A total of 509 women with a history of a low-birthweight (LBW, less than 2500 g) baby were recruited from the antenatal booking clinics of four hospitals and randomized to receive either a social support intervention in pregnancy in addition to standard antenatal care (the intervention group) or standard antenatal care only (the control group). At recruitment to the study, mean gestational age was 6 weeks, mean maternal age was 28.0 years, 86% of the women had one previous LBW baby, 11% had two and 2% had had three or more. The study population was socially disadvantaged: 77% of the women were working class, 18% had unemployed partners and 41% were smoking at booking. Social support was given by four research midwives in the form of 24-h contact telephone numbers and a programme of home visits, during which the midwives provided a listening service for the women to discuss any topic of concern to them, gave practical information and advice when asked, carried out referrals to other health professionals and welfare agencies as appropriate, and collected social and medical information. Pregnancy outcomes were assessed using obstetric case-note data (obtained for 507 women) and a postal questionnaire sent to all mothers 6 weeks after delivery (94% replied). Babies of intervention group mothers had a mean birthweight 38 g higher than that of control group babies; there were fewer very low-birthweight babies in the intervention group.(ABSTRACT TRUNCATED AT 250 WORDS)