Work Package 1
Methods: I used longitudinal data from eight birth cohorts containing participants from twelve countries (Australia, Belgium, Denmark, France, Germany, Greece, Italy, Netherlands, Poland, Norway, Spain and the United Kingdom). The number of included children in each cohort ranged from N=584 (Greece) to N=73,042 (Norway), with a total sample of N=149,604. Child socio‐economic circumstances (SEC) were measured using self‐reported maternal education at birth. Child mental health outcomes were internalising and externalising problems measured using either the Strengths and Difficulties Questionnaire or the Child Behavior Checklist. I modelled the slope index of inequality (SII) using sex‐stratified multi‐level models.
Results: For almost all cohorts, at the earliest age of measurement children born into more deprived SECs had higher internalising and externalising scoress than children born to less deprived SECs. For example, in Norway at age 2 years, boys born to mothers of lower education had an estimated 0.3 (95% CI 0.3 0.4) standard deviation higher levels of internalising problems (SII) compared to children born to mothers with high education. The exceptions were for boys in Australia (age 2) and both sexes in Greece (age 6), where we observed minimal social inequalities. In UK, Denmark and Netherlands inequalities decreased as children aged, however for other countries (France, Norway, Australia and Crete) inequalities were heterogeneous depending on child sex and outcome. For all countries except France inequalities remained at the oldest point of measurement.
Work Package 2
Methods: I used data from 13 EU birth cohorts including up to 36,819 children. Exposure to up to 27 aspects of the urban exposome in pregnancy were estimated (built environment, natural spaces, ambient air pollution, noise) using land-use regression models or other established methods. Three mental health outcomes were used: internalising, aggressive behaviour and ADHD symptoms, square-root transformed to achieve normality and converted to within-cohort z-scores. Associations between each exposure and outcome were estimated using separate linear regression models within each cohort and combined using 2-stage meta-analysis. False-discovery rate was used to correct for multiple testing. In order to test which were the most influential exposures, I also conducted LASSO regression as a variable selection tool. As in work package 1, all analysis was conducted using the DataSHIELD platform.
Results: For child internalising, we found no associations between any environmental exposures and child mental health outcomes. For aggressive behaviour, we found associations between the following aspects of the urban environment and aggressive behaviour: transport land use, connectivity density, facility density, number of bus lines, number of bus stops, NDVI, sive of green & blue space and road traffic noise. Analysis with LASSO regression retained one variable: access to blue spaces. Finally, we found associations between two aspects of the urban environmemt and ADHD symptoms: connectivity density and PM2.5 (air pollution). Analysis with LASSO regression also retained one variable: access to blue spaces.
Work Package 3
Methods: Eight EU cohorts will be included with available data on urbanicity and the personal exposome stressors. The exposure will be level of urbanization at birth, based on the Global Human Settlement Layer (GHSL-SMOD) which stratifies the household residence of participants according to the three levels of the degrees of urbanisation. The outcomes will be exposome variables classified in five subgroups: (i) diet, (ii) lifestyle, (iii) maternal health (fetal period only), (iv) childcare (pre-school period only) and (v) exposure to mental health disorders. The personal exposome will be analysed in three different periods: (i) fetal life, (ii) pre-school age (0-4 years) and (iii) early childhood (5-8 years). For each time period, separate logistic regression models will be fit for each exposome variable with the degree of urbanisation variable as the exposure. Associations will be visualised using volcano plots and/or tabular format.
