Data were collected in four ways. First, 950 hospital records of children admitted to the ED's and PD's of both hospitals were examined for nurses' documentation about child abuse. Second, purposive sampling was used to identify 11 ED and PD nurses at venue A with experience in communicating about abused children. Venue B was only used for the first stage of data collection due to internal management issues. Each participant responded verbally to two vignettes and then participated in an individual, 1-hour semistructured interview conducted by the researcher. Third, the researcher made entries in a journal to document contextual issues, so as to note observations and general conversations during visits to the ED and PD of venue A. Fourth, each form of data was examined separately for themes and patterns of inherent behavior in the nurses' communication practices. The children's records from both venues were examined for nurses' entries about child abuse. The study found that the nurses were influenced by a number of factors when communicating about suspected or confirmed child abuse, including hierarchical structures and protocols; the maintenance of therapeutic relationships; reactions of other professionals; and the loss of control over interventions and consequences. The nurses' concerns about these issues contributed to behavior that initiated and maintained specific departmental and organizational responses to cases of child abuse. Some nurses denied the existence or significance of child abuse. The data from both venues indicate that although the ED nurses acknowledged the existence of child abuse, they were likely to resist being involved by discounting child abuse. If an ED doctor overlooked child abuse, then an ED nurse was unlikely to indicate it either. It is essential that nurses and other mandated reporters of suspected child abuse be made aware of the relevant legislation and the consequences for themselves and the abused children should they fail to report suspected child abuse, regardless of the various work-setting factors that might impede such reporting. 30 references