Earlier this week Metro, The Sun and The Daily Mail all ran a stories on a piece of research showing how women who use mobile phones whilst pregnant could be putting their unborn children at risk of developing behavioural problems later in life(1). Mobile phones are used by a large majority of the population today and good evidence to suggest they caused harm would certainly be important.
Teams based in Denmark and the USA carried out the research covered in these newspaper stories and it was a repeat of a study by the same authors that was published in 2008(2). The pregnant women and children investigated were part of the Danish National Birth Cohort(3). When this cohort was established, the researchers aimed to look at factors during pregnancy and early childhood that may go on to cause diseases later in life. At the time of setting it up they were particularly interested in studying the side effects of medications and infections.
The first study the researchers published in 2008 found an association between mobile usage by pregnant women and behavioural problems in children aged seven. At the time the authors weren’t sure whether the association was a chance finding or due to the fact that other confounding factors (things that could have influenced the results other than mobile phones) might explain association. They therefore set out to repeat the study, but this time look more closely at some additional factors that might explain the associations.
In the latest study the authors asked a different set of women whose children were seven years old, and born between 1998 and 2002, a series of questions about mobile phone usage, including: when they first started regularly using mobile phones; how much they used the phones during pregnancy; whether they used a hands-free device and if so how often; the location of the phone when they carried it (handbag or pant/shirt/jacket pocket); and the current use of cell phones. All the information about how much a mobile phone was used during pregnancy was therefore asked seven years after the event.
The behavioural problems analysed in this study were assessed using a screening tool called the strengths and difficulties questionnaire(4). This questionnaire has been validated and shown to identify childhood psychiatric disorders including: conduct, hyperactivity, depressive and some anxiety disorders; specific phobias; separation anxiety; and eating disorders. In other words it can reliably identify a very wide range of childhood behavioural problems. The researchers combined the overall score from this questionnaire and categorised children into normal, borderline or abnormal for having ‘overall behavioural problems’. They have therefore combined a series of fairly broad range of conditions into one group.
The main findings of the research were that less than 1% of the seven year olds used mobile phone for more than 1 h per week but 35% were using a phone at some point. Around 18% of the children had exposure to mobile phones before and after they were born, and 40% of the children had no exposure. Regarding the behavioural problems, 93% of the children had no recorded behavioural problems, 3% were considered borderline and 3% of children scored an abnormal result. In terms of the associations between mobile phone use and behavioural problems, they found that children whose mothers had used a mobile phone in pregnancy, and children who themselves used mobile phones at the age of seven, were more likely to have behavioural problems.
If these associations were true then this would be a worrying finding. However, as with much of epidemiology, interpreting the results is key. The authors are themselves relatively cautious about whether they think mobile phones do cause all these behavioural problems, however, they feel that because the results of the two studies are consistent, it provides stronger evidence than previously.
I would also be extremely cautious in interpreting these findings. As pointed out above, part of the measures of mobile phone usage (and the ones which had the strongest association with behavioural problems) were based on the recall of information seven years after the pregnancy. There is a potential bias within this data in that women with children that have behavioural problems might remember their use of phones differently to those who don’t. As the cohort study didn’t set out to originally look at the question of mobile phone use, this was probably the best way the researchers could get information on this important question. However, because it was such a long time ago and people aren’t very good at remembering even very important events that have happened to them like being diagnosed previously with cancer, it’s probably not a particularly accurate and reliable measure of mobile phone use.
The behavioural problems that were examined included an extremely broad category of disorders. Because of the groupings we can’t be sure if the results were due to a very strong association with one or two particular behavioural disorders increasing the overall risk of the combined group, or whether it was in fact only present when the groups were combined and as a result of weaker associations with each disorder. If the second scenario were the correct explanation for the findings it would imply that mobile phone use increased the risk of a very different and wide range of conditions from eating disorders through to hyperactivity and depression. This is not my area of expertise, but this scenario seems highly implausible to me given that these are a series of very different disorders and for them to have the same underlying mechanism would be very unusual indeed.
There are a number of other possibilities that might explain the associations found in the studies, including other factors effecting behaviour of the children that have not been measured. As the research was not set up originally to look at these factors we do not have the data to examine these issues with a high level of accuracy. In this second study, the researchers did try to look at one issue that they hadn’t done previously: the level of attention mothers paid to their children. As they didn’t have a direct measure of this, instead they used breastfeeding in the first six months as a way to examine attention level of mothers to the children i.e. women who breast fed their children were more attentive. They found it wasn’t important and didn’t explain the associations, however, I would suggest that this is a fairly rough and ready way to measure attentiveness and it probably doesn’t capture it very well. In epidemiology it’s surprising how sometimes such measures work, so this isn’t a criticism of the researchers, it was a good thing to try to account for, however, I think it didn’t quite work and still other explanations probably exist that haven’t yet been analysed.
By repeating the study the researchers have reproduced the findings from their previous research. In this new study they found similar, although reduced associations, which they feel are likely to reduce the possibility that the first results were due to chance. This does make chance less likely as an explanation, but as the two studies were conducted in virtually identical populations using the same methodologies it is not entirely surprising that the results were compatable. Stronger evidence for this association being causal (i.e. mobile phones creating the behavioural problems) and not due to chance would be provided by carrying out the study, or a similar one, in different settings and countries to see if results are still consistent.
Mobile phones are used so widely today that if there is potential for them to cause disease that public health doctors and scientists would want to know about it and act upon that information. This is therefore an important research question and it’s a good thing that the authors have published their findings – the difficulty is in the interpretation.
Other researchers have looked at whether mobile phones might lead to cancer and again the results had to be interpreted with great caution(5). There was a suggestion of a link, but overall it was felt the current data and analysis could not answer the question. Unfortunately then further studies will need to look at these issues but until we have the results from these we’re probably not going to be able to answer the question definitively. Until then don’t throw away your mobile phone and have a look at the Department of Health advice which strikes a sensible balance on the basis of current evidence with regards to children: they should only use phones when essential and keep calls short.
1. Divan HA, Kheifets L, Obel C, Olsen J. Cell phone use and behavioural problems in young children. Journal of Epidemiology & Community Health [Internet]. 2010 12 [cited 2010 Dec 9];Available from: http://jech.bmj.com/content/early/2010/11/11/jech.2010.115402?q=w_jech_ahead_tab
2. Divan HA, Kheifets L, Obel C, Olsen J. Prenatal and postnatal exposure to cell phone use and behavioral problems in children. Epidemiology. 2008 Jul;19(4):523-529.
3. Olsen J, Melbye M, Olsen SF, Sørensen TI, Aaby P, Andersen AM, et al. The Danish National Birth Cohort–its background, structure and aim. Scand J Public Health. 2001 Dec;29(4):300-307.
4. GOODMAN R, FORD T, SIMMONS H, GATWARD R, MELTZER H. Using the Strengths and Difficulties Questionnaire (SDQ) to screen for child psychiatric disorders in a community sample. The British Journal of Psychiatry. 2000 Dec 1;177(6):534-539.
5. Brain tumour risk in relation to mobile telephone use: results of the INTERPHONE international case-control study. Int J Epidemiol. 2010 Jun;39(3):675-694.