AIMS: To review the characteristics of 462 mothers consecutively admitted to a co-joined mother and baby unit and to examine the incidence and predictors of social services input and separate discharge. METHOD: Admissions to the Birmingham Mother and Baby Unit from 1998 to 2007 were reviewed. Demographic and clinical information, social services involvement and parenting outcome data were collected from case notes and computerized records. RESULTS: One hundred and forty six (146) women received some form of social services input (32%); 44 cases of separate discharge were identifi ed (10%). Diagnosis, involuntary admission, single status, younger age, Afro-Caribbean ethnicity and social services referral predicted parenting concern. Only 9% of separations were to women suffering a new postpartum mental illness; most were to women experiencing the continuation of an existing illness. Poor parenting skills, absence of a good partner relationship and alcohol/substance misuse were frequently reported in those discharged without their baby. CONCLUSION: The incidence of separate discharge is lower than in other studies. Potential reasons for this difference are discussed. Further work must be conducted to understand the social inequities and regional differences in social service intervention during perinatal psychiatric admission and determine if parenting outcomes can be improved in women at risk of separation.