BACKGROUND Length of postnatal hospital stay has declined dramatically in the past thirty years. There is ongoing controversy concerning whether or not staying less time in hospital is harmful or beneficial. OBJECTIVE The objective of this review was to assess the safety, impact and effectiveness of a policy of early discharge for healthy mothers and term infants, with respect to the health and well-being of mothers and babies, satisfaction with postnatal care, overall costs of health care and broader impacts on families. METHODS We searched the Cochrane Pregnancy and Childbirth Group trials register (April 2002), the Effective Practice and Organisation of Care Review Group specialised register of clinical trials, the Cochrane Controlled Trials Register (The Cochrane Library, Issue 4, 2001), MEDLINE (1966 to 2001), CINAHL (1982 to 2001), EMBASE (1988 to 1993) and reference lists of articles. METHODS Randomized trials comparing early discharge from hospital of healthy mothers and term infants, of greater than or equal to 2500 grams, with standard care in the settings in which trials were conducted. METHODS Trial quality was assessed and data were abstracted independently by all five reviewers. RESULTS Eight trials were identified involving 3600 women. There was substantial variation in the definition of 'early discharge', and the extent of antenatal preparation and midwife home care following discharge offered to women in intervention and control groups. Five trials recruited and randomized women in pregnancy, three randomized women following childbirth. Post randomization exclusions were high. Protocol violations occurred in both directions. No statistically significant differences in infant or maternal readmissions were found in six trials reporting data on these outcomes. Three trials had mixed results showing either no significant difference or results favouring early discharge for the outcome of maternal depression although none used a well-validated standardised instrument. The results of six trials showed that early discharge had no impact on breastfeeding although significant heterogeneity was present between studies. CONCLUSIONS The findings are inconclusive. There is no evidence of adverse outcomes associated with policies of early postnatal discharge, but methodological limitations of included studies mean that adverse outcomes cannot be ruled out. It remains unclear how important midwifery support at home is to the safety and acceptability of early discharge. Large well-designed trials of early discharge programs incorporating process evaluation to assess the uptake of co-interventions, and using standardised approaches to outcome assessment are needed.