To systematically locate and review studies that have investigated the incidence of medication errors (MEs) in pediatric inpatients and identify common errors. A systematic search of studies related to MEs in children was performed using the following databases: MEDLINE (1951-April 2006), EMBASE (1966-April 2006), Pharm-line (1978-April 2006), International Pharmaceutical Abstracts (1970-April 2006), Cumulative Index to Nursing and Allied Health Literature (1982-April 2006), and British Nursing Index (1994-April 2006). Studies of the incidence and nature of MEs in pediatrics were included. The title, abstract, or full article was reviewed for relevance; any study not related to MEs in children was excluded. Three methods were used to detect MEs in the studies reviewed: spontaneous reporting (n = 10), medication order or chart review (n = 14), or observation (n = 8). There was great variation in the definitions of ME used and the error rates reported. The most common type of ME was dosing error, often involving 10 times the actual dose required. Antibiotics and sedatives were the most common classes of drugs associated with MEs; these are probably among the most common drugs prescribed. Interpretation of the literature was hindered by variation in definitions employed by different researchers, varying research methods and setting, and a lack of theory-based research. Overall, it would appear that our initial concern about MEs in pediatrics has been validated; however, we do not know the actual size of the problem. Further work to determine the incidence and causes of MEs in pediatrics is urgently needed, as well as evaluation of the best interventions to reduce them.