Parents and guardians were interviewed by trained interviewers to capture sociodemographic and health information when their children were 9 months old, with subsequent follow-up at 3, 5 and 7 years. This study

was a cross-sectional inhibitor purchase survey using the baseline questionnaire (9 months postnatally) of the Millennium Cohort Study. Time spent in care as a child The definition of a looked-after child or a child in care varies between countries due to national legislation. In this cohort, mothers were asked the question: “Before the age of 17, did you spend any time living away from both of your parents?” If they answered yes, they were asked to indicate the nature of the time spent away from home and the amount of time they had spent away. Parents who had spent time in a children’s home or with foster parents, run by either a local authority or a charitable organisation, were coded as having been in care. This group included women who were not sure whether

their placement was managed by the local authority or any another organisation. Although children’s homes and foster placements can be run by voluntary societies, the responsibility for the child still lies with the local authority.27 We classed mothers who had spent some amount of time in foster care or a children’s home under the ‘exposed’ group. The comparison (‘unexposed’) group consisted of all mothers who had answered ‘no’ to the question: “Before the age of 17, did you spend any time living

away from both of your parents?”, or who had only spent time in a boarding school, prison or young-offenders’ institution, or with relatives. Mothers who did not answer the question or who indicated that they were unsure of their answer were excluded. Breastfeeding Mothers were asked if they ever tried to breastfeed their cohort baby. If they answered yes, they were asked when they had last given their baby breast milk. Their answer was converted into breastfeeding duration, and then categorised into: ‘never’, ‘less than 2 months’ ‘over 2 and less than 4 months’ and ‘over 4 months’. The information was also coded into a binary category of ‘never breastfed’ and ‘ever breastfed’. Smoking during pregnancy Maternal smoking was Cilengitide coded as ‘current non-smoker’, ‘smoked during pregnancy’ or ‘gave up smoking during pregnancy’. These categories were recoded as a binary outcome of ‘smoked during pregnancy’ and ‘did not smoke during pregnancy’. Symptoms of depression Symptoms of depression were measured using 9 questions of the validated Malaise Inventory,28 29 a tool used within the Millennium Cohort Study, to provide a measure of depression or psychological distress.30 It is a self-report tool phrased in plain language. There is no specified time frame over which participants are asked to report their symptoms, but the emphasis is on the recent past. Birth outcomes Information on a baby’s birth weight, gestation and delivery method was obtained by self-reporting.