Objective To determine which factors are the best predictors of poor maternal and child outcomes up to age 5, so that parent support program can better target interventions to those who will benefit most.

Methods The Avon Longitudinal Study of Parents and Children is a prospective birth cohort of 14 541 pregnant women. Childhood development was assessed with a parent-reported developmental scale at 18 months (n=7969), the Strengths and Difficulties Questionnaire at 47 months (n=8328) and teacher-reported School Entry Assessment scores at 4–5 years (n=7345). Maternal outcomes were depression at 8 weeks (n=10 070), never breastfeeding up to 6 months (n=7976), feelings of unattachment (n=8253) and hostility (n=8159) at 47 months, and not in employment, education or training (n=8265) at 61 months.

Results Few families with each poor outcome (3% to 9%) had mothers aged <20 years when they were pregnant. Half to three-quarters of families with poor outcomes could be identified if information on all six predictors was used and a woman had at least one of these. Model discrimination (area under the receiver operator characteristic curve) improved from approximately 0.50 for all outcomes using maternal age only, up to 0.80 for postnatal depression when all six predictors were included in the model. Calibration also improved with the model including all six characteristics.

Discussion Factors other than young maternal age, including education, smoking and depression during pregnancy should be considered in identifying those eligible for programs aimed at improving child and maternal outcomes through intensive parent support.