Abstract
The goal of this study was to examine the utility of clinical records of psychiatrically hospitalized youth for predicting critical incidents during treatment. Intake data from inpatients (N = 484, mean age = 14 years) in a secure psychiatric facility were coded for the presence of theoretically based individual and contextual risk factor information and analyzed prospectively to predict youths' involvement in incidents of seclusion and restraint. Findings indicated that whereas several individual and contextual risk factors accounted for the likelihood of a youth becoming involved in seclusion or restraint, only histories of various types of aggression, number of prior residential placements and body mass index could predict the extent of this involvement. The implications of these findings with respect to ecologically valid research and empirically informed practice with high-risk youth are discussed.