A recent study from our group indicated that antenatal anxiety is associated with a significantly increased risk of behavioural/emotional problems in young children (Reference O'Connor, Heron and GoldingO'Connor et al, 2002). The study was based on the ALSPAC cohort, a prospective, longitudinal study of women followed since pregnancy. Analyses indicated that antenatal anxiety at 32 weeks' gestation was associated with an approximately 2-fold increase in behavioural/emotional problems in boys and girls at age 4 years; these associations were observed after accounting for key antenatal, obstetric and psychosocial risks, and postnatal anxiety and depression. The findings are important in providing the strongest evidence to date that the substantial evidence for long-term effects of antenatal stress/anxiety found in numerous animal investigations (e.g. Reference Schneider, Moore and NelsonSchneider & Moore, 2000) may extend to humans.

In our paper, the focus was on whether or not the antenatal environment had a role in the development of behavioural/emotional problems, an issue with substantial implications for our understanding of development, as well as for prevention and public health. Dr Barlow's letter helps draw attention to a separate research base linking behavioural/emotional problems in children with postnatal factors, particularly parent—child relationship quality. Although there remain some controversial matters in that field of research, especially concerning causal mechanisms (see Reference O'ConnorO'Connor, 2002), parent—child relationship quality is certainly a robust predictor of children's psychological development. Given the multiple-risk nature of development and psychopathology, we would agree with Dr Barlow that there is a need to bring together findings from different lines of research and to revise our models and theories that consider multiple levels of risk. Indeed, there are a number of directions for this research to pursue, including the consideration of how postnatal experiences such as parent—child relationship quality moderate the effects of antenatal anxiety/stress and how the role of genetic factors may explain individual differences in response to antenatal anxiety/stress. Research along these lines is underway. Because it has tracked women intensively since pregnancy and has continued to collect information on a wide range of biological and psychosocial variables, the ALSPAC study is an especially important resource for studies of this kind.