Maternal characteristics and pregnancy outcomes in the NICE birth cohort: an assessment of self-selection bias.

Background: Prospective birth cohorts are essential for identifying associations between exposures and outcomes. However, voluntary participation introduces a potential bias due to self selection since the persons that chose to participate may differ in background characteristics and behaviors.

Objectives: To investigate potential bias due to self-selection in the (NICE) birth cohort in northern Sweden.

Methods: Women in the NICE birth cohort ( = 621) were compared to nonparticipating pregnant women in Norrbotten County in northern Sweden who were eligible for participation ( = 4976) regarding maternal characteristics and lifestyle. Maternal characteristics and pregnancy outcomes were compared between the groups and associations between exposures (smoking, folic acid, BMI, parity, education) and pregnancy outcomes (birth weight and gestational age) were analyzed by linear regression analyses, examining any interaction with the group.

Results: NICE participants were more highly educated, older and more likely to cohabit than the non-participants. They more often took folic acid and multivitamin supplements and less often smoked during early pregnancy. Pregnancy outcomes (mode of delivery, gestational age at delivery, birth weight and APGAR score) did, however, not differ significantly between participants and non-participants. Smoking, BMI, education and parity affected gestational age and birth weight, but the associations were of similar magnitude in participants and non-participants, with no significant effect on the group.

Conclusion: Self-selection to the NICE study was evident in some factors related to lifestyle and socioeconomic characteristics but did not appear to skew pregnancy outcomes or alter well-known effects of certain lifestyle parameters on pregnancy outcomes.