A525: Causal inference of smoking, nutrition, alcohol and physical activity on birth weight: a sibling analysis.

Causal inference of smoking, nutrition, alcohol and physical activity on birth weight: a sibling analysis.

Approved: 2014

Collaborators:

AProf Alexis Hure
Prof Julie Byles
Ms Jennifer Powers
Dr Catherine Chojenta
Prof Deborah Loxton
Dr Melissa Harris
Ms Peta Forder
Dr Amy Anderson
Dr Ellie D Arcy
Ms Sheree Harris
Prof Liz Holliday

Synopsis

Residual confounding is a major problem in observational studies, meaning that results must be interpreted with caution. Data from siblings provides the opportunity to better control for family characteristics, thereby reducing residual confounding and increasing causal inference. Sibling analysis allows for a separation of within-family effects from between-family effects and simultaneous estimation of both. We will apply sibling analysis to birth weight data from the state-based Perinatal Data Collections linked to the 1973-1978 ALSWH cohort. We intend to focus on the role of four key modifiable lifestyle factors: smoking; nutrition (diet and weight); alcohol; and physical activity, on birth weight.