Wednesday November 4 2009
“Smoking during pregnancy significantly increases the risk of having a child with behavioural problems,” The Guardian has reported. This finding is based on a study which followed the early years of around 13,000 children.
The study found that, at the age of three, children whose mothers smoked throughout pregnancy were significantly more likely to have behavioural problems than children whose mothers did not smoke. Boys whose mothers smoked were also more likely to have hyperactivity and low attention spans. The study has some limitations, such as using the mothers’ reports as the sole source of data on smoking habits, child behaviour and other social and health factors. In this type of study, it is difficult to remove the influence of factors other than smoking that may have affected the results, although the study did take a number of these into account.
Smoking carries health risks for the smoker, and this study suggests that it may also affect children in the womb.
Where did the story come from?
Dr Jayne Hutchinson and colleagues from the University of York and University of Illinois in Chicago carried out this research. Support for the writing of this paper came from the US National Institute for Drug Abuse, and the study’s researchers were financially supported by grants from various national organisations and charitable foundations. The study was published in the peer-reviewed Journal of Epidemiology and Community Health.
The study was reported in The Guardian and BBC News, which featured generally balanced coverage.
What kind of research was this?
This was a cohort study which looked at the relationship between smoking in pregnancy and child behaviour at the age of three. The researchers report that previous studies have found links between smoking in pregnancy and disruptive behaviour in school-aged boys. However, they say that the evidence involving younger children and girls is less clear, and that previous studies have often not taken into account other maternal factors.
Cohort studies, which follow a group of people over time, have the best design for investigating this type of relationship. It would be unethical to perform a randomised controlled trial in which mothers were asked to smoke during pregnancy.
However, because the groups in cohort studies are not randomly assigned, there can be differences between them other than the factor of interest (smoking in this case). For example, the researchers in this study report that women who smoke during pregnancy are more likely to have lower socioeconomic status and social support, and to experience more stressful life events. These differences may influence the results of the study if they are not taken into account during analysis. The researchers in this study did take into account a range of factors.
What did the research involve?
The researchers used data on around 13,000 children, which had been collected in a large cohort study called the Millennium Cohort Study (MCS). The children in the MCS were all born in 2000 and 2001, and were all on the UK Child Benefit Register. The sample contained a high proportion of families from deprived areas and areas with large numbers of ethnic minorities.
As part of this study, the families filled in questionnaires when the children were aged nine months and three years. At the nine-month assessment, mothers reported whether they had smoked before or during pregnancy, and how much they smoked. Women were classified as: never smoking in pregnancy, quitting smoking in pregnancy, light smoking throughout pregnancy (fewer than 10 cigarettes a day) or heavy smoking throughout pregnancy (10 or more cigarettes a day).
When the children were three years old, the mothers rated their behaviour using a questionnaire. In particular, the researchers were interested in conduct problems (such as temper tantrums or fighting) and hyperactivity–inattention problems (such as restlessness, fidgeting or lack of concentration). The 10% of children who had the highest levels of problems (based on their scores in these questionnaires) were considered to have behavioural problems.
Solely relying on mothers to self-report smoking habits, social factors and children’s behaviour may reduce the reliability of the researchers' findings. Women who smoked during pregnancy may have been reluctant to report it and the interpretation of children’s behaviour is subjective. Obtaining confirmation of data from different sources (for example, medical records, fathers or teachers) could increase the reliability of the results.
The researchers used statistical methods to look for links between maternal smoking and child behaviour. Boys and girls were considered separately to see whether smoking affected them differently. The researchers also looked at whether smoking was linked to having either conduct problems or hyperactivity–inattention problems, or to having a combination of both.
In their analyses, the researchers took into account important social and psychosocial factors that could affect the results. These factors included: mother’s age at the birth, number of children in the family, family stability, household poverty, low maternal education, social class of the parents, problematic relationships and parenting, poor daily life function (such as homelessness), health related behaviours (including postnatal smoking) and measures of mother and child health.
What were the basic results?
According to the mothers’ self-reported smoking habits:
- 65.5% never smoked in pregnancy,
- 12.4% quit smoking in pregnancy,
- 12.5% were light smokers throughout pregnancy, and
- 9.6% of women were heavy smokers throughout pregnancy.
Women who smoked in pregnancy, particularly heavy smokers, were more likely to be working class, to come from poor households and to have lower educational achievement, reduced family stability, more problematic relationships and more health-related problems.
After taking into account other factors that could affect results, the researchers found that boys whose mothers smoked throughout pregnancy were more likely to have conduct and hyperactivity–inattention problems compared to boys whose mothers did not smoke. The risk of the boys having these behavioural problems increased the more their mothers smoked.
Once the researchers had taken into account other factors that could affect results, there was no significant link between maternal smoking throughout pregnancy and behavioural problems in girls.
In a second set of analyses, the researchers excluded children who had both conduct and hyperactivity–inattention problems. In these analyses, they found that heavy smoking in pregnancy was linked to an increased risk of having conduct-only problems in both boys and girls. Both light and heavy smoking were linked to an increased risk of hyperactivity–inattention alone in boys but not girls.
Daughters of women who quit smoking during pregnancy were at lower risk of conduct problems (with or without other problems) or combined conduct and hyperactivity–inattention problems. However, few daughters had these problems, so these results may not be very reliable.
How did the researchers interpret the results?
The researchers concluded that there is an association between smoking throughout pregnancy and problems of conduct and hyperactivity–inattention in three-year-old boys. Smoking throughout pregnancy was also associated with conduct problems alone in three-year-old girls.
Smoking carries health risks for the individual who smokes, and this and other studies suggest that it may also affect children whose mothers smoke while they are pregnant. This is yet another reason for women who smoke to quit.
The study has the advantages of a large sample size and the fact that it assessed a wide range of factors that could have affected the results. There are also some limitations to the study:
- The sample only included children on the UK Child Benefit Register and over-sampled people from poor areas and areas with high proportions of ethnic minorities. In addition, not all the people who were asked to participate in the study did so. This means that the results may not be representative of the population as a whole.
- As with all studies of this type, the groups compared may be imbalanced for factors other than the factor of interest (smoking in this case), and this may affect results. In this study, the researchers took into account a large number of sociodemographic and other factors, which increases the reliability of the results. However, many of the factors that the researchers adjusted for are complex (such as problematic relationships) and it may be difficult to measure these factors and remove their influence entirely. There may also be other unknown or unmeasured factors that have an effect. For example, most of the factors assessed related to the mother only.
- The study carried out multiple statistical tests and this increases the likelihood that significant differences may be found by chance. The fact that some links were consistently significant across analyses, such as the link between smoking and behavioural problems in boys, suggests that the findings may be more likely to be valid.
- Smoking, behaviour and other factors assessed in this study were all based on reports by the mothers. It is possible that the mothers may not have accurately recalled or disclosed details of how much they smoked in pregnancy. In addition, mothers may have different perceptions of what constitutes argumentative or overactive behaviour. In addition, factors such as parenting may be best assessed by looking at both parents. All this could affect the results.
- The study identified problem behaviours by selecting the children with the highest scores on the behavioural questionnaire. Although these children had the highest levels of these behaviours within the sample, it is not clear if their behaviour could be clinically diagnosed as a conduct disorder or attention deficit hyperactivity disorder (ADHD).