Prenatal cocaine exposure and neurobehavioral development: How subjects lost to follow-up bias study results

Abstract

Ninety-four prenatally cocaine-exposed and 50 non-cocaine-exposed infants were enrolled at birth and seen at 3,6, and 12 months. Applying a recently developed model (Cicchetti&Nelson, 1994) produced four independent groups: those infants assessed at birth only (n=61), those returning for at least one (n=29) or two (n=15) assessments after birth, and those returning for all three (n=35). Among those infants having all follow-up assessments, cocaine-exposed infants showed significantly lower psychomotor performance on the Bayley Scales of Infant Development (BSID) than did non-exposed infants. Results approached statistical significance for mental development. Except for birthweight (cocaine-exposed infants weighed significantly less), none of a wide range of maternal, perinatal, or neurobehavioral demographic variables biased study results or the generalization of findings to the total sample from those who returned for all visits.