Project Plan

 

EPoCH aims to address three main research questions:

1) Are maternal and paternal health behaviours in the prenatal period correlated with child health?

2) Are any of these associations causal? (Read more about what we mean by “causal”, and why this question is important here)

3) How do causal maternal/paternal effects interact or mediate each other?

Data

  • EPoCH draws together detailed and diverse data that has already been collected through several large European birth cohort studies. Each of these cohorts has detailed information on thousands of families, collected during or around the time of pregnancy, and at regular intervals ever since:
    • The Avon Longitudinal Study of Parents and Children (ALSPAC)
      • Over 14,500 children born in the Bristol area of the United Kingdom between 1991-1992
    • Born in Bradford (BiB)
      • Nearly 14,000 children born in Bradford, United Kingdom, between 2007-2010
    • The Cleft Collective (CC)
      • Around 3,000 children born with a cleft lip and/or palate anywhere in the United Kingdom from 2013-present
    • The Danish National Birth Cohort (DNBC)
      • Around 100,000 children born anywhere in Denmark between 1996-2003
    • Generation R (GenR)
      • Nearly 10,000 children born in Rotterdam, the Netherlands, between 2002-2006
    • The Norwegian Cohort of Mothers and Children (MoBa)
      • Over 95,000 children born anywhere in Norway between 1999-2008

We are using birth cohort data on:

  • Parental health behaviours – how parents behave before and during pregnancy in terms of:
    • Physical activity,
    • Smoking
    • Alcohol intake
    • Caffeine intake
    • Fat intake
    • Sugar intake
  • Childhood health outcomes – how the health and development of babies and children up to 11 years old in terms of:
    • Body size and composition: birthweight, head circumference, height, body mass index, weight-to-height ratio, fat mass index
    • Neurodevelopmental and psycho-social traits: IQ, social development, educational attainment
    • Congenital anomalies: cleft lip/palate
  • Genetics of parents and children
  • Other factors that might influence our results: age of parents, socioeconomic status, chronic disease, toxins, mental health, stressful life events, relationship between parents, how many children parents have, parents’ behaviours after the birth of their child, etc.

Methods

We’re going to compare and combine (meta-analyse) data from each birth cohort to try to address our three research questions:

1) We’ll use multivariable regression analysis to identify associations between parental health behaviours and childhood outcomes.

2) We’ll combine (“triangulate”) evidence generated using several different approaches to infer whether parental health behaviours actually cause differences in offspring health, or whether the observed correlations arise because of other “confounding” factors. The approaches we will use here are:

  • Mendelian randomization
  • Parental comparison negative control design
  • Discordant sibling comparison

3) Where it looks like the relationship is causal, we will use several several statistical methods to explore whether the childhood health outcome is influenced by the behaviour of mothers, fathers, or a combination of both. The statistical methods we will use are:

  • Conventional mediation and interaction analysis
  • Network Mendelian randomization
  • Counterfactual theory
  • The life course approach

You can read more about EPoCH on our blog.