Predicting maltreatment of children of teenage mothers

To determine the degree to which knowing certain characteristics about young high-risk families can help distinguish those families most likely to maltreat their children from those families at lower risk of maltreating their children.

Observational cohort from which the following predictor variables were gathered when infants were 2 months old: maternal age, depressive symptoms, childrearing attitudes, social support, and living situation (with or apart from related adults). Families were followed up for 24 months to identify the occurrence of maltreatment.

An urban, socioeconomically disadvantaged cohort of teenage mothers and their infants attending a hospital-based special primary care clinic for teen mothers and their infants.

All full-term infants and mothers enrolled into the clinic in 1990 participated in the study. This included 47 mother-infant pairs enrolled when infants were 2 months of age. Forty-five of these pairs were available for follow-up when infants were 24 months of age.

Maltreatment defined as any incident that prompted investigation by the state child protective agency and was found to be a substantiated case of maltreatment by that agency.

Maltreatment occurred in 15 of 45 families before the child's second birthday. Discriminate function analysis produced a model that correctly classified 13 of 15 maltreating mothers and misclassified one of 30 non-maltreating mothers. Stepwise analysis revealed that living situation was by far the strongest predictive variable (R2 = 7).

Maltreatment was a predictable outcome within this extremely high-risk cohort. Living apart from related adults was the strongest risk factor associated with maltreatment. This easily obtainable piece of information may be an important risk marker for practitioners, social service personnel, and others working with this very-high-risk population. It may allow early supportive interventions that might prevent maltreatment.