Dec 08 2010

Cell Phones and Behavior

Cell phones continue to be a focus of epidemiological studies and public concern, despite the fact that so far there is no compelling evidence of any health risk from cell phones. Concerns are likely to be sparked anew with the report of a study linking cell phone use to behavioral problems in children.

The study, by Divan, Kheifets, Obel, and Olsen, is a follow up of a prior study which showed a correlation between cell phone use in pregnant women and behavior problems in their children. They sought to replicate this study with a larger data set and taking into consideration more possible confounding factors. They found:

Results The highest OR for behavioural problems were for children who had both prenatal and postnatal exposure to cell phones compared with children not exposed during either time period. The adjusted effect estimate was 1.5 (95% CI 1.4 to 1.7).

Conclusions The findings of the previous publication were replicated in this separate group of participants demonstrating that cell phone use was associated with behavioural problems at age 7 years in children, and this association was not limited to early users of the technology. Although weaker in the new dataset, even with further control for an extended set of potential confounders, the associations remained.

For children of mothers who used cell phones, and who also used cell phones themselves, there was a 50% relative increase in reported behavioral problems. For children who did not use cell phones but whose mothers did while pregnant, the relative increase was 40% – these relative increases are with an absolute baseline risk of 2% reporting behavioral problems (so a 50% relative increase would be a 1% absolute increase in risk). This is a small, if statistically significant, effect. An odds ratio of 1.5 is a reasonable effect size, depending on what is being measured. In this case the outcomes were based upon questionnaires, and are therefore reliant on subjective reporting.

The results of this study are interesting, but the usual caveats apply. The absolute effect size is small, and the outcomes based upon subjective reporting. This is an observational study, not an experimental study, so no conclusions can be drawn regarding cause and effect. The study authors did not even speculate about a possible mechanism, and would not suggest that cell phone use causes the observed behavior problems.

It should also be noted that while this study replicated a previous study, the effect size was smaller. Also, accounting for other possible correlates, such as social status and maternal history of behavioral problems, tended to decrease the effect size as well (although not eliminate it).

It is possible that the correlation is spurious or is an artifact of reporting bias. If the correlation is real, it seems likely that there are confounding factors that have not been sufficiently accounted for. The authors themselves considered as many as possible. They report:

“One thought was that it was it not cellphone use but mothers’ inattention that led to behavior problems. While it was important, it didn’t explain the association that we found.”

So inattention partially explained the results – perhaps other behavioral or personality attributes also correlate with both cell phone use and behavior problems. Given that the absolute effect size is small (about 1%) there can be many such factors that can exert such an effect, and it is unlikely that the authors thought of all of them.

The plausibility of a direct effect of cell phone use is quite low. Physicists have pointed out that cell phones generate only small amounts of energy, the primary effect of which is to heat up surrounding tissue. But the non-ionizing radiation from cell phones, by definition, is of too low energy to break chemical bonds or cause any cellular damage. Concerns over cell phone use have largely surrounded the risk of brain cancer, and epidemiological studies have not shown increase in brain cancer overall or an increased risk correlating with cell phone use. The data does not entirely clear cell phones either, but so far the evidence does not support a link.

Given the low biological plausibility and generally negative epidemiological data, it is unlikely that cell phones directly cause any biological harm. This also applies to behavioral problems, as investigated in this study.


While sure to garner headlines, this study is not compelling evidence that cell phones are a health concern with regard to behavioral problems in children. The study found a small absolute effect size which is very likely to be due to confounding factors that were insufficiently accounted for in this study. It is very difficult to derive any firm conclusions from observational studies such as this. Many such studies are required to pinpoint a likely cause and effect relationship – no one observational study can do this.

I do not even think that this study warrants precautionary measures, as the likelihood of a real risk from cell phones is small and implausible. Cell phone use is increasingly common, and epidemiological studies such as this are certain to turn up a great deal of noise in the data – apparent correlations or the downstream effects of complex behavioral influences. There are likely to be more studies showing similar apparent correlations and it would be premature to respond to each one with calls for precautionary measures.

At best this study suggests questions for future research. But the results are within the predictable noise of preliminary epidemiological data that is best ignored by the public. Most such correlations do not hold up with future research and the ultimate interpretation of such correlations is often different from the most direct cause and effect that grabs headlines and public attention. Basing decisions on preliminary epidemiological evidence such as this is likely to be unnecessarily disruptive and inconvenient and is unlikely to provide any health benefit.

11 responses so far