Prediction of Cognitive Ability With Social Determinants in Children of Low Birth Weight.

Only 133 of 204 children who met initial criteria completed at least one outcome measure at about age 9, and the authors note possible attrition bias. No individual predictor was statistically significant, so the effect sizes for single social factors cannot be treated as reliable stand-alone markers. The model used only information available at birth or before hospital discharge, so later family, neighbourhood, and health experiences were not captured. Early intervention services received by children were not controlled for, which may have affected the apparent relationships with birth weight and gestational age. The cohort was born in 1998 to 2000, and changes in social stigma and neonatal care may limit generalisation to current patients. Estimates for some racial and ethnic groups may be less stable because few children belonged to those groups.

The authors report no conflicts of interest.

Do not treat the final model as a validated screening tool for individual children. No single predictor was statistically significant, and the authors caution that the reported effect sizes may capitalize on chance in a small sample.