OBJECTIVE: To determine the strength of the association between antenatal psychosocial risk factors and adverse postpartum outcomes in the family, such as assault of women by their partner, child abuse, postpartum depression, marital dysfunction and physical illness. DATA SOURCES: MEDLINE, Cinahl, Famli, Psych Abstracts and the Oxford Database of Perinatal Trials were searched from relevant articles published from Jan. 1, 1980, to Dec. 31, 1993, with the use of MeSH terms "depression, involutional," "child abuse," "child neglect," "domestic violence," "family," "marital adjustment," "family health," "newborn health," "child health," "physical illness," "social support," "psychosocial risk," "prediction," "risk factors," "obstetrics" and "prenatal care." Further articles were identified from bibliographies. STUDY SELECTION: Of the 370 articles identified through the search, 118 were included for review. Studies were included if they examined the association between psychosocial risk factors and the outcomes of interest. Articles were excluded if they were reviews of poor quality or they had one or more of the following features: insufficient description of the sample, a high attrition rate, a lack of standardized outcome measures, outcomes other than the ones of interest or results that had already been reported in a previous study. DATA EXTRACTION: The strength of evidence of each study was evaluated. On the basis of the evidence, each risk factor was assigned a rating of the strength of its association with each of the postpartum outcomes. The ratings were class A (good evidence of association), class B (fair evidence) and class C (no clear evidence). Of the 129 antenatal psychosocial risk factors studied, 15 were found to have a class A association with at least one of the postpartum outcomes. DATA SYNTHESIS: Child abuse and abuse of the mother by her partner were most strongly correlated (class A evidence) with a history of lack of social support, recent life stressors, psychiatric disturbance in the mother and an unwanted pregnancy. Child abuse was also strongly associated with a history of childhood violence in the mother or her partner, previous child abuse by the mother's partner, a poor relationship between the mother and her parents, low self-esteem in the mother and lack of attendance at prenatal classes. Postpartum abuse of the mother was also associated with a history of abuse of the mother, prenatal care not started until the third trimester and alcohol or drug abuse by the mother or her partner (class A evidence). Child abuse had a fair (class B) association with poor marital adjustment or satisfaction, current or past abuse of the mother and alcohol or drug abuse by the mother or her partner. There was class B evidence supporting an association between abuse of the mother and poor marital adjustment, traditional sex-role expectations, a history of childhood violence in the mother or her partner and low self-esteem in the mother. Postpartum depression was most strongly associated with poor marital adjustment, recent life stressors, antepartum depression (class A evidence), but was also associated with lack of social support, abuse of the mother and a history of psychiatric disorder in the mother (class B evidence). Marital dysfunction was associated with poor marital adjustment before the birth and traditional sex-role expectations (class A evidence), and physical illness was correlated with recent life stressors (class B evidence). CONCLUSIONS: Psychosocial risk factors during the antenatal period may herald postpartum morbidity. Research is required to determine whether detection of these risk factors may lead to interventions that improve postpartum family outcomes.