fact that there's a lot of uncertainty about whether a risk exists does not, on its own, tell us much about the size of the risk if it does exist. (It could happen that some substance has a big effect on the chances of cancer, but that this substance has not been studied much, so that there is still uncertainty about whether the risk does exist.)

But with mobile phones and brain tumours, the evidence that we have so far suggests that the risk is pretty small. This is partly because, despite the fear that most people have for these particular diseases, they are actually uncommon. You are not at all likely to suffer from a brain tumour. It's possible (though very far from certain), IARC say, that using a mobile phone a lot may increase your chances of a brain tumour a bit, but the vast majority of phone users will still not get one. And if in the future the existence of the risk is eventually more firmly established, you would still have to think about how an increased risk of a rare disease balances out again the benefits of using a mobile phone.

Most of the media reports said nothing about this aspect. There were, however, sensible and balanced discussions on the Cancer Research UK website and the NHS Choices website. If you want to get some idea of how big the risks might actually, be, look at these. And Ben Goldacre has explained why it's very difficult to measure this sort of risk here.

I'm not going to throw my mobile away yet...

(Updated to add Goldacre link, 4 June 2011)