April 28, 2021

expert reaction to study looking at childhood air pollution exposure and mental health at age 18

A study published in JAMA Network Open looks at air pollution exposure in childhood and adolescence and mental health at the transition to adulthood.

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“The results of this study do provide evidence of an association between air pollution levels where young people lived, and measures of their mental health, with higher pollution levels being associated with worse mental health.  What they can’t do is to show that it’s the high air pollution that actually causes the poorer mental health.  Maybe it does, in part anyway, or maybe it doesn’t – the research results can’t tell us.  So that leaves the position unclear on what, if anything, should be changed as a result of these findings.  That said, there are plenty of other reasons for wanting to clean up our air anyway, and if it does also turn out that such changed could also improve young people’s mental health, that’s all to the good.

“The main issue in interpreting the findings is that the study was observational.  It involved over 2,000 people born in England and Wales in 1994-5 who were originally recruited for a study of environmental risks using twins, though the fact that the participants were twins seems not to have been used in this particular study.  Put simply, the researchers used a (good) data-based model to estimate the air pollution levels outside the places where the young people lived at the ages of 10 and 18.  Two pollutants were measured: nitrogen oxides (NOx) and very small particles (PM2.5).  When they were 18, the young people were interviewed about a wide range of aspects of mental health issues, and these were statistically processed by the researchers to provide an overall general measure of mental health for each of them, as well as three separate measures of more specific aspects of mental health.  The researchers then carried out analyses that showed that higher levels of nitrogen oxide air pollution (NOx) were associated with higher scores on the overall measure of mental health (indicating worse mental health).  There was also some evidence of an association between tiny particles in the air (PM2.5) and worse mental health, particularly in comparing the young people who lived in areas with the highest particulate pollution with those who lived in areas with cleaner air on this measure.

“The problem in interpretation is that there will be a great number of differences between the young people, and their families, who live in areas with different pollution levels, apart from just the differences in pollution.  Places with high air pollution levels are likely to be in different parts of towns and cities than those with lower pollutant levels, and indeed are more likely to be in large towns and cities than in the country.  The people who live in more polluted areas would generally differ from those in less polluted areas in terms of the jobs they do, how well off they are, and many more aspects.  So any differences in the average mental health of young people living in areas with high and low air pollution might be caused by some combination of all these other differences, rather than by the pollution differences.  The researchers were, of course, aware of that, and they made statistical adjustments to try to take account of some of the differences, on which they had collected data.  So they made adjustments for the young person’s gender, the socioeconomic status of the family the psychiatric history of the family, whether the young person had previous emotional or behavioural problems, and whether they smoked.  In additional analyses they also made adjustments for various characteristics of the neighbourhoods where people lived.  Those adjustments generally didn’t make major differences to the patterns of associations they had observed.  But those aren’t all the ways in which people living under different pollutant levels might differ, so there’s no way to be sure about what’s causing what.  The researchers can’t adjust for factors on which they do not have data – for instance, they did not adjust for smoking by the parents (or other people the participants lived with), or indeed any other aspects of air pollution inside people’s homes.  The researchers are, rightly, explicit in their report about saying that their study can’t establish cause and effect.

“They also point out a range of other limitations of their findings.  They didn’t measure pollutant levels in the young people’s early lives, before school (and they took into account pollutant levels below age 10 only indirectly).  Evidence on other possible harmful effects of air pollution have strongly suggested that exposures at those young age might have important health effects of various kinds.  The fact that the study looked at only two air pollutants means that it can’t say much about what specific aspects of the air pollution might be important.  Areas that have high levels of one air pollutant usually have high levels of many others as well, so the evidence of association between nitrogen oxides and mental health could be there because of an association with a different pollutant.  For instance, the researchers point out that lead was not banned from petrol in the UK till these young people were about six years old, in 2000, so that people who were exposed to high NOx levels, which very often arise from road traffic, may well have been exposed to high lead levels as young children – and it’s well known that exposure to lead has a strong impact on some mental issues.  And high levels of pollutants caused by traffic are associated with high levels of traffic noise, and some other studies have linked traffic noise to various mental health issues.

“It’s difficult to relate the numerical results of this study to measures of mental health from other studies, because the researchers used a measure specifically constructed for their study from a range of other psychiatric data.  So how big is the problem anyway?  The researchers do report that the association they found was ‘modest’, and say that this indicates that outdoor air pollution is therefore “unlikely to be a major etiological factor in an individual’s psychiatric illness risk across short time spans” – in other words, that outdoor air pollution isn’t likely to be a decisive influence in causing mental illnesses in individual young people.  But, if the association does turn out to be one of cause and effect, air pollution could still be important in psychiatric public health, simply because so many people are exposed to air pollution levels above current guidelines.  So, while this single study can’t give clear evidence on exactly what kind of association there is between poor air quality and mental health, it’s yet another potential reason for being concerned about air pollution levels.”

‘Association of air pollution exposure in childhood and adolescence with psychopathology at the transition to adulthood’ by Aaron Reuben et al. was published in JAMA Network Open at 16:00 UK time on Wednesday 28 April 2021.

DOI: 10.1001/jamanetworkopen.2021.7508

Declared interests

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  However, my quote above is in my capacity as an independent professional statistician.”